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 |
|---|---|---|---|---|---|
Nasal colonization by both gram-positive and gram-negative pathogens induces expression of the innate immune protein lipocalin 2 ( Lcn2 ) . Lcn2 binds and sequesters the iron-scavenging siderophore enterobactin ( Ent ) , preventing bacterial iron acquisition . In addition , Lcn2 bound to Ent induces release of IL-8 from cultured respiratory cells . As a countermeasure , pathogens of the Enterobacteriaceae family such as Klebsiella pneumoniae produce additional siderophores such as yersiniabactin ( Ybt ) and contain the iroA locus encoding an Ent glycosylase that prevents Lcn2 binding . Whereas the ability of Lcn2 to sequester iron is well described , the ability of Lcn2 to induce inflammation during infection is unknown . To study each potential effect of Lcn2 on colonization , we exploited K . pneumoniae mutants that are predicted to be susceptible to Lcn2-mediated iron sequestration ( iroA ybtS mutant ) or inflammation ( iroA mutant ) , or to not interact with Lcn2 ( entB mutant ) . During murine nasal colonization , the iroA ybtS double mutant was inhibited in an Lcn2-dependent manner , indicating that the iroA locus protects against Lcn2-mediated growth inhibition . Since the iroA single mutant was not inhibited , production of Ybt circumvents the iron sequestration effect of Lcn2 binding to Ent . However , colonization with the iroA mutant induced an increased influx of neutrophils compared to the entB mutant . This enhanced neutrophil response to Ent-producing K . pneumoniae was Lcn2-dependent . These findings suggest that Lcn2 has both pro-inflammatory and iron-sequestering effects along the respiratory mucosa in response to bacterial Ent . Therefore , Lcn2 may represent a novel mechanism of sensing microbial metabolism to modulate the host response appropriately .
Bacteria utilize iron for electron transport , amino acid synthesis , DNA synthesis , and protection from superoxide radicals [1] . Under aerobic conditions , iron is primarily in the ferric ( Fe ( III ) ) oxidation state and readily forms insoluble complexes . Sequestration of the scarce quantities of soluble iron is a prototypical protective response against invading bacteria , mediated by the iron-binding proteins transferrin and lactoferrin and the storage protein ferritin . To acquire iron within the host , bacteria secrete siderophores such as enterobactin ( Ent ) that bind ferric iron with greater affinity than mammalian proteins . In the thrust and parry between bacteria and their host to obtain iron , a new form of competition has been identified . Lipocalin 2 ( Lcn2 , also known as NGAL , siderocalin and 24p3 ) is a member of the lipocalin family of small-molecule transport proteins [2] . Lcn2 specifically binds Ent with an affinity similar to the Escherichia coli Ent receptor [2] and competes with bacteria for Ent binding . Lcn2 is able to bind both ferric and aferric Ent , thereby depleting Ent from the microenvironment and inhibiting bacterial uptake of Ent-bound iron . As a result , Lcn2 is bacteriostatic . Bacterial growth can be restored by the addition of excess iron [2] or Ent [3] . In a murine sepsis model , serum Lcn2 is protective against an E . coli strain that requires Ent to obtain iron . Accordingly , Lcn2-deficient mice ( Lcn2−/− ) succumb to infection [3] . Conversely , co-injection of E . coli and a siderophore to which Lcn2 cannot bind causes lethal infection in Lcn2+/+ mice [3] . Originally isolated from the specific granules of neutrophils , Lcn2 is also found in mucus producing cells of the respiratory tract [4]–[6] . In mice , the nasal mucosa responds to Streptococcus pneumoniae colonization by increasing Lcn2 mRNA expression ( 65-fold ) [6] . Consequently , protein levels increase in olfactory glands and respiratory and olfactory epithelial cells . Lcn2 is secreted into the nasal lumen and bathes the colonized mucosa . Lcn2 is also induced by Haemophilus influenzae colonization , suggesting its production is a general response to colonizing bacteria [6] . In addition to sequestering iron , Lcn2 can act as a signaling molecule . A murine Lcn2 receptor , 24p3R , has been identified and is widely expressed in tissues including the lung and in lymphoid and myeloid cells [7] . In lymphocytic cells , 24p3R is able to internalize Lcn2 alone or Lcn2 bound to a siderophore . Internalization of Lcn2 bound to an iron-loaded siderophore increases the intracellular iron concentration . However , internalization of Lcn2 bound to an iron-free siderophore depletes intracellular iron levels by binding to cellular iron followed by export from the cell through recycling endosomes . In respiratory epithelial cells , Lcn2 elicits chemokine release [8] . In A549 and other human respiratory epithelial cell lines , incubation with aferric Ent produces a dose-dependent increase in the secretion of the chemokine IL-8 [8] . This response is potentiated by the addition of Lcn2 to form aferric Ent-Lcn2 . In contrast , ferric Ent ( Fe-Ent ) does not elicit significant IL-8 release . Whether Lcn2 induces chemokine release and neutrophil recruitment during bacterial infection is unknown . Perhaps due to the actions of Lcn2 , successful pathogens do not typically depend solely on Ent for iron [9] . Pathogens such as Klebsiella pneumoniae often produce additional siderophores [9] , [10] that Lcn2 is not predicted to bind [11] . Salmonella enterica and certain K . pneumoniae and E . coli isolates also encode the iroA locus that can counteract Lcn2 [12]–[17] . This cluster encodes the Ent glycosylase IroB that prevents Lcn2 binding , and IroC to export , IroN to import , and IroD to linearize glycosylated Ent [18] , [19] . Transformation of E . coli with the iroA locus is sufficient to cause lethal infection in Lcn2+/+ mice [19] . Conversely , disruption of either the iroC exporter or iroB glycoslyase attenuates virulence in a mouse model of systemic Salmonella infection [20] . In K . pneumoniae , this locus is associated with strains causing invasive disease ( pyogenic liver abscesses ) in patients [17] . Lcn2 could represent a new paradigm of mucosal immunity based on recognition of bacterial siderophores leading to direct iron sequestration and subsequent pro-inflammatory signaling . To test the antibacterial effects of mucosal Lcn2 , a respiratory pathogen that produces Ent is required . A gram-negative , non-motile , encapsulated member of the family Enterobacteriaceae , K . pneumoniae colonizes both the nasopharynx and large intestine of humans [21] , and is a common cause of bacterial pneumonia and sepsis . The wild-type K . pneumoniae strain ATCC 43816 produces Ent , contains the iroA locus , and produces a second siderophore , Yersiniabactin ( Ybt ) [22] . By exploiting defined K . pneumoniae siderophore mutants and Lcn2-deficient mice , we determined whether Lcn2 inhibits bacterial colonization by sequestering iron , promoting inflammation , or both .
Previously a murine model of K . pneumoniae pneumonia was established and characterized [22] , [23] . However , in order to test the role of Lcn2 in the upper respiratory tract we needed to develop a nasal colonization model . To do this , C57BL/6 mice were inoculated intranasally with increasing concentrations of K . pneumoniae in 10 µL of PBS . To facilitate recovery from the non-sterile nasopharynx , the strain KPPR1 , a rifampin resistant-derivative of K . pneumoniae ATCC 43816 was used . Mice were sacrificed at day 1 , 3 , 7 , or 14 and colonization was measured by tracheal cannulation and quantitative culture of nasopharyngeal lavage , as previously described [24] . Using 2×106 cfu as the inoculum , sustained colonization at >5×103 cfu/ml was observed for at least 7 days ( Figure 1 ) . Whereas instillation under anesthesia causes pneumonia [22] , [23] , inoculation of awake mice caused nasal carriage without systemic disease . K . pneumoniae was recovered only sporadically and at low density ( <7×102 cfu/organ ) from the lungs and spleen ( data not shown ) . Furthermore , all mice appeared healthy throughout the course of the experiment despite a 1000× larger inoculum compared to the pneumonia model . To characterize the cellular response to K . pneumoniae nasal colonization , mice were inoculated as above and sacrificed at day 1 and day 3 , skulls were dissected and decalcified , and saggital sections of the nasopharynx were examined by histology . On day 1 , an acute inflammatory infiltrate in the paranasal airspaces was apparent in hematoxylin-eosin stained sections ( Figure 2A ) . The lumen of the nasopharynx harbored dense clusters of neutrophils , as judged morphologically by high power light microscopy ( Figure 2B ) . On day 3 , an infiltrate of similar density and location was seen . No neutrophils were seen in the lumens of unchallenged mice ( data not shown ) . To directly quantify the neutrophil influx in response to K . pneumoniae , flow cytometry was performed on nasal washes of KPPR1-colonized and PBS mock-colonized mice . Activated neutrophils were identified by the following markers: CD45 ( pan-hematopoietic marker ) , Ly6G ( Granulocyte receptor-1 ) and CD11b ( Mac-1 ) . Aliquots of 100 µL/animal were stained with fluorophore-conjugated antibodies , and total numbers of neutrophils were compared ( Figure 2C–F ) . On day 3 , a marked increase in the percentage of total events that were CD45+ , CD11b+ , Ly6G+ neutrophils was seen in colonized versus uncolonized mice ( Figure 2G , p<0 . 01 ) . To determine if neutrophils are protective during K . pneumoniae colonization , mice were depleted of neutrophils by the Ly6G-specific antibody RB6-8C5 . One day prior to K . pneumoniae inoculation , mice were injected intraperitoneally with RB6-8C5 or control Rat IgG . Mice were inoculated intranasally with KPPR1 and subsequently sacrificed for nasopharyngeal lavage and quantitative spleen culture . At day 1 post-inoculation , RB6-8C5-treated mice had significantly greater colonization density compared to control-treated mice ( p<0 . 01 , Figure 3A ) . Four out of five RB6-8C5-treated mice were bacteremic , as judged by recovery of K . pneumoniae from the spleen , compared to none ( 0/5 ) of the control-treated mice ( p<0 . 05 , Figure 3B ) . By day 2 , RB6-8C5-treated mice colonized with K . pneumoniae became moribund ( data not shown ) . These data indicate that neutrophils control intranasal density of colonizing K . pneumoniae and prevent bacteremia and sepsis . By exploiting isogenic siderophore mutants , K . pneumoniae can be used to evaluate two non-mutually exclusive functions of Lcn2: iron sequestration and pro-inflammatory activation . K . pneumoniae KPPR1 makes glycosylated Ent ( Gly-Ent ) and Ybt , neither of which Lcn2 is predicted to bind [11] . Constructing mutants in the iroA locus will allow study of the effects of Lcn2 binding to non-glycosylated Ent ( Ent ) . However , whether Ybt can compensate for the loss of Ent bound by Lcn2 is unknown . The role of Ybt and Gly-Ent in a mouse model of K . pneumoniae pneumonia was tested previously [22] . To determine the requirement of each siderophore during nasal colonization , cfu recovery of single and double mutants was compared to wild-type K . pneumoniae on day 3 post-inoculation ( Figure 4 ) . As determined by the density of ybtS and entB mutants , Gly-Ent or Ybt were sufficient to support K . pneumoniae nasal colonization . In contrast , the entB ybtS mutant strain lacking both siderophores was deficient in colonization ( p<0 . 001 at day 3 ) . These data indicate that siderophores are required for persistence on the nasal mucosa , but either Gly-Ent or Ybt is sufficient to scavenge the necessary iron . This is in contrast to the pneumonia model where the ybtS mutant had a distinct disadvantage compared to the entB mutant at later stages of infection [22] . To generate K . pneumoniae producing unmodified Ent that can be bound by Lcn2 , the iroA locus was disrupted in the wild type and the ybtS mutant . The resulting iroA ybtS mutant is predicted to depend on unmodified Ent to acquire iron and therefore to be susceptible to Lcn2-mediated interference with iron acquisition . In contrast , the iroA mutant is predicted to make unmodified Ent , which may initiate Lcn2-mediated inflammation , but to be able to acquire iron using Ybt . To test the requirement of the iroA locus for growth in the presence of Lcn2 , wild-type and siderophore mutant K . pneumoniae were grown in serum from Lcn2−/− mice with or without recombinant murine Lcn2 ( rLcn2 , 1 . 6 µM ) . Serum was chosen as a growth medium where Ent is required to obtain iron from transferrin [25] , [26] . Production of either Ent or Gly-Ent was sufficient for maximal growth in Lcn2−/− serum ( Figure 5 , white bars ) . In contrast to colonization , Ybt was unable to support growth in mouse serum in vitro ( see entB mutant ) . Addition of rLcn2 ( shaded bars ) inhibited Ent-producing K . pneumoniae ( iroA , iroA ybtS mutants ) by approximately 1000-fold compared to Gly-Ent-producing K . pneumoniae ( wild type , ybtS mutant ) . To confirm that Lcn2-sensitivity of the iroA and iroA ybtS mutants is attributable to disruption of the iroA locus , complementation studies were performed . The mutants contain a disruption in the initial gene iroB that may have polar effects on the rest of the operon . Transformation with plasmids containing iroB from K . pneumoniae KPPR1 or E . coli χ7122 was toxic and inhibited growth even in the absence of Lcn2 ( data not shown ) . Therefore polar effects could not be ruled out . Instead , the mutants were transformed with a plasmid pIroA ( pIJ137 ) encoding iroBCDN from E . coli χ7122 . In E . coli , this construct confers the ability to produce Gly-Ent as determined by liquid chromatography-mass spectrometry [27] . Transformation with pIroA , but not the control plasmid , rescued growth of the iroA and iroA ybtS mutants in Lcn2-containing serum ( Figure 5B ) . This indicates that the iroA locus is required to prevent Lcn2-dependent growth inhibition . In E . coli and Salmonella strains encoding iroA , only a subset of enterobactin is glycosylated [20] , [27] . Comparison of wild-type K . pneumoniae growth in the presence or absence of Lcn2 suggested that it produces a mixture of Lcn2-resistant and sensitive Ent ( Figure 5A ) . To examine this possibility in more detail , culture supernatants from the wild type grown in iron-limited M9 broth were used to stimulate growth of the entB mutant in Lcn2-containing serum . As controls , purified Ent and Gly-Ent ( Salmochelin S4 ) were used to stimulate growth . Gly-Ent supported identical levels of growth in the presence and absence of Lcn2 , whereas Ent only stimulated growth in the absence of Lcn2 . Culture supernatant from the wild type was partially sensitive to Lcn2 , indicating that K . pneumoniae secretes a mixture of Lcn2-sensitive and resistant Ent ( Figure 5C ) . To test whether the iroA locus is required for bacterial colonization , C57BL/6 mice were inoculated with 2×106 cfu of K . pneumoniae producing different combinations of Gly-Ent , Ent and Ybt . To enhance siderophore expression prior to inoculation , K . pneumoniae was grown in the presence of the iron chelator 2 , 2′-dipyridyl ( 200 µM ) . At this concentration of chelator , transcription of Ent-synthesis enzymes is increased ∼8 fold compared to rich media , based on a transcriptional GFP fusion to entC ( data not shown ) . Gly-Ent-producing K . pneumoniae colonized efficiently with or without producing Ybt ( Figure 6 , compare wild type and ybtS ) . In contrast , K . pneumoniae producing only Ent ( iroA ybtS mutant ) was deficient for colonization ( p<0 . 001 at day 3 ) . This defect in colonization could be due to a defect in iron acquisition , or through additional host defenses including pro-inflammatory responses triggered by Ent-Lcn2 signaling . As shown above , Ybt is able to support colonization in the absence of Ent ( Figure 4 ) . To test for defects in colonization independent from iron acquisition , carriage of K . pneumoniae making Ent plus Ybt ( iroA mutant ) was examined . The ability of the iroA mutant to produce Ybt restored maximal colonization . These data indicate that iron sequestration is an important host mechanism for inhibiting colonization of Ent-dependent K . pneumoniae . To determine whether Lcn2 is required to inhibit Ent-dependent bacteria , colonization of Lcn2-producing ( Lcn2+/+ ) and Lcn2-deficient littermate mice ( Lcn2−/− ) was compared ( Figure 7A ) . Wild-type K . pneumoniae producing Gly-Ent and Ybt maximally colonized both Lcn2+/+ and Lcn2−/− mice . Conversely , the iroA ybtS mutant K . pneumoniae dependent on Ent was significantly inhibited in Lcn2+/+ mice ( p<0 . 01 ) . This inhibition was absent in Lcn2−/− mice and was similar to the defect seen in the entB ybtS mutant . Therefore , Lcn2 is required to inhibit Ent-dependent strains and acts predominantly by interfering with iron acquisition . Two potential sources of Lcn2 in the nasal lumen are the mucosa and infiltrating neutrophils . To determine whether mucosal Lcn2 is able to inhibit Ent-dependent K . pneumoniae colonization , mice were depleted of neutrophils using the RB6-8C5 antibody one day prior to inoculation . Despite RB6-8C5 treatment , colonization was significantly inhibited in Lcn2+/+ compared to Lcn2−/− mice at day 1 ( p<0 . 05 , Figure 7B ) . These data indicate that mucosal sources of Lcn2 are sufficient to provide its antibacterial effects during colonization . Although iron sequestration is the predominant effect observed during colonization , Lcn2 may have additional pro-inflammatory effects that are blocked by glycosylation of Ent . In cultured respiratory cells , the combination of Lcn2 and aferric Ent induces synergistic IL-8 release [8] . To determine whether glycosylation of Ent prevents this synergistic activation of chemokine release , A549 human respiratory cells were incubated with combinations of purified Ent , Gly-Ent , or rLcn2 . Whereas co-incubation of Ent with Lcn2 induced IL-8 release greater than 20-fold , co-incubation of Gly-Ent and Lcn2 did not stimulate IL-8 release above the level induced by Lcn2 alone ( Figure 8 ) . No significant cellular cytotoxicity , as measured by LDH release , was detected for any combination of stimuli ( data not shown ) . Consistent with the observation that expression of the iroA locus prevents Lcn2 binding , Lcn2 does not stimulate IL-8 release from cultured respiratory cells in response to Gly-Ent . The induction of the neutrophil chemoattractant IL-8 in vitro suggests that Ent release by colonizing bacteria may stimulate neutrophil recruitment during infection . To measure the effect of Ent on neutrophil influx , flow cytometry was performed on nasal lavage fluid from mice colonized with wild type , iroA and entB mutants . The iroA mutant is predicted to produce Ent that interacts with Lcn2 , is not predicted to be susceptible to iron sequestration since it makes Ybt , and colonizes similarly to the wild type ( Figure 7 ) . In contrast , the entB mutant produces no Ent but also colonizes similarly to the wild type ( Figure 4 ) . Comparison of the wild type and iroA mutant demonstrated no difference in neutrophil influx ( Figure 9A ) . This suggests that the subset of unmodified Ent produced by the wild type is sufficient to induce a robust inflammatory response . To measure cellular inflammation directly attributable to unmodified Ent , neutrophil counts were compared between iroA and entB-colonized mice . The iroA and entB mutants achieved a similar level of colonization ( data not shown ) , but the iroA mutant elicited significantly more neutrophils ( Figure 9B ) . To determine if the neutrophil influx in response to Ent-producing Klebsiella is Lcn2-dependent , Lcn2+/+ and Lcn2−/− littermates were colonized with iroA mutant K . pneumoniae . Lcn2+/+ mice had a significantly greater neutrophil influx than Lcn2−/− littermates in response to the Ent-producing iroA mutant ( Figure 9C , p<0 . 05 Wilcoxon matched pair test ) . The lower number of neutrophils in Lcn2−/− mice could not be attributed to decreased colonization density ( data not shown ) . Concurrent with its predominant iron sequestration effect , Lcn2 also appears to induce greater neutrophil influx in response to bacteria producing unmodified Ent .
This work establishes an animal model to study the function of Lcn2 in the upper respiratory tract . As well as being an important human pathogen , K . pneumoniae is a tractable model organism that colonizes the nasopharynx and can be genetically manipulated to produce siderophore mutants that do or do not interact with Lcn2 . Colonization with wild-type K . pneumoniae persists for at least seven days , and induces a robust acute inflammatory response . The fact that intranasal inoculation of awake mice produces colonization , and not pneumonia , is likely because the mice do not aspirate nasopharyngeal contents into their lungs [28] . Similarly to pneumonia , persistence of K . pneumoniae in the upper respiratory tract requires ongoing iron acquisition [22] . Either Gly-Ent or Ybt are required for maximal colonization . This confirms that the nasal mucosa is an iron-limited environment for bacteria , presumably due to the high lactoferrin concentration in nasal secretions [29] , [30] . Furthermore , this suggests that colonization by K . pneumoniae requires bacterial replication , and not simply persistence of the originally inoculated organisms . Having established the need for K . pneumoniae to acquire iron in the nasopharynx , this model can be exploited to examine how host Lcn2 and the bacterial iroA locus affect the level of colonization . By using Lcn2−/− mice and K . pneumoniae siderophore mutants , this study demonstrates that Lcn2 inhibits nasopharyngeal colonization by bacteria producing unmodified Ent . Although a seemingly narrow target for antimicrobial activity , Ent is produced by many members of the gram-negative Enterobacteriaceae family [1] . Therefore , Lcn2 may contribute to the tropism of enteric commensals for the gut rather than the respiratory tract . The respiratory tract secretes Lcn2 at basal levels and rapidly upregulates Lcn2 expression during colonization [4] , [6] . In contrast , the large intestine does not express basal Lcn2 despite the presence of huge numbers of colonizing bacteria [4] . However , Salmonella infection induces Lcn2 production in the intestine , and the ability to utilize Gly-Ent confers a competitive advantage over an iroN mutant during intestinal inflammation [31] . Therefore , pathogenic Enterobacteriaceae such as Salmonella and K . pneumoniae can use Gly-Ent to obtain iron in an otherwise restricted environment . In the respiratory tract , Lcn2 may provide selective pressure for Ent-independent methods of iron acquisition . Many clinical isolates of K . pneumoniae produce either aerobactin or Ybt [9] , [10] in addition to Ent . Lcn2 binds bacillibactin of Bacillus anthracis , but B . anthracis also produces the unusual siderophore petrobactin that Lcn2 cannot bind [32] . Likewise , pathogens such as S . pneumoniae and H . influenzae appear to acquire iron in the respiratory tract without producing siderophores [33] , [34] . In the pathogenic K . pneumoniae strain used here , either Gly-Ent or Ybt can support colonization of the nasopharynx . However , Ybt cannot support growth in serum . This defect could be due to an inability of Ybt to strip iron off of transferrin , or a serum component other than Lcn2 that inhibits Ybt-mediated iron acquisition . In contrast , the iroA locus supports robust growth in the presence of Lcn2 . Therefore , Ybt and Gly-Ent are not functionally redundant in K . pneumoniae , and likely reflect adaptation to growth in the disparate environments of the respiratory , urinary , and intestinal tracts and the bloodstream . K . pneumoniae nasal colonization induces a robust inflammatory response characterized by a rapid influx of neutrophils . Neutrophils limit colonization of wild-type K . pneumoniae and prevent hematogenous spread to the spleen . Despite producing Lcn2 , neutrophils likely inhibit K . pneumoniae by predominantly Lcn2-independent mechanisms based on the following observations . In the presence of neutrophils , colonization by wild-type K . pneumoniae is the same in Lcn2+/+ and Lcn2−/− mice . In contrast , depletion of neutrophils causes a 10-fold increase in wild-type colonization . Finally , mucosal Lcn2 is able to inhibit colonization of iroA ybtS K . pneumoniae despite depletion of neutrophils . Together , these data indicate that iron sequestration by mucosal Lcn2 is complementary to the antimicrobial actions of neutrophils . Lcn2 also appears to induce a pro-inflammatory response from respiratory cells when bound to Ent . In vitro , Ent combined with Lcn2 causes a synergistic release of IL-8 from human respiratory cells , but Gly-Ent combined with Lcn2 does not . IL-8 is a neutrophil attracting chemokine , suggesting that signaling by Ent-Lcn2 leads to the recruitment of neutrophils . Accordingly , Ent-producing K . pneumoniae induce nasopharyngeal neutrophil influx in an Lcn2-dependent manner . The potential signaling pathway between Lcn2 , Ent and neutrophil recruitment is unknown . There is no direct IL-8 ( CXCL-8 ) homologue in the mouse [35] , [36] , although there are several CXC chemokines such as Mip-2 , KC , and LIX that have been shown to be induced by K . pneumoniae respiratory infections [37] . Studies to determine the effects of Ent and Lcn2 on murine CXC chemokine production from the murine respiratory mucosa are underway . The data from this study suggest the following model in which Lcn2 monitors iron utilization by bacteria and activates the immune response when iron stores become depleted: At low bacterial density , secreted Ent binds Fe and Lcn2 in turn binds Fe-Ent to prevent delivery of iron to bacteria . The Fe-Ent-Lcn2 complex is internalized by respiratory epithelial cells [8] , and could serve as both a signal of controlled colonization and a mechanism of iron recycling . We hypothesize that as bacterial density increases Lcn2 becomes a pro-inflammatory signal . When bacterial growth outpaces Fe availability , an increased proportion of Ent will be aferric . This aferric Ent-Lcn2 complex could be internalized and serve as a signal of uncontrolled bacterial replication to the respiratory epithelium . In vitro , respiratory cells respond to Ent-Lcn2 by induction of chemokines , an effect that could explain the increased neutrophil influx seen in response to Ent-producing K . pneumoniae in vivo . The combined data from neutrophil and bacterial counts are consistent with the hypothesis that Lcn2 has a continuum of iron-sequestering and pro-inflammatory activities . Specifically , Ent-producing K . pneumoniae elicit neutrophils in Lcn2+/+ mice without showing a defect in colonization . The bacterial density of colonization by K . pneumoniae is relatively low ( ∼1e4 CFU/ml ) compared to counts from the pneumonia model [23] . If Fe is not depleted during colonization , then Lcn2 may bind primarily Fe-Ent . The small percentage of aferric Ent-Lcn2 could be sufficient to produce a modest increase in the number of neutrophils elicited by K . pneumoniae colonization . Whereas total depletion of neutrophils leads to increased bacterial numbers ( Figure 3 ) , this incremental increase in neutrophils is not sufficient to affect the density of colonizing organisms . If a perturbance in the microenvironment caused a large increase in bacterial levels , we would predict a greater pro-inflammatory response elicited by Ent-Lcn2 . Alternatively , if K . pneumoniae reaches the lower respiratory tract where it can replicate to high numbers ( >109 CFU/gm ) [22] , there may be a dramatic increase in Ent-Lcn2 formation with a more significant effect of Lcn2 on the immune response . Consistent with this model , Chan and colleagues report that Lcn2 limits growth of K . pneumoniae ATCC 43816 ( the parent strain of our wild type KPPR1 ) during pneumonia [38] . Since Ent is dispensable for growth during pneumonia from KPPR1 [22] , iron sequestration by Lcn2 cannot account for the observed growth inhibition . This suggests that Lcn2 controls bacterial growth by an additional mechanism such as neutrophil recruitment . Whether Ent production stimulates Lcn2-dependent inflammation during pneumonia remains to be determined . Studying the pro-inflammatory effects of Lcn2 on the mucosal surface could reveal a new paradigm of innate immune signaling in response to bacterial metabolism .
KPPR1 , a rifampin-resistant derivative of K . pneumoniae subsp . pneumoniae ( ATCC 43816 ) , with a type 1 O antigen and type 2 capsule was used as the wild-type strain in these studies [23] . Measurement of entC expression was performed using KPPR1 harboring the entC promoter-GFP reporter plasmid pE2 ( VK096 ) [22] . Siderophore mutant strains of KPPR1 contain in-frame deletions of the enterobactin synthesis gene entB encoding 2 , 3-dihydro-2 , 3 dihydroxybenzoate synthase ( VK087 ) , the yersiniabactin synthesis gene ybtS encoding salicylate synthase ( VK088 ) or both ( VK089 ) [22] . Unless otherwise noted , strains were grown overnight in Luria-Bertani ( LB ) media , either at 30°C on agar or at 37°C shaking in broth . Media was supplemented with rifampin ( 30 µg/ml ) or kanamycin ( 50 µg/ml ) as needed . To stimulate siderophore production strains were grown overnight in LB broth supplemented with 200 µM 2 , 2′-dipyridyl ( Acros Organics , Geel , Belgium ) , back diluted to OD600 0 . 1 , and grown an additional two hours to mid-logarithmic phase . PCR primers specific for conserved regions of the iroB glycosylase gene ( iroBfor 5′-GTGATGCAAACCGTCGGCTTC , iroBrev 5′-ACCATCGGTTTGACGGTGCCGAG ) were constructed by comparing DNA sequences from Salmonella typhimurium LT2 , E . coli CFT073 , E . coli UT189 , and K . pneumoniae virulence plasmid pLVPK . An internal 0 . 3 kb iroB PCR fragment was amplified , sequenced , and found to be 96% identical to iroB from pLVPK . This fragment was then cloned into the TA-based PCR cloning vector pCR2 . 1 ( Invitrogen , Carlsbad , CA ) and transferred to a kanamycin-resistant derivative of the λpir-dependent suicide vector pGP704 [39] . This iroB suicide vector was transformed into E . coli strain BW20767 ( ATCC 47084 , RP4-2tet::Mu-1kan::Tn7-integrant uidA ( ΔMlu1 ) ::pir+ recA1 creB510 leu-63 hsdR17 endA1 zbf-5 thi ) and subsequently conjugated into wild-type ( KPPRI ) and ybtS mutant ( VK088 ) K . pneumoniae to generate iroA and iroA ybtS mutants . Integration of the suicide vector into the iroB gene was confirmed by generation of a novel PCR product using one primer on the vector ( pGP704 MCS Pst>Xba 5′-GGTCGACGGATCCCAAG ) and an iroB primer 5′ of the insertion site ( iroBORF for 5′ ATGCGTATTTTATTTATAGGTCC ) . For complementation studies , each mutant was transformed by electroporation with either pACYC184::iroB ( pIJ53 ) or pACYC184::iroBCDN ( pIJ137 , referred to as pIroA in this study ) containing iroA genes from E . coli χ7122 [27] . The vector pACYC184 was transformed as a negative control . All animals were handled in strict accordance with good animal practice as defined by the relevant national and/or local animal welfare bodies , and all animal work was approved by the University of Pennsylvania Institutional Animal Care and Use Committee ( Assurance # A3079-01 ) . For nasal colonization experiments , six to eight week-old C57BL/6 ( Jackson Labs , Jackson , ME ) mice were atraumatically inoculated intranasally without anesthesia with 2×106 cfu of K . pneumoniae . LB broth grown cultures were centrifuged , resuspended in phosphate buffered saline ( PBS ) , and 10 µL of the suspension was applied equally to both nares . To determine colonization density , mice were sacrificed by CO2 asphyxiation , the trachea was exposed and cannulated , 200 µL PBS was instilled , and lavage fluid was collected from the nares . Aliquots of lavage fluid were plated on LB agar supplemented with rifampin with a lower limit of detection of 20 cfu/ml . To measure lung or spleen infection , the organ was removed , weighed , homogenized in 500 µL PBS , plated on LB agar supplemented with rifampin , and quantified as cfu/gm . To examine the infection by histology , skulls were removed by decapitation , fixed by 48 h incubation in 10% neutral buffered formalin ( Sigma-Aldrich , St . Louis , MO ) , and decalcified for 48 h in Cal-EX decalcifying solution ( Fisher Scientific , Fair Lawn , NJ ) . Saggital sections of the nasal cavity were paraffin embedded and processed for hematoxylin and eosin ( H&E ) staining . To deplete neutrophils , 145 µg of the rat anti-mouse IgG2B mAb RB6-BC5 directed against Ly-6G on the surface of mouse granulocytes was injected intraperitoneally 24 h prior to intranasal bacterial inoculation . This dose has been shown previously to result in peripheral blood neutropenia for at least 96 h ( <50 granulocytes/ml , [6] ) . Intraperitoneal injection of 145 µg of rat total IgG was used as a control . To ensure similar endogenous nasal flora , Lcn2−/− mice [3] ( provided by Shizuo Akira via Alan Aderem ) and Lcn2+/+ littermates were compared . Offspring from heterozygous breeding pairs were genotyped using DNA extracted from tail samples with the DNeasy Blood and Tissue Kit ( Qiagen , Germantown , MD ) . For each mouse , paired PCR amplifications were performed with a common primer ( common 5′-CACATCTCATGCTGCTGAGATAGCCAC ) and a primer specific for either the intact Lcn2 locus ( wild-type 5′-GTCCCTCTCACTTTGACAGAAGTCAGG ) or the neomycin-cassette disrupted Lcn2 locus ( neo1500 5′-ATCGCCTTCTATCGCCTTCTTGACGAG ) . Recombinant mouse lipocalin 2 ( rLcn2 ) was purified as previously described [40] . Briefly , E . coli strain BL-21 expressing plasmid-encoded mouse lipocalin 2 as a glutathione S-transferase fusion protein ( a gift from J . Barasch ) was grown to mid-logarithmic phase in Terrific Broth supplemented with 50 µM ferrous sulfate and induced with 0 . 2 mM IPTG . Cells were harvested and lysed by sonication , and rLcn2 was purified on a Glutathione Sepharose 4B bead column ( GE Amersham , Piscataway , NJ ) followed by digestion with human thrombin ( Sigma-Aldrich , St . Louis , MO ) and elution in PBS . Purified rLcn2 was quantified using the Micro BCA Protein Assay Kit ( Pierce , Rockford , IL ) and siderophore-binding activity was confirmed by incubation with Fe-Ent followed by measurement of absorbance at 340 nm . Bacterial growth using mouse serum as an iron source was assayed as previously described [3] . RPMI supplemented with 10% heat-inactivated Lcn2−/− -mouse serum was inoculated with 1×103 cfu/ml of an overnight culture of K . pneumoniae and incubated an additional 24 h at 37°C with 5% CO2 in 96-well plates . Bacterial growth was quantified by plating serial dilutions on LB agar . Where indicated , RPMI was supplemented with 1 . 6 µM recombinant Lcn2 . To determine if the wild type produces Lcn2-resistant and sensitive Ent , sterile-filtered culture supernatants were collected from wild-type KPPR1 grown overnight in iron-limited M9 minimal media . RPMI supplemented with 3% heat-inactivated mouse serum was inoculated with 5×103 cfu/ml of an overnight culture of entB mutant K . pneumoniae , supplemented with 4-fold serial dilutions of purified Ent or Gly-Ent ( Salmochelin S4; EMC microcollections , Tuebingen , Germany ) or culture supernatant , incubated overnight and plated for bacterial cfu . The dilution of culture supernatant ( 1∶640 ) where Lcn2-resistant growth matched that of Gly-Ent was used for comparisons . Antibody staining was performed at 4°C in 96-well plates using cells from 100 µL of nasal lavage fluid resuspended in PBS with 1% bovine serum albumin ( BSA ) . For each mouse , cells were spun at 1500 rpm for 10 minutes , washed once with 200 µL PBS+BSA , centrifuged at 1500 rpm for 2 minutes , resuspended in the same volume and blocked for 10 minutes . Cells were spun as above , resuspended in 25 µL of a 1∶200 dilution of Fc Block ( BD Pharmingen , San Diego , CA ) , and incubated an additional 10 minutes . Then , four-color staining was performed for 30 minutes by addition of 25 µL of an antibody mixture of fluorophore-conjugated Ly6G , CD11b , CD4 , and CD45 antibodies ( BD Pharmingen ) at the following final concentrations: FITC 1∶200 , PE 1∶300 , PerCP 1∶100 , APC 1∶300 . Cells were washed twice , centrifuged and fixed by resuspension in 300 µL PBS+BSA with 0 . 5% paraformaldehyde . For fluorophore compensation , mouse splenocytes were harvested . Spleens were homogenized by passage through a sterile screen , washed in 15 ml RPMI , and red blood cells were lysed by resuspension in 2 ml 0 . 83% ammonium chloride and incubation at room temp for 2 minutes . The suspension was neutralized by addition of RPMI , incubated an additional 3 minutes , and the supernatant was removed , centrifuged , washed in RPMI and resuspended in PBS+BSA . 200 µL aliquots were blocked for 10 minutes with PBS+BSA , and individually stained with FITC , PE , PerCP , or APC-conjugated anti-CD4 antibody . Flow cytometry data was collected using a BD FACSCaliber and analyzed using FlowJo software ( Tree Star , Ashland , OR ) . For nasal lavage fluid , the entire sample was analyzed and normalized to events per ml . Lcn2+/+ and Lcn2−/− littermates were compared pair-wise to account for potential variation in baseline neutrophil counts that may be caused by differences in the endogenous nasal flora . The human type II pneumocyte A549 cell line ( ATCC CCL-185 ) was propagated in minimal essential medium ( Invitrogen , Carlsbad , CA ) supplemented with 10% fetal bovine serum as described previously [8] . Near-confluent monolayers in 24-well plates were weaned from serum and antibiotics overnight . Then , A549 cells were stimulated with combinations of 50 µM purified siderophores ( EMC microcollections , Tuebingen , Germany ) or 25 µM lipocalin 2 overnight as indicated . Purified diglycosylated Ent ( Salmochelin S4 ) , which is the major product of IroB [41] , or purified Ent was used . The next day , culture supernatants were collected and stored at −20°C or used immediately for IL-8 ELISA . The BD OptEIA IL-8 ELISA assay ( BD Biosciences , San Diego , CA ) was performed according to the manufacturers instructions and detected using TMB substrate ( Zymed/Invitrogen , Carlsbad , CA ) and a Bio-Rad Model 680 microplate reader ( Bio-Rad , Hercules , CA ) . As a positive control for IL-8 release , cells were stimulated with 100 pg/ml of IL-1β ( Peprotech , Rocky Hill , NJ ) . A vehicle control was included containing the solvents for each reagent used . Cytotoxicity was evaluated by LDH measurement of supernatants using the Cytotoxicity Detection Kit Plus ( Roche , Mannheim , Germany ) according to the manufacturers directions . All statistical analysis was performed and graphs were generated using Prism 4 for Macintosh ( GraphPad Software , Inc ) . For colonization data , non-parametric analysis of two groups was performed using the Mann-Whitney test and analysis of multiple groups was performed using the Kruskall-Wallis Test with Dunn's post-test . The presence or absence of splenic bacteria was analyzed by Fisher's Exact test . IL-8 ELISA data was analyzed by one-way ANOVA with Tukey's multiple comparison test . Cytotoxicity data was analyzed by one-sample t-test for significant increase above the control and one-way ANOVA with Tukey's multiple comparison test for differences between samples . For neutrophil quantification , non-parametric analysis of mouse pairs was performed using the Wilcoxon matched pair test . Genbank nucleotide accession numbers for the iroA loci used for iroB sequence comparisons are NC_005249 ( K . pneumoniae CG43 pLVPK ) , NC_003197 ( Salmonella Typhimurium LT2 ) , NC_004431 ( E . coli CFT073 ) , and NC_007946 ( E . coli UTI89 ) . Genbank accession number for the iroBCDN genes from χ7122 used for complementation is AF449498 . The Genbank protein accession number for Lcn2 is NP_032517 for 24p3R is NP_067526 . | Bacterial pathogens such as Klebsiella pneumoniae require iron and use secreted molecules called siderophores to strip iron from mammalian proteins . When bacteria colonize the upper respiratory tract , the mucosa secretes the protein lipocalin 2 ( Lcn2 ) that binds to the siderophore enterobactin ( Ent ) and disrupts bacterial iron acquisition . In addition , Lcn2 bound to Ent stimulates release of the neutrophil-recruitment signal IL-8 from cultured respiratory cells . Some pathogens avoid Lcn2 binding by attaching glucose to Ent ( to make Gly-Ent ) or by making alternative siderophores . To determine the effect of Lcn2 on bacterial colonization , we colonized mice that express or lack Lcn2 with K . pneumoniae mutants that express or lack Ent , Gly-Ent and the alternative siderophore Yersiniabactin ( Ybt ) . Our results indicate that mucosal Lcn2 inhibits colonization through iron sequestration and increases the influx of neutrophils in response to K . pneumoniae producing Ent . Therefore , Lcn2 acts as a barrier to colonization that pathogens must overcome to persist in the upper respiratory tract . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"microbiology/innate",
"immunity",
"microbiology/immunity",
"to",
"infections",
"microbiology"
] | 2009 | Mucosal Lipocalin 2 Has Pro-Inflammatory and Iron-Sequestering Effects in Response to Bacterial Enterobactin |
The pilus 2a backbone protein ( BP-2a ) is one of the most structurally and functionally characterized components of a potential vaccine formulation against Group B Streptococcus . It is characterized by six main immunologically distinct allelic variants , each inducing variant-specific protection . To investigate the molecular determinants driving the variant immunogenic specificity of BP-2a , in terms of single residue contributions , we generated six monoclonal antibodies against a specific protein variant based on their capability to recognize the polymerized pili structure on the bacterial surface . Three mAbs were also able to induce complement-dependent opsonophagocytosis killing of live GBS and target the same linear epitope present in the structurally defined and immunodominant domain D3 of the protein . Molecular docking between the modelled scFv antibody sequences and the BP-2a crystal structure revealed the potential role at the binding interface of some non-conserved antigen residues . Mutagenesis analysis confirmed the necessity of a perfect balance between charges , size and polarity at the binding interface to obtain specific binding of mAbs to the protein antigen for a neutralizing response .
The bacterial surface is the foremost interface between host and pathogen , and recognition of the specific epitopes by the immune system provides the host a key signature to initiate microbial clearance . Identification and characterization of antigenic epitopes is a rapidly expanding field of research with potential contributions to the tailored design of improved , safe and effective vaccines [1] , [2] . A number of approaches are currently being used that require atomic-level information in understanding the rules governing antibody/antigen interaction , in particular it is the degree of complementarity between surfaces on epitope and paratope that determines the affinity and specificity of this interaction [3]–[5] . To date , the concept of complementariness is directly related to the conservation of the amino acid sequence on a specific neutralizing epitope . A single amino acid change resulted crucial to alter the surface antigenic properties of a specific epitope of Neuraminidase ( NA ) in Influenza virus [6] . Streptococcus agalactiae ( also known as Group B Streptococcus or GBS ) is a Gram-positive pathogen causing severe diseases in newborn and young infants worldwide [7] . Pilin proteins , structural components of cell surface-exposed appendages , have been discovered in GBS as important virulence factors as well as promising vaccine candidates [8] . These high molecular weight structures are made by a major shaft subunit ( named backbone protein , BP ) , a major ancillary protein ( named AP1 ) , and a minor ancillary protein ( named AP2 ) . BP is distributed regularly along the pilus structure and is fundamental for pilus assembly whereas the two ancillary proteins are dispensable [9] . AP1 is thought to be located at the tip of the assembled pilus structure , while AP2 is involved in pilus attachment to the cell wall [10]–[12] . In GBS three pathogenicity islands , named Pilus Island-1 ( PI-1 ) , Pilus Island-2a ( PI-2a ) and Pilus Island-2b ( PI-2b ) , each encoding pilin protective subunits , have been identified [9] . Among them , the backbone protein of Pilus Island 2a ( BP-2a ) is a key component of a promising pilus-based vaccine formulation against Group B Streptococcus infections [13] , [14] . However , this protein showed the highest level of gene variability among all pilin antigens , characterized by six non cross-protective allelic variants [13] . Each variant identified was able to induce protective immunity in mouse models and opsonophagocytosis killing of live bacteria , but only against GBS strains expressing the homologous variant [13] , [14] . A recent Structural Vaccinology approach applied to the BP-2a protein led to the identification of the minimal protein domain carrying the protective epitopes [14] . By using in vitro opsonophagocytosis assays and in vivo animal infectious models , we demonstrated that , within each variant , the domain D3 is responsible for eliciting neutralizing antibodies against pathogen homologous infections [14] . This structure-based approach combined with immunological assays succeeded in the generation of an easy-to-produce chimeric antigen capable to elicit protection against the majority of circulating GBS serotypes [14] . However , the specific mechanisms by which antibodies raised against each variant can mediate a neutralizing response only against GBS strains expressing the homologous variant are not completely understood . The aim of this work is to contribute a deeper understanding of the molecular basis driving the immunogenic specificity of single BP-2a variants , explaining the mechanism by which amino acid variability on the antigen surface may allow the bacteria to adapt to the host environment and/or escape its immune system . To investigate the variant-specific immunogenicity of BP-2a at the molecular level , we generated neutralizing monoclonal antibodies ( mAbs ) raised against a specific allelic variant of the protein , the 515 allele [13] , [14] . The produced mAbs were functionally screened according to their ability to recognize the polymerized pilus structure on the bacterial surface and to mediate opsonophagocytosis GBS killing in vitro . By Surface Plasmon Resonance ( SPR ) technology we determined their binding affinity . An approach based on partial digestion , immunocapture and mass spectrometry was used to identify the epitope region in the antigen . Finally , docking and molecular simulation prediction studies were used to elucidate the intrinsic and functional affinity between mAbs and antigen at a residue level . This work describes a potential strategy to investigate the immunological and structural properties of a surface virulence factor , by elucidating at the molecular level the chemical-physical properties directly related to an effective neutralizing response against pathogen infection .
Recent advances in monoclonal antibodies ( mAbs ) technology suggested us to use them as tool to investigate the principles governing functional antibody/antigen interactions [6] . So , to understand the immunological differences among the six different variants of the highly immunogenic GBS protein BP-2a through the identification of neutralizing epitope ( s ) , mouse monoclonal antibodies ( mAbs ) against the 515 allelic variant were generated following standard procedures ( see Materials and Methods ) . Since surface accessibility of bacterial proteins is a fundamental pre-requisite of antibodies for mediating an effective humoral response against bacterial infections , selection criteria for mAb identification were based on variant-specificity and on bacterial surface staining , which was investigated by Flow Cytometry ( FACS ) analysis . The screening procedure resulted in the identification of six different monoclonal antibodies ( named 4H11/B7 , 17C4/A3 , 27F2H2/H9 , 14F6/A1 , 25B7/D7 , 28E7/E4 ) able to recognize only the polymeric pilus structure on the bacterial surface of their homologous strain 515 ( Figure 1A ) . In fact , they were not able to stain the surface of GBS strains expressing a different BP-2a variant ( Figure 1A ) . To assess if the six selected mAbs could also mediate a functional immunogenic response against GBS we performed an in vitro opsonophagocytosis assay , using as effector cells differentiated HL60 cells , as described in the Materials and Methods section , and GBS strain 515 . We analyzed each monoclonal antibody at three different dilutions in presence of baby rabbit complement . As shown in Figure 1B , only three out of six mAbs ( 4H11/B7 , 17C4/A3 and 27F2H2/H9 ) were able to mediate an effective complement-dependent opsonization and killing of GBS bacteria , meaning that these mAbs can recognize and bind neutralizing epitopes exposed in the pilin protein on the bacterial surface . Classes and subclasses of the six monoclonal antibodies were determined as described in the Materials and Methods section . Clone 4H11/B7 secreted the IgG2b subclass; clone 17C4/A3 and 27F2H2/H9 secreted the IgG2a subclass whereas clones 14F6/A1 , 25B7/D7 and 28E7/E4 had the IgG1 isotype . To evaluate the interaction between BP-2a 515 variant and the selected mAbs , we conducted SPR ( Surface Plasmon Resonance ) analyses . A convenient strategy to study this interaction is to capture the antibody on a surface containing Fc-receptors in order to place the antibody in a well-defined orientation for binding analysis . Two CM5 biosensors , one coated with Protein A and the second with Protein G , were prepared in order to steadily capture the different isotypes of the mAbs and study their interaction with BP-2a 515 variant in terms of association ( ka ) and dissociation ( kd ) rate constants , and binding affinity ( KD = kd/ka ) . The monoclonal antibodies 17C4/A3 and 27F2/H2/H9 were captured on both Protein A and Protein G biosensors while 4H11/B7 mAb was stably captured only in presence of Protein G . Two out of the three IgG1 mAbs , 25B7/D7 and 28E7/E4 , were captured by Protein G , increasing the RU ( refractive unit ) of capture according to the concentration ( 5 or 15 nM ) , while 14F6/A1 was not captured by protein G up to the concentration of 15 nM . After the capture , the mAbs ( 4H11/B7 , 17C4/A3 and 27F2H2/H9 ) that were the same antibodies that were able to mediate opsonophagocytic killing of GBS cells could bind BP-2a 515 variant , while the two IgG1 mAbs captured by the Protein G biosensor did not bind to the protein in the range of 0 . 5 to 2 . 5 µM . For the mAbs which were able to bind the BP-2a 515 variant single cycle kinetics were performed on both biosensors when possible . The average of three independent runs is reported in Table 1 . Data showed that the association phase ( ka ) resulted comparable for all the tested mAbs , within a range of ∼2 . 5-fold ( ka max/ka min ) . Larger differences were measured in the kinetic of dissociation kinetics ( kd ) , with the most stable binding observed for 4H11/B7 , kd approximately 10-fold slower than for 17C4/A3 and 5 . 5-fold slower than for 27F2/H2/H9 mAb . Nevertheless , the corresponding thermodynamic dissociation constants ( KD ) did not differ among the three mAbs , which showed to strongly bind the BP-2a 515 variant . To identify the neutralizing epitope on BP-2a 515 variant , an epitope mapping with the three functionally active monoclonal antibodies ( 27F2/H2/H9 , 17C4/A3 and 4H11/B7 ) was performed . Two different MS-based approaches were used , one of them allowing the identification of conformational epitopes ( see Materials and Methods ) . With either approach , the experiments were performed six times , using the proteases trypsin , LysC and GluC , and using the mutant form of the entire protein lacking the three isopeptide bonds ( BP-2a-515K199A/K355A/K463A ) previously generated [14] . It is well-known that the presence of internal isopeptide bonds is important for the stability and resistance to proteolysis of single structurally independent domains in which the protein is organized . The results indicate that the three monoclonal antibodies recognize the same region of BP-2a-515 in domain D3 , with sequence 411-TYRVIERVSGYAPEYVSFVNGVVTIK-436 . The domain D3 was the same protein portion previously characterized as the domain carrying most of the epitopes inducing protective antibody responses [14] . Figure 2A shows the mass spectrum of the total LysC digestion of BP-2a-515 ( upper panel ) and that of the peptides immunocaptured with 4H11/B7 ( lower panel ) . The two labeled signals in the lower panel correspond to the fragments of BP-2a-515 sharing the 411–436 sequence ( Figure 2B ) . To confirm the sequence of these peptides , MS/MS spectra were obtained for the peak with a m/z = 2946 . 530 Da . For the peak with a m/z = 4156 . 461Da no MS/MS was recorded due to the low intensity of the signal . When using trypsin and GluC no immunocaptured peptide fraction could be detected . Since the peptides retained by the antibodies after LysC digestion contained R and E residues , potential cleavage sites for trypsin and GluC , respectively , these results suggest that the residues R and E and their immediate neighbors may play a role either in the interaction with mAbs or in the structural arrangement of the epitope , since cleavage at either site prevents binding . In the BP-2a-515 structure ( PDB code: 2XTL ) [14] it can be observed that the residues that are exposed at the surface are comprised between Glu424 and Lys 436 ( Figure 2C ) . The fact that the two approaches tested ( see Materials and Methods ) lead to the same result indicates that epitope recognition is primarily based on sequence . To investigate the mode of action of neutralizing mAbs on the antigen BP-2a 515 , we performed mAbs-protein docking . Monoclonal antibodies sequences were obtained by isolation of total RNA from each hybridoma cell line and reverse-transcription . Then , using the generated cDNA as template , the heavy ( VH ) and light ( VL ) chains were amplified using specific PCR primers . Sequence comparison of the three neutralizing mAbs is showed in Figure 3 and all light chains were of the κ-type . To elucidate the residue-specific interaction between antigen and antibody at the binding interface , after mAbs sequencing , a structural model of the Fv domain of two of them ( 17C4/A3 and 4H11/B7 ) was developed using Modeler 9v8 [15] . Template crystal structures for mAb 17C4/A3 were selected from PDB showing 80% sequence identity for an antibody variable heavy chain ( PDB entry 1H3P ) and 79% sequence identity for an antibody variable light chain ( PDB entry 2ROW ) . The same procedure led to the selection of two template crystal structures for mAb 4H11/B7 sharing 90% sequence identity for the light chain ( PDB entry 1I9J ) and 76% sequence identity for the heavy chain ( PDB entry 3O6M ) . In both cases , light and heavy chains were packed together and energy minimized before proceeding with docking studies . To investigate at the amino-acid level the molecular interactions between neutralizing mAbs and BP-2a antigen , the modeled structures of antibodies were docked against the partial crystal structure of the antigen [14] using ATTRACT [16] . Knowing that mAbs bind to the D3 domain of the antigen in the region T411-K436 , we used this information to screen the most accurate complexes within a range of 15000–20000 complexes generated by the docking program . To validate the stability and reliability of the best selected complexes , we performed explicit solvent molecular dynamics simulation using GROMACS 4 . 0 . 5 simulation package [17] . Molecular dynamics results of best docked solutions of mAbs/BP-2a antigen revealed different binding orientations of the neutralizing mAbs against the target protein which showed the importance of specific residues both in the epitope and in the paratope . Molecular simulation results indicated that a shorter portion of the previous identified epitope might be necessary at binding interface: P423-K436 in 17C4/A3-complex ( Figure 4A ) and V426-K436 in 4H11/B7-complex ( Figure 4B ) . In particular , during the course of the simulation , the distance between those residues and the CDR remained at a contact distance of around 4 Angstrom , indicating their importance for complex interaction and stability . Although two different mAb binding orientations were identified as sterically possible , two amino acid residues located on the target antigen were identified as fundamental at the binding interface in both cases: Val429 and Asn430 . Remarkably , residue 429 had been identified as being under selective pressure , which is consistent with a direct role in the interaction with the antibody . Both mAb binding interfaces form a deep cleft filled by a loop region of domain D3 . In the docking models residues Val429 and Asn430 fill the deeper cavity of both clefts in a water-free environment ( Figure 5 ) . Val429 is involved in hydrophobic interaction with Val206 , Arg208 and Ser204 in the 17C4A3-antigen complex ( Figure 6A ) , whereas it interacts with Val207 , Leu317 , Asn320 and Tyr321 in the 4H11/B7-antigen complex ( Figure 6B ) . Asn430 is predicted to interact with residue Glu269 and Arg208 through H-bonds and makes polar interactions with Ala252 , Ser254 and Ile318 in the 17C4A3-complex ( Figure 6A ) , while it establishes through polar interactions with Tyr321 and Tyr277 residue in the 4H11/B7-complex ( Figure 6B ) . The level of conservation of the epitope was analyzed in a set of 144 BP-2a sequences from different GBS isolates . An alignment of the protein sequences revealed six main variants , as described in previous work [13] . In the full protein alignment , inter-variant variability is high ( p distance = 0 . 31 ) with relatively few conserved positions ( 25% ) , while intra-variant variability is low , suggesting a mosaic structure ( see full alignment in Supplementary Material ) . Comparative sequence analysis of the identified epitope region among the BP-2a protein sequences divided the isolates into the same variants ( Figure 7A ) with the same variability pattern of the full protein ( p distance = 0 . 37 and 26% conserved sites ) . The alignment also showed that the epitope region has a more conserved first half , residues 411 to 423 ( sequence 515 ) , and a more divergent second half , residues 424 to 436 ( Figure 7B ) . In addition , the unique BP-2a nucleotide sequences were aligned to study the genetic events causing the observed variability . A recombination analysis using GARD [18] identified two statistically significant break points located at codon positions 198 and 636 ( sequence 515 ) . To distinguish the effect of mosaicism and point mutations in epitope variability , results from the GARD algorithm were taken into account for the estimation of positive selection . Thus , the REL algorithm implemented in HyPhy [19] identified 16 sites under selective pressure in BP-2a ( Supplementary Material ) , one of them ( Val429 , sequence 515 ) located in the epitope region ( Figure 7A ) . These results suggest that both recombination and selection of advantageous mutations have acted to generate the six BP-2a variants for both the full protein and the epitope region . In particular , residue 429 could be changing in response to the pressure of the immune system and be thus a key element for immunological specificity . To further confirm single amino acid contributions to mAb-epitope binding and elucidate molecular determinants of the immunological specificity of each BP-2a allelic variant , a peptide dot blot immunoassay was performed . Combining epitope mapping , docking and molecular dynamics results with alignment data and detected signatures of positive selection ( Figure 7A ) , four mutated peptides were synthesized ( Figure 7B ) and used to evaluate the contribution of single point mutations at the mAb-antigen binding interface . In particular , major consideration was reserved on those residues found to reach the deeper cavity of the mAbs cleft: Val429 and Asn430 . The designed mutations were aimed at testing the effect of size and charge of the side chains of these two residues on the binding to the antibodies . Mutating both the Asn430 and Val429 to an alanine residue resulted in a conserved capability of binding to the two mAbs , even with increased affinity in the case of Val429 substitution . On the other hand , mutating either residue to lysine , that carries a bulkier and positively charged side chain , substantially reduced or completely abolished the binding of the mAbs to the spotted peptides ( Figure 7B ) . The results from lysine substitution support the presence of residues 429 and 430 at the interaction interface ( highly perturbed by a lysine ) while those from alanine substitution suggest that these two residues are not responsible for specificity but their nature is probably limited by their fit in the cavity .
The ability of the host to identify microbial molecular determinants that are unique to pathogens has a crucial role in host defense . The recognition by the immune system of the host of surface exposed components , such as proteins and polysaccharides represents the start signal for microbial clearance . Characterization studies of vaccine formulations require a deep knowledge of the interactions between pathogen and host immune system and vaccine components should include the molecular determinants able to stimulate an effective immune response against a specific pathogen . The data reported in this work show a significant correlation between the molecular interactions of monoclonal antibody/target protein with successful neutralizing response against bacterial infection . The protection against GBS has been associated with the production of high levels of neutralizing antibodies which specifically recognize the antigens exposed on bacterial surface [20] , [21] . However , the specific mechanisms , in terms of single molecular determinants , by which antibodies neutralize GBS infections , are not completely understood . Moreover , it is well-known that GBS as well as many other bacteria have evolved a wide range of mechanisms to escape the immune system of their hosts or to adapt to environmental variation , for instance , adopting the strategy of gene variability and/or differential gene expression . These strategies play a crucial role in the capacity of pathogens to trigger disease and also explain why it is so difficult to develop vaccines against these microorganisms . Thus , positive selection and recombination have played an important role in adaptation of the core-genome of different Streptococcus species to different hosts [22] . In this context , the identification of different allelic variants of a key vaccine candidate , such as the pilin protein BP-2a , able to induce variant-specific protection [13] clearly reflects a typical strategy of the bacterium to escape the immune system of the host and , at the same time , represent an additional confirmation of the important role of this protein in GBS virulence . Recent data showed that the majority of protective epitopes of the different BP-2a alleles are located in a single structurally independent domain , called D3 domain [14] . A synthetic chimeric protein constituted by the protective D3 domain of the six BP-2a variants was able to protect mice against the challenge with all of the type 2a pilus-carrying strains [14] . In this work we have investigated the contribution of single amino acid residues within the immunodominant domain D3 of BP-2a , able to drive the neutralizing host humoral response . As a tool to investigate the principles governing functional antibody/antigen interactions at the amino acid level , we successfully selected functionally active monoclonal antibodies targeting the 515 allelic variant of the pilin BP-2a . The selected mAbs were able both to recognize the polymerized pilus structure on bacterial surface and to mediate complement-dependent opsonophagocitic killing of live bacteria . Epitope mapping analysis of two of the neutralizing mAbs identified showed that the mAbs bind to the same region of BP-2a-515 in domain D3 , the same domain previously identified as the immunodominat protein region carrying protective epitopes [14] . Although these results confirmed the importance of D3 domain for immunogenicity and protection capacity of BP-2a , to elucidate the specific affinity of the antibodies versus their protein target a structural analysis of the mAbs alone and in complex with the target antigen has been performed . Molecular docking and MD simulation studies indicated that only two specific residues on the target protein , Val429 and Asn430 , were able to reach the deepest cavity formed by the antibody binding interface , mediating specific hydrophobic and polar/H-bond interactions , respectively . Previous studies support the importance of single amino-acid residues at the binding interface in mAb/antigen interactions , responding to a strict balance of shape and energetics . In the case of Influenza virus ( H3N2 vaccine strains 1968–2007 ) , modeling and antigen/antibody docking analyses revealed the molecular basis of the interactions between Hemagglutinin ( HA ) protein , the primary target of the human immune system , and monoclonal antibodies [6] . Specific mutations both in the neutralizing epitopes and in their vicinity altered the protein surface and the surface electrostatics of the virus , leading to the loss of recognition by the antibody [6] . Though the epitopes responsible for immunity were very similar in successive variants of HA , the simulations could explain the antigenic drift of pathogen surface determinants that has been responsible for the loss of immunity against Influenza infection even in vaccinated population [6] . It has been also shown that single-residue mutants of an antigen may prevent docking by increasing the free energy barrier to conformational rearrangements required for binding to the antibody [23] . In light of this data , the variability of the epitope region identified in BP-2a has been analysed , including the detection of events of recombination and positive selection . Both factors have been found to be significant players in epitope variability . Thus , recombination is likely to be at the basis of the six allelic variants known today and , in addition , residue 429 is predicted to be under positive selection . Interestingly , this residue is located in a part of the epitope region that has low conservation , strengthening the hypothesis of a GBS antigenic drift to escape the immune response and to adapt to the host . To further characterize the epitope recognized by 4H11/B7 and 17C4/A3 mAbs , a functional dot blot assay using mutated peptides was performed . To identify functional residues within the neutralizing epitope , we mutated those residues predicted to be fundamental at the binding interface ( Val429 and Asn430 ) to lysine and alanine . Mutating Val429 and Asn430 into alanine did not drastically affect mAb surface electrostatics and did not generate steric interferences that could inhibit the binding of antibody . This indicates that a perfect surface antigen-antibody complementarity in this region is not necessary for binding . Conversely , changing the same residues in lysine resulted in a decreased , in the case of Val429 , and in a complete abolishment , in the case of Ans430 , of mAb binding . The substitution of valine or asparagine by lysine introduces a drastic perturbation of the shape and electrostatics of the antigen surface at this region . The fact that this perturbation inhibits binding supports the presence of these two residues at the interface . Overall , this study provides new insights into mAbs-BP-2a 515 variant interactions and highlights the molecular correlation between BP-2a variability and its immunological specificity . Moreover , the identification of a neutralizing epitope of a highly immunogenic antigen could be useful for a knowledge-based design of effective vaccines , avoiding the side effects of unfavorable epitope ( s ) and stringently targeting the immune response only on those one ( s ) , belonging either to the same or to different antigenic protein ( s ) , responsible of pathogenic clearance . Knowing the native molecular architecture of protective determinants could be possible to selectively engineer the antigens for including them in a more effective vaccine formulation .
Animal treatments were performed in compliance with the Italian laws and approved by the institutional review board ( Animal Ethical Committee ) of Novartis Vaccines and Diagnostics , Siena , Italy . GBS strains used in this work are 515 ( serotype Ia , expressing BP-2a-515 allele ) ; CJB111 ( serotype V , expressing BP-2a-CJB111 ) ; H36B ( serotype II , expressing BP-2a-H36B ) ; 3050 ( type II , expressing BP-2a-2603 ) ; CDC84 ( serotype II , expressing BP-2a-DK21 ) ; and strain CDC89 ( serotype Ia , expressing BP-2a-CJB110 ) . Bacteria were grown at 37°C in Todd Hewitt Broth ( Difco Laboratories ) or in trypticase soy agar supplemented with 5% sheep blood . The full-length BP-2a 515 variant and the mutated form of BP-2a-515 ( BP-2a-515K199A/K355A/K463A ) were produced as previously reported [14] . Recombinant proteins were expressed in E . coli BL21 ( DE3 ) ( Novagen ) cells as His-tagged fusion proteins and purified by affinity chromatography and gel filtration . Mouse monoclonal antibodies ( mAbs ) were generated by Areta International ( Varese , Italy ) using standard protocols . Briefly , B-cell hybridoma clones were isolated from spleen cells of immunized CD1 mice with the purified recombinant BP-2a-515 protein . Positive clones were first selected by ELISA and then culture supernatants were screened for binding to the surface of GBS 515 strain by flow cytometry . Positive primary hybridoma clones were subjected to single cell cloning and sub-cloning by limiting dilution . Monoclonality of a clone was accepted only when all the wells of a microtitre plate with growing cells gave positive reaction in indirect ELISA after repeated sub-cloning . The selected mAbs were finally purified by protein G affinity chromatography . Classes and subclasses of the monoclonal antibodies were determined by IsoQuick Mouse Monoclonal Isotyping Kit ( Sigma ) . Flow Cytometry Analysis ( FACS ) analysis was performed as described elsewhere [9] . Briefly , mid-exponential phase bacterial cells were fixed in 0 . 08% ( wt/vol ) paraformaldehyde and incubated for 1 hour at 37°C . Fixed bacteria were then washed once with PBS , resuspended in Newborn Calf Serum ( Sigma ) and incubated for 20 min . at 25°C . The cells were then incubated for 1 hour at 4°C in presence of mAbs diluted 1∶200 in dilution buffer ( PBS , 20% Newborn Calf Serum , 0 . 1% BSA ) . Cells were washed in PBS-0 . 1% BSA and incubated for a further 1 h at 4°C with a 1∶100 dilution of R-Phycoerythrin conjugated F ( ab ) 2 goat anti-mouse IgG ( Jackson ImmunoResearch Laboratories; Inc . ) . After washing , cells were resuspended in PBS and analyzed with a FACS CANTO II apparatus ( Becton Dickinson , Franklin Lakes , NJ ) using FlowJo Software ( Tree Star , Ashland , OR ) . The assay was performed using differentiated HL-60 as phagocytic cells and live bacteria as target cells . GBS strain 515 was grown to mid-exponential growth phase ( A650 nm = 0 . 3 ) , harvested by centrifugation , and , after washing in cold saline solution , was resuspended in HBSS buffer ( Invitrogen ) . Promyelocytic HL-60 cells ( ATCC , CCL-240 ) were expanded in RPMI 1640 ( Invitrogen ) containing 10% Fetal clone I ( HyClone ) at 37°C with 5% CO2 and differentiated into granulocyte-like cells to a density of 4×105 cells/ml by the addition of 100 mM N , N dimethylformamide ( DMF , Sigma ) to the growth medium . After 4 days , cells were harvested by centrifugation and resuspended in HBSS buffer . Briefly , the reactions took place in a total volume of 125 µl containing ≈3×106 differentiated HL-60 , ≈1 , 5×105 CFU of GBS cells , 10% baby rabbit complement ( Cedarlane ) , and different dilutions of purified mAbs . Immediately before and after 1 h of incubation at 37°C with shaking at 350 rpm , a 25-µl aliquot was diluted in sterile distilled water and plated onto trypticase soy agar plates with 5% sheep blood . A set of negative controls consisted of reactions without phagocytic cells or with heat-inactivated complement . The amount of opsonophagocytic killing ( log kill ) was determined by subtracting the log of the number of colonies surviving the 1 h assay from the log of the number of CFU at the zero time point . Surface plasmon resonance ( SPR ) analyses were performed using a Biacore X100 instrument ( GE Healthcare ) . Protein A and Protein G ( Sigma ) were immobilized on CM5 biosensors ( Biacore ) using standard primary amine coupling ( Amine Coupling Kit , GE Healthcare ) in which the carboxymethylated CM5 dextran layers were activated by mixing equal volumes of 0 . 4 M N-ethyl-N′- ( 3-dimethylaminopropyl ) carbodiimide ( EDC ) and 0 . 1 M N-hydroxysuccinimide ( NHS ) at a flow rate of 10 µL/min for 7 min injection . Protein A ( 250 µg/mL ) and Protein G ( 150 µg/mL ) in 10 mM sodium acetate pH 4 . 5 , were immobilized on the activated biosensors using a contact time of 9 min at 10 µL/min flow rate . Unreacted NHS-esters were blocked with three injections ( 4 min each ) of 1 . 0 M ethanolamine hydrochloride , pH 8 . 5 . The immobilization procedure allowed obtaining a Protein A and a Protein G coated biosensors of ∼3500 RU and ∼1000 RU respectively . Untreated flow cell 1 was used as reference . PBS buffer pH 7 . 2 with 0 . 005% ( v/v ) Tween 20 was used as running buffer for protein immobilization and binding experiments . To perform single cycle kinetics ( SCK ) , monoclonal antibodies ( ranging from 5 to 15 nM in running buffer ) were captured onto the Protein A and Protein G surfaces , according to their isotype , at a flow rate of 2 µL/min for 4 min injection . The analyte BP-2a 515 variant in 2-fold serial dilutions in running buffer ( starting from 500 or 250 nM , five concentrations in total ) was injected over the captured antibody for 2 min at 45 µL/min followed by a 5 or 10 min dissociation . Biosensor regeneration was performed after each cycle and achieved using urea 8 M , pH 10 . 5 ( 4 min , 10 µL/min ) . This treatment did not damage the biosensor surface as shown by equivalent signals of capturing ligand on different runs . Each kinetic experiment was preceded by an identical binding-regeneration cycle of buffer as analyte after mAb capture . This cycle was used as blank and subtracted from all the active curves to correct background effects . The association , dissociation and affinity constants ( ka e kd and KD respectively ) were determined by a simultaneous fitting of the kinetic curves with a model of equimolar stoichiometry ( 1∶1 ) using the BIAevaluation X100 software version 1 . 0 ( GE Healthcare ) . The epitope-mapping approach was based on the method described by Peter and Tomer [24] , which we adapted to the following two protocols [25] , [26]: Approximately 5×106 monoclonal antibody-secreting hybridoma cells were collected . Poly ( A ) + RNA was isolated using RNeasy Mini Kit according to the manufacturer's instructions ( QIAGEN ) . cDNA was produced via reverse transcription using ∼2 ug of poly ( A ) + RNA template and oligo- ( dT ) 12–18 primer using First Strand cDNA Synthesis kit ( Novagen ) . The resulting cDNA was used as a template for PCR amplification using PfuUltra High-Fidelity DNA Polymerase ( Stratagene ) and degenerated primers specific for FvH and FvL gene fragments ( Mouse Ig-Primer Set , Novagen ) . PCR timing has been set according to the manufacturer's instruction ( Mouse Ig-Primer Set , Novagen ) . Positive PCR products have been purified using a QIAprep Spin Miniprep Kit ( QIAGEN ) and sequenced . A total of 144 S . agalactiae BP-2a sequences were examined both from GenBank , including previously published sequences [13] , and from complete genome sequences . Codon alignments and phylogenies were constructed using the coding region for each of the 22 unique BP-2a gene sequences with MEGA5 [27] . To detect codons that show signs of adaptive evolution the program HyPhy [19] was used , as implemented in Datamonkey [28] . The codon-based maximum likelihood method REL ( Random Effects Likelihood ) [29] was used to estimate the dN/dS ratio at every codon in the alignment . The REL method can also take recombination into account , provided that prior to the selection analysis a screening of the sequences for recombination breakpoints is performed . The recombination analysis was performed with GARD [18] , using the HKY85 substitution model . The corresponding amino acid sequences of monoclonal antibodies were used to search the Protein Data Bank ( PDB ) in order to retrieve suitable templates for modeling . Ten models for each antibody have been obtained with Moleder 9v8 [30] . The best models were selected according to the objective function scoring . The quality of the refined structures obtained was checked with verify Profile-3D module of Discovery Studio 3 . 0 ( Accelrys ) . Molecular docking has been performed with ATTRACT docking program [31] . The docking protocol of ATTRACT has already been described in previous publications [16] , [31] . Briefly , the antibodies and the target protein ( BP-2a 515 variant ) coordinates are translated into a reduced protein presentation made up to three pseudo atoms per amino acid residue: the protein backbone is represented by one pseudo atom , small aminoacid side chains ( Ala , Asp , Asn , Cys , Ile , Leu , Pro , Ser , Thr , Val ) are represented by one pseudo atom and larger and more flexible side chains are represented by two pseudo atoms , to better describe the shape and dual chemical character of side chains . The contacts between pseudo atoms are described by different interaction: Lennard–Jones ( LJ ) -type potentials ( A/r8-B/r6-potential ) , repulsive and attractive LJ-parameters describing approximately the size and physico-chemical character of the side chain chemical groups [31] . For systematic docking studies , the antibody , called the ligand protein , was used as probe and placed at various positions and various orientations on the surface of the domain D3 of BP-2a 515 crystal structure . We also took into account the experimental data from Mass analysis , setting a weight of 1 . 5 on exposed aminoacid residues identified and on the surface area ( 4A ) around them . Best docked complexes were selected according to energy scoring function and were finally energy-minimized using the Sander program from the Amber8 package . 18 . During energy minimization , a Generalized Born ( GB ) model was employed to implicitly account for solvation effects as implemented in Amber8 . All molecular dynamics ( MD ) simulations were performed with the GROMACS 4 . 0 . 5 simulation package [17] using the AMBER99SB-ILDN force field [32] with explicit water ( TIP3P ) [33] . The selected energy minimized complexes served as starting structure for MD simulations . After stepwise heating of the systems to 310 K production runs were performed for up to 20 ns with a time step of 2 fs in the NPT ensemble at 310 K and 1 bar . Temperature and pressure were controlled by Nosé-Hoover [34] , [35] ( coupling constant tt = 2 . 5 ) and Parrinello-Rahman [36] , [37] ( tp = 5 . 0 ps ) schemes , respectively . Figures of the molecular structures were generated with VMD [38] and Discovery Studio 3 . 0 ( Accelrys ) . Amounts of 5 – 2 – 0 . 2 µg of purified peptides ( Thermo scientific ) were spotted on nitrocellulose membrane ( 0 . 45 µm pore size , Biorad ) and left to dry for at least 30 minutes at room temperature . The spotted membranes were washed three times with PBST ( 0 . 05% Tween 20 in phosphate-buffered saline or PBS pH 7 . 4 ) applying a constant vacuum flow using SNAP i . d . Protein Detection System ( Millipore ) and blocked for 1 h at room temperature in PBST buffer containing 10% of non-fat-dry milk ( Biorad ) . The membranes were then probed 1 h at room temperature with specific anti-BP2a mAb ( diluted ∼4 . 5 µg/mL in PBST/1% non-fat-dry milk ) and washed 5 minutes ( 3X ) with PBST and further incubated in PBST/1% non-fat-dry milk for 1 h containing a dilution of 1∶1000 goat anti-mouse horseradish peroxidase-conjugated secondary antibody ( Dako , Glostrup , Denmark ) . Subsequently , the filters were washed 15 minutes ( 2X ) with PBST and developed by enhanced chemiluminescence ( ECL ) detection assay ( Pierce ECL Western blotting substrate , Thermo Fisher Scientific Inc . ) following manufacturer's protocols . | Group B Streptococcus ( GBS ) is the leading cause of neonatal invasive diseases and pili , as long filamentous fibers protruding from the bacterial surface , have been discovered as important virulence factors and potential vaccine candidates . The bacterial surface is the main interface between host and pathogen , and the ability of the host to identify molecular determinants that are unique to pathogens has a crucial role for microbial clearance . Here , we describe a strategy to investigate the immunological and structural proprieties of a protective pilus protein , by elucidating the molecular mechanisms , in terms of single residue contributions , by which functional epitopes guide bacterial clearance . We generated neutralizing monoclonal antibodies raised against the protein and identified the epitope region in the antigen . Then , we performed computational docking analysis of the antibodies in complex with the target antigen and identified specific residues on the target protein that mediate hydrophobic interactions at the binding interface . Our results suggest that a perfect balance of shape and charges at the binding interface in antibody/antigen interactions is crucial for the antibody/antigen complex in driving a successful neutralizing response . Knowing the native molecular architecture of protective determinants might be useful to selectively engineer the antigens for effective vaccine formulations . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"sequence",
"analysis",
"computer",
"science",
"computer",
"modeling",
"antigen",
"processing",
"and",
"recognition",
"streptococci",
"protein",
"structure",
"immunology",
"host-pathogen",
"interaction",
"biology",
"computational",
"biology",
"microbiology",
"computerized",
... | 2013 | Understanding the Molecular Determinants Driving the Immunological Specificity of the Protective Pilus 2a Backbone Protein of Group B Streptococcus |
Genetic transformation , in which cells internalize exogenous DNA and integrate it into their chromosome , is widespread in the bacterial kingdom . It involves a specialized membrane-associated machinery for binding double-stranded ( ds ) DNA and uptake of single-stranded ( ss ) fragments . In the human pathogen Streptococcus pneumoniae , this machinery is specifically assembled at competence . The EndA nuclease , a constitutively expressed virulence factor , is recruited during competence to play the key role of converting dsDNA into ssDNA for uptake . Here we use fluorescence microscopy to show that EndA is uniformly distributed in the membrane of noncompetent cells and relocalizes at midcell during competence . This recruitment requires the dsDNA receptor ComEA . We also show that under ‘static’ binding conditions , i . e . , in cells impaired for uptake , EndA and ComEA colocalize at midcell , together with fluorescent end-labelled dsDNA ( Cy3-dsDNA ) . We conclude that midcell clustering of EndA reflects its recruitment to the DNA uptake machinery rather than its sequestration away from this machinery to protect transforming DNA from extensive degradation . In contrast , a fraction of ComEA molecules were located at cell poles post-competence , suggesting the pole as the site of degradation of the dsDNA receptor . In uptake-proficient cells , we used Cy3-dsDNA molecules enabling expression of a GFP fusion upon chromosomal integration to identify transformed cells as GFP producers 60–70 min after initial contact between DNA and competent cells . Recording of images since initial cell-DNA contact allowed us to look back to the uptake period for these transformed cells . Cy3-DNA foci were thus detected at the cell surface 10–11 min post-initial contact , all exclusively found at midcell , strongly suggesting that active uptake of transforming DNA takes place at this position in pneumococci . We discuss how midcell uptake could influence homology search , and the likelihood that midcell uptake is characteristic of cocci and/or the growth phase-dependency of competence .
Bacterial transformation , a programmed mechanism for genetic exchange , is based on the uptake and integration of exogenous DNA into the recipient genome . This exogenous DNA is captured in double-stranded ( ds ) form and internalized as single strands ( ss ) . Gram-positive and Gram-negative microorganisms use related assemblies of proteins to internalize , protect and process transforming DNA [1]–[4] . Internalizing DNA is a complex process as it must cross the outer membrane ( in Gram-negative bacteria ) , the cell wall and the cytoplasmic membrane . Uptake of exogenous DNA depends on the formation of a structure evolutionarily related to type IV pili and type-2 secretion systems [3] , originally called the transformation pseudopilus [5] . A large macromolecular complex containing ComGC was found at the surface of competent Bacillus subtilis cells [6] . Most recently , a bona fide type IV transformation pilus was detected at the surface of competent Streptococcus pneumoniae [7] . Assembly of this pilus requires the traffic ATPase ComGA [5] , [7] . Although the mode of action of the transformation pilus is still poorly understood , it binds DNA [7] and is known to be required for access of exogenous dsDNA to its membrane-bound receptor , ComEA , DNA binding being abolished in comGA mutant cells [8] . While evidence that DNA enters the cell in ss form has been obtained for only a very small number of species , all models for DNA uptake postulate entry of ssDNA into the cytosol through a transmembrane channel formed by ComEC [9] with ComFA acting as a translocase [10] . However , it is only with the Gram-positive human pathogen S . pneumoniae that the protein responsible for converting dsDNA into ssDNA is known . This sequence non-specific endonuclease , EndA , was the first component of the DNA-uptake apparatus identified [11] . This membrane protein [12] , [13] , is not required for initial dsDNA binding [11] but was proposed to degrade the non-transported strand [13] . Consistent with this role , degradation and uptake were shown to proceed with opposite polarities , and occur at similar rates ( ∼100 nt s−1 at 31°C ) , suggesting the functional coupling of degradation of one strand with import of its complement [14] . EndA is also a virulence factor; it enables pneumococci to escape the innate host immune response by degrading the DNA-scaffold trap elaborated by neutrophils [15] . EndA pre-exists in cells at the time they abruptly and simultaneously develop the ability to internalize ssDNA or competence . Competence is induced in nearly all cells of an exponentially growing culture in response to a peptide pheromone , CSP ( competence-stimulating peptide ) [16] , for a period of time as short as ∼20 min [17] . CSP induction ultimately activates the synthesis of the comX encoded competence-specific alternative σX factor [18] required for expression of the late competence ( com ) genes including those that encode the transformation machinery [19] , [20] . The fact that EndA , uniquely among proteins of the uptake apparatus , is already present in cells prior to competence raises the question of its subcellular localization in noncompetent cells and its recruitment to the transformation machinery . Previous experiments had indicated that EndA-dependent degradation required the presence of the transformation pilus and the dsDNA receptor ComEA but not the transmembrane channel ComEC or the putative translocase ComFA [8] . Furthermore , endA mutant cells were found to accumulate DNA at the cell surface , presumably as a consequence of drastically reduced uptake . Here , we investigate EndA deployment and use it as a tool to document the subcellular localization of both bound transforming DNA and the DNA uptake machinery in competent pneumococci .
We first compared the localization of EndA in competent and noncompetent cells using strains expressing EndA fluorescent protein fusions . The rationale for designing the fusions , the chromosomal location of chimeric genes , and the evaluation of functionality and impact on transformation frequency are summarized in Figure S1 . Briefly , both GFP-EndA and YFP-EndA fusions were fully active , and modulation of their expression levels had no detectable effect on transformation frequency . Fluorescence microscopy analysis revealed that YFP-EndA is distributed all around the membrane in noncompetent cells ( Figure 1A ) . Ten min after CSP addition and irrespective of the presence of exogenous DNA a different localization pattern emerged , with a number of cells exhibiting one or two discrete YFP-EndA foci ( Figure 1A ) . Automated focus detection of over 2 , 500 cells using SpotFinder ( MicrobeTracker image analysis software [21] ) revealed that 6% of the cells contained foci , most of them ( 185/201 ) containing just one ( Figure 1B ) . In cells without constriction ( the most abundant cell type representing 77% of the population ) , about half the foci ( 47% ) were at midcell ( Figure 1C ) and only 10% near a pole . Two widely different interpretations could account for the competence-dependent clustering of EndA . The straightforward explanation is that the protein is recruited at the entry pore as an active component of the uptake machinery . An alternative , which takes into account EndA's excess amount for transformation ( Figure S1 ) and role as a virulence nuclease degrading dsDNA , would be that its clustering reflects the need to keep it away from the entry pore to protect transforming DNA . Detection of foci in only a small fraction of the cells appeared paradoxical since all pneumococci in a culture are known to develop competence [22] . To determine whether this reflected brevity of focus life-time relative to the ∼20 min [17] competence period , we measured focus frequency as a function of time after CSP addition . This analysis revealed that the proportion of positive cells increased rapidly and reached a maximum ( ∼20% ) 8 min after CSP addition , almost coinciding with maximal transformation proficiency in the population ( Figure 1D ) . A 2-fold decrease was then observed over the next 10 min , and ∼60 min after CSP addition the frequency of cells with one focus was close to that measured in noncompetent cultures . The frequency of cells with foci paralleled the kinetics of transformation , suggesting a link between the capacity of YFP-EndA to concentrate into foci and DNA-uptake potential . Time-lapse microscopy analysis then indicated that foci eventually formed in all cells during the competence window but lasted for only a few minutes , sometimes reforming at a different position in the same cell ( Figure 1E and Movie S1 ) . The disappearance of foci followed by their reformation at a different position strongly suggests that at the single-cell level competence is lasting longer than the lifespan of an individual focus . This dynamic behavior of EndA is consistent with a model in which the nuclease is transiently recruited at specific locations during competence development . It is of note that relocalization appeared to involve a large fraction of EndA molecules , despite evidence that only a small fraction ( <10% ) of them is required for DNA uptake ( [23] and Figure S1 ) . To investigate whether competence-dependent localization of EndA required expression of a late com gene , we examined the distribution of YFP-EndA in a comX mutant strain after CSP addition . The subcellular localization of YFP-EndA appeared indistinguishable from that seen in noncompetent cultures with no sign of clustering ( Figure S2A ) . To identify the late com gene involved , we investigated YFP-EndA localization in a series of mutants lacking components of the transformation machinery . YFP-EndA was still able to concentrate into foci 10 min after CSP addition in ∼5% of cells lacking ComGA , ComEC or ComFA , with foci preferentially localized at midcell ( Figure S2 , A and B ) , as observed in wildtype cells . In contrast , YFP-EndA remained uniformly distributed around the membrane and failed to form foci in the absence of ComEA ( Figure S2A ) . To rule out the possibility that the failure to form foci was due to degradation of the fusion protein , we monitored YFP-EndA levels in all mutants . Immunoblot analysis demonstrated that the amount of YFP-EndA in the absence of ComX , ComGA , ComEC , ComFA or ComEA was similar to that in wildtype cells ( Figure S2C ) . These results demonstrate that competence-induced clustering of EndA relies on the dsDNA receptor ComEA , and is independent of the presence of other components of the DNA uptake machinery such as the transformation pilus , the transmembrane channel or the DNA translocase . To help distinguish between a clustering of EndA indicative of its recruitment to the DNA uptake machinery or sequestration away from the site of uptake , we investigated the location of DNA bound at the surface of competent pneumococci . We took advantage of the documented accumulation of DNA molecules at the surface of endA− cells , resulting from reduced uptake in the mutant [8] . Short DNA fragments ( 285 bp; to lessen spreading at the cell surface ) harboring a single Cy3 label at the 5′ end ( Cy3-DNA ) were readily detected as fluorescent signals bound to competent endA− cells ( Figure 2A ) . In contrast , we detected no fluorescence with noncompetent cells . Signal intensity was proportional to donor DNA concentration ( Figure 2B ) , suggesting that the fluorescence detected represents bona fide competence-specific DNA binding . Further support for this conclusion was provided by the failure to detect bound Cy3-DNA in the absence of transformation pilus ( i . e . , in endA comGA mutant cells ) . In addition , no Cy3-DNA signal was observed with endA comEA double mutant cells ( Figure 2A ) , strongly suggesting that DNA is retained at the surface of endA− cells through binding to the dsDNA receptor ComEA . Bound Cy3-DNA was preferentially located at midcell ( Figure 2C ) , with essentially no polar location detected , suggesting that the site of DNA binding could coincide with EndA location as determined in wildtype cells . To investigate the possible co-localization of EndA and bound DNA , we constructed a GFP fusion with an EndA protein devoid of nuclease activity , GFP-EndAH160A ( referred to as EndA0 ) , based on the recent identification of the active site of EndA [24] , [25] . Co-localization of Cy3-DNA and EndA0 was detected ( Figure 2D ) . We also analyzed the possible co-localization of ComEA with Cy3-DNA using a strain harboring a functional GFP-ComEA fusion at the comEA locus . Similarly to EndA , ComEA frequently colocalized with DNA ( Figure 2D ) . We conclude from these observations that midcell clustering of EndA corresponds to its recruitment to the DNA uptake machinery . We then examined the subcellular localization of the GFP-ComEA fusion protein throughout a competence cycle ( Figure 3A ) . As observed with EndA , ComEA appeared clustered in a subset of the competent population and the number of cells containing GFP-ComEA foci increased progressively until transformability reached a maximum in the culture . A closer examination of the positioning of EndA and ComEA foci suggested that these two proteins occupy similar locations at early time points following CSP addition ( compare Figures 1D and 3A ) . Specifically , about half the first foci detected at 2 min post-CSP addition were close to midcell , consistent with emergence of foci at this position . This observation , together with the dependency of EndA clustering on ComEA and the co-localization of bound DNA with both proteins ( Figure 2D ) , suggests a direct interaction of EndA and ComEA , an hypothesis which received indirect support ( see Text S1 ) . However , ComEA and EndA did not colocalize permanently . Thus , while the fraction of cells containing EndA foci rapidly decreased 10 to 20 min after CSP addition and virtually disappeared from the culture by 60 min ( Figure 1D ) , the fraction of cells containing GFP-ComEA foci remained steady until 40 min after CSP addition , and persisted in cells that were no longer competent ( 60 min time point; Figure 3A ) . A re-treatment of data shown in Figure 3A , involving recording also as polar foci those located in the section immediately next to the pole , revealed that the proportion of cells with GFP-ComEA foci near the poles increased significantly post-competence , i . e . , 40 and 60 min after CSP addition ( Figure S4 ) . One interpretation of this localization pattern is that ComEA diffuses in the membrane from midcell toward the poles . Yet , ComEA did not accumulate further at the poles . Thus , the fraction of cells containing ComEA foci fell to 6% at 90 min ( none with a focus at midcell ) and vanished by 120 min , suggesting that ComEA could be progressively degraded once it reached the pole . Consistent with this hypothesis , increasing amounts of degradation products of GFP-ComEA protein appeared as competence declined ( Figure 3B ) . Post-competence ComEA , i . e . , molecules present 60 min after CSP addition were unable to bind freshly added Cy3-DNA , possibly because of the disappearance of the transformation pilus . Nevertheless , a significant fraction of Cy3-DNA added early to endA− cells ( i . e . , at the time of CSP addition ) remained bound to the surface after 60 min ( Figure 3C ) . Examination of the location of this persistent material revealed again clear signs of migration toward the pole ( Figure 3C , yellow arrows ) . The persistence of DNA bound to the cell surface was accompanied by stabilization of ComEA ( Figure 3D ) , presumably a result of the interaction of the receptor with its dsDNA substrate . Evidence that DNA binding occurs at midcell was obtained under somewhat artificial conditions involving the use of endA− cells , which accumulate DNA molecules at the surface because of reduced uptake . We therefore wished to obtain evidence for midcell DNA binding in wildtype cells . Because multiple fluorophores internal to DNA are likely to interfere with its uptake [26] , [27] , and since uptake proceeds with 3′-5′ polarity [28] , we first verified that pneumococcal cells could internalize DNA bearing a single Cy3 at its 5′ end . A Cy3-labelled 4 . 3-kb PCR fragment containing an SmR mutation transformed as efficiently as the unlabeled control ( Figure 4A ) , indicating that the bulky terminal fluorophore did not interfere with uptake; this result did not , however , show that the fragment had been internalized with Cy3 still attached . We then directly visualized the transformation process by fluorescence microscopy in living wildtype cells . The recipient strain harbored the gfp orf joined in frame to the 3′ end of ftsZ but separated from it by a stop codon . The Cy3-labelled donor fragment did not contain the stop codon , enabling expression of the FtsZ-GFP fusion upon integration into the chromosome ( Figure 4B ) . Cy3-DNA was added to cells together with CSP and samples were taken 5 min later for recording at time intervals both DNA fluorescence and the eventual appearance of FtsZ-GFP . The latter allowed detection of transformed cells 60–70 min after sampling ( Figures 4B and S5A ) . Looking back to the early period of DNA contact with these cells , we saw some Cy3-DNA foci at the surface 10–11 min post-sampling ( Figures 4B and S5A ) . All foci were found exclusively at midcell , strongly suggesting that binding and presumably active uptake of transforming DNA takes place at this position in wildtype pneumococci . This interpretation was further supported by time-lapse microscopy of wildtype cells using Cy3-DNA fragments of widely different sizes ( 285 bp and 12 kb ) . Importantly , midcell fluorescence disappeared faster with the shorter fragment , which was detected at a single time-point , whereas the 12 kb fragment remained visible for about 2 min , a figure fully consistent with the rate of DNA uptake measured previously ( ∼100 nt s−1 at 31°C ) [14] ( Figure S5B ) . We conclude that the rapid disappearance of midcell fluorescence in wildtype cells results from active uptake of transforming DNA .
The EndA nuclease has a dual role in S . pneumoniae . It is a key component of the transforming DNA uptake machinery in competent cells [11] and a virulence nuclease [15] . It is unclear however how this membrane-bound protein [12] , [13] fulfill the latter role since previous experiments showed a complete lack of degradation of transforming DNA added to comGA mutant cells [8] . This implied that despite its membrane location , EndA is unable to access exogenous DNA by itself . An explanation would be that EndA molecules responsible for degradation of the DNA-scaffold trap elaborated by neutrophils are not membrane-bound . Consistent with this view , cell-free filtrates of noncompetent endA+ cultures were able to degrade DNA , strongly suggesting that some EndA molecules are liberated into the medium during normal growth [29] . EndA release could occur by autolysis , possibly subsequent to fratricide [30] , [31] , or by an unknown mechanism . The present study establishes two different subcellular localizations for EndA which is uniformly distributed in the membrane of noncompetent cells and relocalizes at midcell during competence . Recruitment at midcell is dependent on the dsDNA receptor ComEA . However , the observation that EndA no longer clusters in post-competent cells despite the persistence of ComEA foci suggests that ComEA alone might not be sufficient for clustering of the nuclease . Thus EndA clustering may require , in addition to ComEA , either competence-specific physiological conditions or another com gene product , different from the proteins constituting the transformation pilus , the transmembrane channel or the DNA translocase . Nevertheless , ComEA also localizes at midcell , together with bound dsDNA , and evidence suggests that transforming DNA uptake occurs at this position . It was suggested that EndA is fixed asymmetrically near the DNA entry pore and that degradation of the nontransported strand occurs by successive endonucleolytic cleavages [13] . While EndA cleaves ssDNA slightly faster than dsDNA in vitro [24] , [25] , the interaction with ComEA could modulate EndA activity so as to degrade only one strand to allow import of its complement . Alternatively , another component of the DNA uptake machinery could be responsible for the necessary adaptation of EndA activity to the uptake process . Midcell uptake in S . pneumoniae contrasts with the situation in B . subtilis where several lines of evidence indicate that the subcellular site for active uptake is the cell pole [26] , [32]–[35] . Proteins involved in integration of entering ssDNA into the B . subtilis chromosome , including the recombinase RecA , are also located at the pole . Extension towards the nucleoid of RecA nucleocomplexes formed at the pole is expected to be seen in this species , and threads emerging from polar RecA foci and extending into the cytosol were interpreted as such [32] . What could be the consequence of midcell uptake in S . pneumoniae ? If accompanied by co-localization of RecA , this would result in formation of RecA nucleofilaments very close to the nucleoid . This situation may favor immediate inclusion of ssDNA-RecA complexes in the nucleoid , a ‘caging’ representing the most favourable situation for RecA-driven homology search . This suggestion is based on previous findings that an increase in DNA concentration ( including in the form of completely heterologous DNA ) accelerates homologous pairing through facilitated diffusion within nucleoprotein networks , also known as intersegmental homology search [36] , [37] . Midcell uptake in S . pneumoniae could thus help accelerate chromosomal integration of internalized ssDNA , contributing to the rapid kinetics of a process known to be entirely completed within ∼10 min [1] , in a species in which competence develops suddenly and disappears almost as rapidly owing to intrinsic shut-off mechanisms [38] , [39] . What could account for the different locations of the uptake machinery in B . subtilis and S . pneumoniae ? One major difference is seen in the growth-phase dependence of spontaneous competence development . The two species have evolved specific regulatory cascades , best adapted to their lifestyles , for tightly controlling assembly of orthologous DNA-uptake machines [40] . Under laboratory conditions , these cascades lead to competence induction during the transition to stationary phase in B . subtilis cultures , whereas pneumococcal competence develops in early exponential growth phase , i . e . , in actively dividing cells . These differences could introduce species-specific constraints that account for the different implantation of transformation machineries . Another noticeable difference between these two bacteria is their shape . B . subtilis is a straight-rod shaped bacterium whereas S . pneumoniae is a rugby-ball shaped coccus . The morphology of bacterial cells is ultimately determined by the peptidoglycan cell wall . Large transenvelope complexes such as the transformation machinery presumably require the formation of gaps in the peptidoglycan for their assembly . An attractive idea is that remodeling by lytic enzymes [41] , [42] occurring at the site of nascent peptidoglycan synthesis enables assembly of this apparatus . In support of this hypothesis , peptidoglycan synthesis occurs at midcell in actively growing pneumococci [43] . In B . subtilis , since competence develops at the onset of stationary phase , when the rate of peptidoglycan synthesis is slowing [44] , it is tempting to speculate that the transformation machinery assembles at the septum during the last division cycle . Subsequent daughter cell separation would then result in the observed location , thus predicted to correspond to new poles . This interpretation would fit nicely with a recent report that ComN , which is involved in posttranscriptional control of comEA and comEC expression [45] , localizes in a DivIVA-dependent manner at midcell where it promotes accumulation of the corresponding mRNA [46] . While the biological significance of this observation could not be established , the intriguing possibility that ComN could be required for proper localization of ComEC was not ruled out [46] . Nevertheless , the observation of competence complexes at both poles in B . subtilis cells and of polar complexes at non-adjacent poles in divided cells that have not separated ( doublets ) argues against assembly of the machinery at the septum during the last division . On the other hand , the existence of a significant fraction of non-polar foci ( up to 30% ) was reported in B . subtilis [26] , [35] and it remains unclear whether they represent intermediates moving toward the poles for subsequent functioning , or degradation as we suggest for S . pneumoniae . In any case , investigations of species that exhibit different morphology and develop competence at different growth stages will help establish whether location of the transformation machinery at the site of peptidoglycan disruption is a general rule . If so , one would predict physical and possibly functional interaction between components of the transformation and division assemblies .
S . pneumoniae strains , plasmids and oligonucleotide primers used for PCR and site-directed mutagenesis are listed in Table S1 . Stock cultures of pneumococal strains were routinely grown at 37° in Todd–Hewitt ( BD Diagnostic System ) plus yeast extract ( THY ) medium to OD550 = 0 . 3; after addition44 of 15% ( vol/vol ) glycerol , stocks were kept frozen at −70°C . These precultures were used to initiate cultures in C+Y at 6×106 cells mL−1 . E . coli strains used were LE392 and BL21 DE3 pLysS . Transformation was performed as described previously [22] . To monitor transformation proficiency ( Figure 1D and Figure 3A ) , at the indicated times after CSP1 addition 1 mL competent cells were incubated for 3 min at 37°C with 100 ng of R304 chromosomal DNA harboring the rpsL41 point mutation conferring resistance to streptomycin ( SmR ) . Uptake was terminated by addition of DNase I ( 50 mg mL−1; SIGMA ) and incubation was continued for 17 min at 30°C before plating . To investigate the impact of 5′ terminal label with a single Cy3 dye molecule on transformation ( Figure 4A ) , a 4 . 2-kb fragment carrying the rpsL41 SmR allele was amplified from R304 chromosomal DNA using the OCN79-OCN80 primer pair to generate a Cy3-labelled fragment or the RpsL5-RpsL6 primer pair to generate unlabelled control fragment . Details of plasmid and strain constructions and transformation procedures are described in Text S2 . At the indicated times after CSP1 addition , the OD550 was measured ( for equivalent loading ) and samples ( 3 mL ) were collected by centrifugation . Cell pellets from the CSP-induced and control cultures were stored at −80°C . Whole-cell extracts were prepared by resuspension of cell pellets in 50 µl lysis buffer [10 mM Tris pH 8 . 0 , 1 mM EDTA , 0 . 01% ( wt/vol ) DOC , 0 . 02% ( wt/vol ) SDS] and incubation at 37°C for 10 min followed by addition of 50 µl loading buffer [0 . 25 M Tris pH 6 . 8 , 6% ( wt/vol ) SDS , 10 mM EDTA , 20% ( vol/vol ) Glycerol] containing 10% ( vol/vol ) β-mercaptoethanol . Samples were heated for 15 min at 50°C prior to loading . Proteins were separated on pre-cast 4–12% NuPage Bis-Tris gels ( Invitrogen ) with MOPS-SDS running buffer , transferred to a nitrocellulose membrane using an iBLOT apparatus ( Invitrogen ) and blocked in 8% ( wt/vol ) skimmed milk in Tris-buffered saline ( TBS ) ( 50 mM Tris-HCl , 150 mM NaCl , pH 8 ) containing 0 . 1% ( vol/vol ) Tween-20 . The blocked membrane was probed with anti-GFP , or anti-SsbB antibodies [47] . Primary antibodies were diluted 1∶10 , 000 into 5% ( wt/vol ) skimmed milk in TBS supplemented with 0 . 01% ( vol/vol ) Tween-20 . Primary antibodies were detected using peroxidase-conjugated goat anti-rabbit immunoglobulin G ( Sigma ) with ECL Western Blotting Detection System ( GE Healthcare® ) and a luminescent image analyzer ( LAS-4000 , Fuji ) . Signals were quantified with a MultiGauge V3 . 0 Software ( Fugifilm ) . Pneumococcal precultures grown in C+Y medium at 37°C to an OD550 of 0 . 06 were induced to develop competence . At indicated times post CSP addition , 1 mL samples were collected , cooled down by addition of 500 µL cold medium , pelleted ( 3 min , 3 , 000 g ) and resuspended in 50 µL C+Y medium . 2 µL of this suspension were spotted on a microscope slide containing a slab of 1 . 2% C+Y agarose as described previously [48] before imaging . Images were captured and processed using the Nis-Elements AR software ( Nikon ) and deconvolution of fluorescent images was carried out using the SVI HuygensEss software v . 4 . 4 ( Scientific Volume Imaging B . V . , VB Hilversum , Netherlands ) . Images were further analyzed using the MATLAB-based , open-source software MicrobeTracker [21] . For more details , see Text S2 . Cy3-DNA fragments used for microscopy analysis were generated by PCR reaction . The 285-bp , 4 . 3-kb and 12-kb Cy3-DNA fragments were amplified from R304 chromosomal DNA using the OCN75-OCN76 , OCN79-OCN80 and OCN77-OCN78 primer pairs respectively . The 2 . 2-kb Cy3-DNA fragment encoding the functional ftsZ-gfp fusion was amplified from strain R3702 with primer pair OMB99-OMB100 . To directly visualize the transformation process , we used strain R3708 harboring the gfp orf joined in frame to the 3′ end of ftsZ but separated from it by a stop codon ( TAA ) as a recipient strain . This strain was induced to develop competence and simultaneously incubated for 5 minutes with the 2 . 2-kb Cy3-DNA fragment containing a functional ftsZ-gfp fusion construct ( TAA replaced by CTC , coding for Leu ) . Samples were subsequently spotted on a microscope slide and time lapse experiments were performed capturing images in the red ( Cy3 ) and green ( GFP ) channels . | Natural genetic transformation , a programmed mechanism for horizontal gene transfer , permits the passage of environmental double-stranded ( ds ) DNA through the bacterial membrane and its subsequent integration into the recipient chromosome by homology . In the human pathogen Streptococcus pneumoniae , it requires development of a physiological state termed competence , which develops transiently in nearly all cells of an exponentially growing culture . Expression of a specific set of genes then allows assembly of a large membrane-associated machinery for binding exogenous dsDNA and internalizing single-stranded ( ss ) DNA fragments . The key role of converting dsDNA into ssDNA is fulfilled by EndA , a membrane-located endonuclease which is also a pneumococcal virulence factor pre-existing in noncompetent cells . Here , we report that EndA is uniformly distributed in the membrane of noncompetent cells and relocates into clusters during competence . We show that this relocalization is dependent upon the dsDNA-receptor ComEA and that ComEA and EndA are preferentially located at midcell in cultures exhibiting maximal transformation proficiency . Finally , using fluorescence microscopy , we visualize the transformation process in living cells providing evidence that DNA binding and presumably uptake occur at midcell . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"streptococci",
"genetics",
"microbial",
"pathogens",
"molecular",
"genetics",
"biology",
"microbiology",
"molecular",
"cell",
"biology",
"bacterial",
"pathogens",
"gram",
"positive"
] | 2013 | Midcell Recruitment of the DNA Uptake and Virulence Nuclease, EndA, for Pneumococcal Transformation |
Aedes aegypti is the major vector of yellow and dengue fevers . After 10 generations of adult selection , an A . aegypti strain ( SP ) developed 1650-fold resistance to permethrin , which is one of the most widely used pyrethroid insecticides for mosquito control . SP larvae also developed 8790-fold resistance following selection of the adults . Prior to the selections , the frequencies of V1016G and F1534C mutations in domains II and III , respectively , of voltage-sensitive sodium channel ( Vssc , the target site of pyrethroid insecticide ) were 0 . 44 and 0 . 56 , respectively . In contrast , only G1016 alleles were present after two permethrin selections , indicating that G1016 can more contribute to the insensitivity of Vssc than C1534 . In vivo metabolism studies showed that the SP strain excreted permethrin metabolites more rapidly than a susceptible SMK strain . Pretreatment with piperonyl butoxide caused strong inhibition of excretion of permethrin metabolites , suggesting that cytochrome P450 monooxygenases ( P450s ) play an important role in resistance development . In vitro metabolism studies also indicated an association of P450s with resistance . Microarray analysis showed that multiple P450 genes were over expressed during the larval and adult stages in the SP strain . Following quantitative real time PCR , we focused on two P450 isoforms , CYP9M6 and CYP6BB2 . Transcription levels of these P450s were well correlated with the rate of permethrin excretion and they were certainly capable of detoxifying permethrin to 4′-HO-permethrin . Over expression of CYP9M6 was partially due to gene amplification . There was no significant difference in the rate of permethrin reduction from cuticle between SP and SMK strains .
The yellow fever mosquito Aedes aegypti inhabits tropical and subtropical regions worldwide and is the major vector of dengue fever ( DF ) and yellow fever . DF is a rapidly growing health issue , with the average annual number of cases being approximately 100 million and a 30-fold increase in the past 50 years [1] , [2] . The disease is endemic in at least 112 countries , especially in south and Southeast Asia , and an estimated 2 . 5 billion people are currently living in risk areas [1] . At present , DF causes more illness and death than any other arbovirus disease in humans [3]-[5] . Successful population control of vector insects is the key to prevent transmission and epidemics of infectious diseases . This strategy relies heavily on insecticides , particularly pyrethroid , a popular class of insecticides with high and rapid toxic activity toward insects and low toxicity to mammals [6] . Pyrethroids are used to control and/or prevent adult mosquitoes by ultra-low volume sprays , thermal fogging , pyrethroid-impregnated nets , etc . However , many dengue endemic areas are now facing the problem of pyrethroid resistance due to frequent and intensive use of these chemicals [7] . Resistance of A . aegypti to pyrethroids has been reported from various countries [8] . Understanding the level and mechanism of resistance to insecticides is essential for developing appropriate vector control measures . A . aegypti is present in most residential areas of Singapore . Despite a well-established national vector control program that includes community engagement , law enforcement , and inter-sectional coordination , Singapore continues to face the risk of DF resurgence [9]–[11] . During an outbreak in 2005 , more than 14 , 000 DF cases were reported [12] . Since 2005 , surveillance for dengue control has been based on four pillars: ( 1 ) case surveillance through mandatory notification of dengue cases to the Ministry of Health , by all medical practitioners; ( 2 ) vector surveillance through premises checks by vector control officers from National Environment Agency ( NEA ) ; ( 3 ) virus genotype surveillance at Environmental Health Institute of NEA; and ( 4 ) monitoring of other environmental parameters such as weather factors and population density . Clustering of cases and development of risk maps using the surveillance data allows prioritization of vector control operations . Though the program is largely based on source reduction and large scale fogging has not been conducted by the authorities since 2005 , insecticides continue to be used by the authority for indoor misting in areas with dengue transmission , by private arrangement and household use of aerosol cans . Regular monitoring for insecticide has revealed that Singapore's A . aegypti larvae have developed resistance to synthetic pyrethroids , with resistance ratio to permethrin in the range of 29–47 times [13] . Generally , insecticide resistance in insects is caused by three major mechanisms: ( 1 ) reduced sensitivity of the target site , ( 2 ) reduced penetration of the insecticide due to altered cuticles , and ( 3 ) increased activity or level of detoxification enzyme ( s ) . Insects develop resistance to insecticides by obtaining one or more of these mechanisms . Amino acid substitution in the voltage-sensitive sodium channel ( Vssc ) , the target site of DDT and pyrethroids , is the most well-understood mechanism conferring resistance against pyrethroid insecticides [14] , [15] . This mechanism is called “knockdown resistance , ” with its inherited trait , “kdr , ” being first identified in the housefly as a leucine–phenylalanine substitution at position 1014 ( L1014F ) of Vssc . Recently , many other mutations that cause reduced sensitivity to Vssc have been identified from many arthropod species and are called either “kdr-like factor” or “knockdown resistance . ” Knockdown resistance is unaffected by synergists that inhibit the activity of metabolic enzymes such as carboxyl esterases or cytochrome P450 monooxygenases ( P450s ) . Several amino acid substitutions have been detected in A . aegypti Vssc: S989P , I1011M , I1011V , V1016G , V1016I , F1534C , and D1763Y [16]–[22] . Of these , only substitutions V1016G , V1016I , and F1534C have been shown to strongly correlate with pyrethroid resistance [17] , [23]–[27] . The dominance of these three mutations on pyrethroid sensitivity , however , is not well understood . Metabolism-mediated insecticide resistance is now considered a key mechanism in insects [7] , [28] , [29] . Three families of metabolic enzymes have been implicated in the metabolism of insecticides: esterases , glutathione transferases ( GSTs ) , and P450s . The genome project of A . aegypti identified 26 esterases , 49 GSTs , and 160 P450s [30] . Identification of the specific enzymes involved in insecticide resistance , however , has proven challenging . P450s has been shown to be the metabolic enzyme most strongly linked to the development of pyrethroid resistance [31]–[33] . However , due to the large number of P450 genes and the structural similarity among different isoforms , identification of the isoform ( s ) associated with resistance has been difficult , with only few exceptions [34] , [35] . Recently , microarray analysis that compares gene expression between susceptible- and resistant-strains unearthed candidate genes that confer insecticide resistance in different insect species [36]–[38] . Although molecular diagnosis of insecticide resistance that targets Vssc mutations is widely reported , such a system has not been reported for P450 genes in Aedes mosquitoes . Since insecticide resistance in a mosquito population could be concurrently affected by more than one mechanism , an accurate molecular diagnosis thus needs to consider the various possible mechanisms . Metabolism-mediated pyrethroid resistance of A . aegypti requires further study especially in the populations collected from Southeast Asia , the largest endemic area of DF . Reduced cuticle penetration is the least understood mechanism among the three . Though it may have a primary role in resistance [39]–[41] , it more often acts in combination with the other mechanism ( s ) . This study examined the mechanisms conferring pyrethroid resistance in an A . aegypti strain collected from Singapore and investigated three major mechanisms of resistance . The results are expected to lead to the development of more accurate resistance monitoring systems and contribute to the discovery of new insecticide target sites to overcome the challenge of insecticide resistant mosquitoes .
Mice were used as the blood source for mosquitoes . This study complies with the guidelines for animal experiments performed at National Institute of Infectious Diseases , Japan . The protocol for the utility of mice was approved by the Animal Use Committee of National Institute of Infectious Diseases ( registration numbers: 209052 , 210044 , 211031 and 112029 ) . A laboratory-susceptible reference strain of A . aegypti , SMK , was obtained from Sumitomo Chemical Co . , Ltd . in 2009 . This strain was originally from USA and has been maintained in the laboratory for at least 20 years without exposure to insecticides . The pyrethroid-resistant population , SPS0 , was collected from Singapore in 2009 . The SPS0 population was selected by exposure to permethrin for 10 generations as described below and the SP strain ( = SPS10 ) was established . Permethrin doses used for the selections are listed in Table 1 . Larvae were fed a ground diet of insect food ( Oriental Yeast Co . , Ltd . , Tokyo , Japan ) , and the adults were maintained on 10% sucrose . Females were given blood meals from mice . Both larvae and adults were reared at 27 ± 1°C with a photoperiod of 16∶8 ( L∶D ) h . Three technical grade insecticides were used: permethrin ( 94 . 1% , Sumitomo Chemical Co . , Ltd . , Osaka , Japan ) , temephos ( 94 . 9% , Wako Chemical Pure Industries , Ltd . , Osaka , Japan ) , and pirimiphos methyl ( 98 . 0% , Wako Chemical Pure Industries , Ltd . , Osaka , Japan ) . An inhibitor of P450s , piperonyl butoxide ( PBO , 98 . 0% , Wako Chemical Pure Industries , Ltd . , Osaka , Japan ) , was used for the adult and larval bioassays . [14C]- ( 1RS ) -trans-permethrin ( specific activity , 4 . 44 GBq/m mol , purity >99% ) labeled at the phenoxy benzyl group and 4′-hydroxy-3-phenoxy-benzyl 3- ( 2 , 2-dichlorovinyl ) -2 , 2-dimethylcyclopropane-carboxylate ( 4′-HO-permethrin ) were generously provided by Sumitomo Chemical Co . , Ltd . ( Osaka , Japan ) . [14C]- ( 1RS ) -trans-permethrin was purified by thin layer chromatography ( TLC ) according to the modified method described by Shono et al . [42] . We used a solvent system of toluene/carbon tetrachloride ( 1∶1 ) instead of benzene/carbon tetrachloride ( 1∶1 ) in order to separate trans-permethrin from other degraded compounds . Three- to seven-day-old adult male and virgin females were used for the permethrin selections . In order to avoid measurement error caused by insect locomotive activity , we used a topical application method , instead of filter paper method , for pyrethroid selections and adult bioassays . The mosquitoes were anaesthetized with CO2 and placed in a 200-ml Erlenmeyer glass flask . They were then re-anaesthetized for 3 min with 100 µl diethyl ether absorbed onto cotton . A volume of 0 . 22 µl permethrin in acetone solution was dropped onto the thoracic notum of each mosquito with a Hamilton repeating dispenser PB-600 equipped with 10 µl syringe 701SNR ( Hamilton , Reno , NV , USA ) . The insects were placed in groups of 40 in paper cups sealed with a nylon mesh . Water-absorbed cotton was placed on the top of the mesh . Approximately 1000 mosquitoes were selected from each generation ( Table 1 ) . The doses of permethrin used for the selections were determined by a preliminary experiment for each gender . Twenty-four hours after treatment , all the mosquitoes were put into a cage , provided with sucrose water , and maintained for the next generation . A bioassay to monitor permethrin susceptibility was conducted for each selected generation as described below . Adult bioassays were conducted by topical application using 3- to 7-day-old female mosquitoes and consisted of at least five doses of permethrin causing >0% and <100% mortality . All bioassays were run at 25°C with 40 mosquitoes per dose . Mortality was assessed 24 h after treatment , with mosquitoes that could not stand on the bottom of cup counted as dead . The LD50 values for each strain were calculated using log-probit mortality regression analysis [43] . To estimate the contribution of P450s to resistance , a 0 . 22 µl ( 5 µg ) aliquot of piperonyl butoxide ( PBO ) was placed on the thoracic notum of the mosquitoes 24 h prior to dosing with permethrin . A preliminary experiment confirmed that this PBO dose did not kill any of the mosquitoes . The resistance ratios were calculated based on the ratio of LD50 values for the SMK and SP strains . To obtain F1 progeny , 100 virgin females and 100 males were randomly crossed and the offspring was used in the permethrin bioassays . BC1 progeny was also obtained by backcrossing 100 females F1 ( SP♀ × SMK♂ ) and 100 SMK♂ . Opposite backcrossing was also conducted . Permethrin bioassays were conducted for both BC1 progenies as described above . Larval bioassays were carried out as described elsewhere [44] . Twenty early fourth instar larvae were exposed to different concentrations of insecticides in 50 ml distilled water and mortality was counted after 24 h exposure at 25°C . Alcohol solutions of the insecticides ( ×200 of final conc . , 250 µl ) were added to the water . For estimating synergistic effects , the larvae were treated with a sublethal dose of PBO ( 5 µg/ml ) in combination with permethrin . Mortality was assessed 24 h after treatment , with larvae that could not swim to the surface counted as dead . Three replications were run for each insecticide concentration , and the LC50 values for each strain and insecticide were calculated using log-probit mortality regression analysis [43] . DNA was extracted from individual male mosquitoes using the REDExtract-N-Amp Tissue PCR Kit ( Sigma , MO , USA ) . We used male mosquitoes to avoid gene contamination derived from sperm DNA in females . In A . aegypti , sex is determined by M-locus which is mapped on the 1st chromosome [45] . Since Vssc gene is located on the 3rd chromosome [24] , theoretically , genotyping of Vssc genes from males does not cause any gender bias . The extraction and tissue preparation solutions were mixed , and each mosquito was homogenized in a 200-µl PCR tube for 1 min at 1500 rpm using a shaking homogenizer ( MM300 , Retsch Co . , Haan , Germany ) in 62 . 5 µl of the mixture containing a zirconia ball ( 4 mm in diameter ) . After homogenization , the samples were incubated at room temperature for 30 min , followed by incubation at 95°C for 3 min . A 50-µl aliquot of neutralization solution was then added and mixed by vortexing . The extracted mixture was stored at −20°C . Male mosquitoes were genotyped for their Vssc alleles as described previously [46] . We targeted six amino acid positions to identify the candidate ( s ) for knockdown resistance: the typical kdr , L1014F , and five amino acid positions in Vssc previously identified from pyrethroid resistant A . aegypti ( i . e . , S989P , I1011M or V , V1016G or I , F1534C , and D1763Y ) [16] , [17] , [19] , [20] , [22] . In this study we numbered the amino acid position according to the sequence of the most abundant splice variant of the house fly Vssc ( GenBank accession nos . AAB47604 and AAB47605 ) . Partial DNA fragments of domains II , III , and IV were amplified by PCR using TaKaRa Ex Taq Hot Start Version ( Takara Bio , Shiga , Japan ) and three primer sets: AaSCF20 and AaSCR21 ( for domain II ) , AaSCF7 and AaSCR7 ( for domain III ) , and AlSCF6 and AlSCR8 ( for domain IV ) . The primer sequences are listed in Table S1 . The cycling conditions for PCR were as follows: initial denaturation at 95°C for 5 min , followed by 35 cycles of 94°C for 30 s , 55°C for 30 s , 72°C for 1 min , and a final extension step at 72°C for 5 min . The PCR products were then treated with illustra ExoStar ( GE Healthcare UK Ltd . , Little Chalfont , England ) to remove unincorporated primers and dNTPs , followed by sequencing with the following primers: AaSCF3 ( forward primer for domain II ) , AaSCR22 ( reverse primer for domain II ) , AaSCR8 ( reverse primer for domain III ) , and AlSCF7 ( forward primer for domain IV ) using an ABI 3130 Genetic Analyzer ( Applied Biosystems , Foster City , CA , USA ) . The sequences were assembled and aligned using the GENETYX software ( GENETYX Corporation , Tokyo , Japan ) . A dose of 600 dpm ( ca 0 . 88 ng ) [14C]-permethrin in 0 . 22 µl acetone was administered to the thoracic notum of female mosquitoes . Each treated mosquito was isolated in a scintillation vial ( 6 ml ) with a water-absorbed small cotton pad . The mosquitoes were anaesthetized with diethyl ether at 0 . 75 , 1 . 5 , 3 , 6 , 12 , 24 , and 48 h after treatment and rinsed in methanol ( 0 . 5 ml × 2 ) . The methanol was then added to 4 ml of scintillation cocktail Ultima Gold fluid ( PerkinElmer Inc . , Waltham , MA , USA ) and analyzed in a liquid scintillation counter LSC-3100II ( Hitachi Aloka Medical , Ltd . , Tokyo , Japan ) . To study effect of synergist on permethrin penetration , PBO ( 5 µg/♀ ) was placed on the thoracic notum or thoracic sternum 1 h before permethrin treatment . To assess the effects of solvent on permethrin penetration , 0 . 22 µl of acetone was applied to the thoracic notum 1 h prior to permethrin treatment ( negative control ) . Each time point was replicated four times , with 32 mosquitoes ( four mosquitoes × eight time points ) being used for each experiment . After being rinsed in methanol , each mosquito was placed into a 2-ml safe-lock tube ( Eppendorf Co . , Ltd . , Hamburg , Germany ) with a zirconia ball ( 4 mm in diameter ) and homogenized for 1 min at 1500 rpm with 0 . 5 ml of scintillation cocktail Ultima Gold fluid ( PerkinElmer Inc . , Waltham , MA , USA ) in a shaking homogenizer ( MM300 , Retsch Co . , Haan , Germany ) . The homogenized solution was transferred to a scintillation vial and rinsed again with 1 ml of scintillation cocktail . An additional 2 . 5-ml of cocktail was added to the vial ( total volume 4 ml ) and the radioactivity inside the mosquito was counted using a liquid scintillation counter . To measure the rate of permethrin excretion , the holding vial was extracted with 4 ml of scintillation cocktail and the radioactivity was counted using a liquid scintillation counter . Each time point was replicated four times ( four individual mosquitoes ) in each experiment . The rate constants for the in vivo dynamics of topically applied permethrin were calculated using the linear one compartment model described previously [47] , [48] . To identify the metabolites excreted into the vial , 10 SP females had ca 17 ng ( 11 , 800 dpm ) of [14C]-trans-permethrin in 0 . 44 µl acetone applied to the thoracic notum and were then placed into 20 ml glass vials . Three replicates were conducted . Forty-eight hours after treatment , the mosquitoes were anaesthetized with diethyl ether , removed from the vial , and the residual isotope in the vial was extracted with methanol ( 3 ml × 2 ) . The methanol was evaporated by N2 gas , and aliquots ( 5000 dpm ) were spotted onto silica gel plates 60 F254 ( HPLC , 0 . 2 mm , Merck KGaA , Damstadt , Germany ) , followed by development in a solvent of toluene/ethyl acetate ( 6∶1 ) . The HPTLC plate was auto-radiographed , scanned with a Bio-Imaging Analyzer BAS2500 ( Fuji Photo Film , Tokyo , Japan ) , and the signal intensities of each metabolite were quantified . Two-dimensional development was conducted for water soluble metabolites using a solvent of chloroform/methanol/water ( 65∶25∶4 ) and auto-radiographed as described above . Unlabeled authentic 4′-HO-permethrin was co-chromatographed and identified by viewing under ultraviolet light at 254 nm . Microsomes were prepared from the abdomens of 3-7-old female mosquitoes using a modified procedure as described previously [49] . Mosquitoes were anaesthetized with CO2 and transferred to a glass Petri dish placed on crushed ice . Two hundred abdomens were separated from the thoraces of mosquitoes using two sets of forceps and put into a glass container on ice containing 5 ml of homogenization buffer [49] . The collected abdomens were homogenized with glass–teflon Dounce homogenizer for 20 strokes using homogenizing mixer HK-1 ( Asone , Osaka , Japan ) and filtered through a layer of nylon wool . Homogenates were then centrifuged at 4°C at 10 , 000 xg for 15 min in an Eppendorf 5804R ( Eppendorf Co . , Ltd . , Hamburg , Germany ) equipped with a F-34-6-38 rotor . The 10 , 000 xg supernatant was centrifuged again at 4°C at 100 , 000 xg for 1 h in a Beckman Optima MAX-XP ( Beckman Coulter Inc . , Brea , CA , USA ) equipped with a MLS-50 rotor . The microsomal pellets were resuspended by homogenizing in 2 ml resuspension buffer [49] . The supernatant of the 100 , 000 xg centrifugation was also collected and used for an in vitro metabolism study . In vitro metabolism was studied as described previously with some modifications [44] . The 2-ml reaction mixture contained incubation buffer [0 . 1 M sodium phosphate buffer ( pH 7 . 5 ) containing 1 mM EDTA , 0 . 1 mM DTT , and 1 mM PMSF dissolved in ethylene glycol monomethyl ether] , microsomes or supernatant equivalent to 10 abdomens , 0 . 2 ml of 10 mM β-NADPH , and 100 , 000 dpm ( 0 . 147 µg ) of [14C]-trans-permethrin in 10 µl ethanol . An incubation mixture without β-NADPH served as the control . For the inhibition study , 10 µl of PBO ( 20 mM ) or 4′-HO-permethrin ( 0 . 5 , 5 . 0 , and 10 . 0 µg/ml equivalent to ×7 , ×70 , and ×140 final concentration of permethrin , respectively ) in ethanol was added to the mixture . The mixtures were incubated for 5 , 30 , 60 , 120 , and 360 min at 25°C with shaking , and then incubation was terminated by adding 0 . 2 ml of 1 N HCl , followed by 1 g of ( NH4 ) 2SO4 . Each sample was extracted with diethyl ether ( 4 ml × 3 ) , by vortexing for 1 min and then centrifuged at 4000 xg for 1 min , evaporated by N2 gas , and redissolved in 100 µl methanol . Aliquots of the extracted compounds ( 5 , 000 dpm ) were spotted on HPTLC plates and developed in a solvent of toluene/ethyl acetate ( 6∶1 ) . The HPTLC plate was auto-radiographed , scanned with a Bio-Imaging Analyzer , and the signal intensities were measured as described above . Each time point for each strain was replicated three times using different enzyme sources . All the experiments involved the simultaneous use of the SP and SMK strains . The metabolites being stuck at the spotting position of the plate after development were collected , extracted with methanol ( 4 ml × 2 ) , evaporated , and then spotted onto HPTLC plates . The high polar compounds were developed with a solvent of chloroform/methanol/water ( 65∶25∶4 ) and auto-radiographed as described above . The microarray used in this study was designed using the Agilent eArray platform ( Agilent Technologies , CA , USA ) and contained 60 mer oligo-probes representing >15 , 000 A . aegypti Liverpool transcripts identified in the genome project ( https://www . vectorbase . org/content/aedes-aegypti-liverpooltranscriptsaaegl13fagz ) . Each probe was spotted at least two times at different positions on each array . A probe of each P450 gene was spotted six times . Microarrays in a 4 × 44 k format were constructed using contract manufacturing carried out by Agilent Technologies . The entire design of the 44 k array used in this study is available from the NCBI Gene Expression Omnibus ( GEO ) site as accession # GPL17604 . The strains were reared in parallel in order to minimize variation resulting from breeding conditions . For each life stage ( larvae , adult males and females ) , four RNA sources were prepared from each of the SP and SMK strains reared in different trays and cages ( = four biological replicates ) . Total RNA was extracted from 10 fourth instar larvae or 10 three-day-old adults using ISOGEN ( Nippon Gene Co . , Ltd . , Tokyo , Japan ) . Genomic DNA was removed by digesting the total RNA samples with DNase I using TURBO DNase ( Life Technologies Co . , Carlsbad , CA , USA ) . The quality and quantity of total RNA were assessed by spectrophotometry using Nanodrop ND-1000 ( Thermo Fisher Scientific Inc . , Waltham , MA , USA ) and a bioanalyzer MultiNA ( Shimadzu Co . , Kyoto , Japan ) . The purified RNA was mixed with the internal control RNA supplied in a RNA Spike-In Kit ( One-color , Agilent Technologies ) . Fluorescein-labeled cRNA were synthesized via a double-stranded cDNA intermediate using a Low Input Quick Amp Labeling Kit ( Agilent Technologies ) . The cRNA derived from the SP and SMK strains were differentially labeled with cyanine-3 ( Cy-3 ) dye-conjugating CTP ( PerkinElmer Inc . , Waltham , MA , USA ) . cRNA was then purified with Qiagen RNeasy Mini Kit ( Qiagen , Venlo , Netherlands ) , and the overall efficacy of cRNA synthesis and fluorescein-labeling was measured using a Nanodrop ND-1000 spectrophotometer . The labeled cRNA were pooled and hybridized to microarray probes in a hybridization oven at 65°C for 17 h under rotation at 10 rpm using a Gene Expression Hybridization Kit ( Agilent Technologies , Santa Clara , CA , USA ) . After hybridization washing , the fluorescent dyes were stabilized against ozone oxidization with the Gene Expression Wash Buffer ( Agilent Technologies ) and the microarray plate was then dried in nitrogen gas . The fluorescence of Cy-3 on each spot was scanned using a DNA Microarray Scanner G565BA ( Agilent Technologies , Santa Clara , CA , USA ) . Spot identification and quantification were performed using Feature Extraction Software v . 7 . 5 ( Agilent Technologies , Santa Clara , CA , USA ) in the default setting . A linear and LOWESS algorithm ( a combination of the linear method and traditional LOWESS method ) was used for dye normalization . Flagged spots ( such as saturated , low intensity , and statistical outlier ) were ignored in the final data analysis . The ratio of transcription levels for each gene in the microarray experiment for the SP and SMK strains was called the “microarray ratio . ” A representative microarray ratio in each hybridization experiment was expressed as the average of all spots for a gene on an array , where the spot number for a gene is calculated as “unique probe number” x “replicating spot number ( ideally n = 4 ) . ” The P-value for every spot was calculated by Agilent's Universal Error Model . The raw results of the 10 microarray hybridizations are available from GEO ( series accession# GSE50069 ) . The genes over expressed in the SP strain were selected using a cut-off of >3-fold relative change in expression and a P<0 . 01 . In order to design primers for real time quantitative PCR , a partial or full length sequence of the P450 genes ( CYP6BB2 , CYP6Z7 , CYP6Z8 , CYP9M4 , CYP9M5 , CYP9M6 , CYP9M7 ) and an internal control gene ( Ribosomal protein S3 gene , RPS3 ) were amplified and sequenced from eight male mosquitoes for each of the SP and SMK strains . We focused on these P450 genes according to the following criteria: over expressed in females ( >5 ) and in males ( >3 ) in SP relative to SMK in microarray analysis . CYP9M7 was also analyzed as it forms a gene cluster with CYP9M4 , CYP9M5 , and CYP9M6 within 65 kbp on the same supercontig ( 1 . 29 ) and is structurally related to these genes . The primer sequences and their regions are shown in Table S1 and Figure S1 . Genomic DNA was extracted from individual male mosquitoes as described above . The PCR was conducted using high fidelity polymerase ( PrimeSTAR GXL DNA Polymerase; Takara Bio Inc . , Shiga , Japan ) , treated with illustra ExoStar ( GE Healthcare UK Ltd . , Little Chalfont , England ) to remove unincorporated primers and dNTPs , followed by direct sequencing with the primers listed in Table S1 . Because two haplotypes of CYP9M6 , namely CYP9M6v1 and CYP9M6v2 , were identified from the SP strain , the primer sets , 9M6F31/9M6R21 and 9M6F23/9M6R35 for CYP9M6v1 , and 9M6F30/9M6R22 and 9M6F18/9M6R35 for CYP9M6v2 were used to amplify each variant , followed by direct sequencing of the PCR products ( Figure S2 and Table S1 ) . cDNA was synthesized using total RNA isolated for microarray analysis , and the QT' primer ( Table S1 ) and reverse transcriptase ReverTra Ace ( Toyobo , Osaka , Japan ) according to the manufacturer's instructions . Real time quantitative PCR was performed using a PikoReal Real Time PCR System ( Thermo Fisher Scientific Inc . , Waltham , MA , USA ) . The PCR primers were designed using Primer Express software ( Applied Biosystems , Foster City , CA , USA ) . Each PCR reaction of 10 µl final volume contained 5 µl SYBR Premix Ex Taq II ( Takara Bio Inc . , Shiga , Japan ) , 1 µl cDNA ( equivalent to 10 ng total RNA ) , 0 . 4 µl of each forward and reverse primer ( 10 µM , Table S1 ) , and 3 . 2 µl ddH2O . The PCR reactions were performed under the following conditions: 95°C for 2 min , followed by 40 cycles of 95°C for 10 s , and 60°C for 30 s . The 2−ΔΔCt method [50] was used to quantify the relative expression level of P450s , with RPS3 gene acting as the internal control . For each gene analyzed , serial dilutions of cDNA showed that the efficacy of amplification for all P450 genes and the RPS3 gene were >0 . 998 . Three replicates of the PCR reactions were performed for each sample ( technical replications ) and each experiment was repeated four times using an independent RNA source ( biological replications ) . Gene copy number was determined by quantitative PCR using the same primer sets described above . Genomic DNA was individually extracted from eight virgin females of each strain using Get Pure DNA Kit-Cell , Tissue ( Dojindo Molecular Technologies , Inc . , Kumamoto , Japan ) according to the manufacturer's instructions . The quantity and quality of the DNA was assessed using a spectrophotometer Nanodrop ND-1000 ( Thermo Fisher Scientific Inc . , MA , USA ) . The DNA samples were diluted to 100 ng/µl and 1 µl used as a template . Data were normalized using RPS3 gene and the exact single gene copy number reported by the genome project . Three replicates of the PCR reactions were performed for each sample ( technical replications ) and each experiment was conducted using 8 individual DNA source ( biological replications ) . One hundred virgin males and females of the SP and SMK strains were cross-mated . The offspring were then interbred to obtain F2 mosquitoes . Three-day-old F2 females were dosed with a sub-lethal amount of [14C]-permethrin ( ca 600 dpm , 0 . 88 ng ) as described above . Each treated mosquito was isolated in a scintillation vial and the amount of radioisotope excreted was quantified 24 h later . Genomic DNA was isolated from six legs of each mosquito as described above using a REDExtract-N-Amp Tissue PCR Kit . Both CYP9M6 and CYP6BB2 were genotyped . Genotyping of CYP9M6 was carried by real time quantitative PCR using three sets of primer: 9M6F88/9M6R89 ( common ) , 9M6F95/9M6R97 ( specific to SP ) , and 9M6F96/9M6R97 ( specific to SMK ) . The primer sequences are listed in Table S1 . Genotyping of CYP6BB2 was carried out by analysis of the melting curve of different peaks in the post-PCR assay . PCR was performed using a TaKaRa Ex Taq Hot Start Version ( Takara Bio , Shiga , Japan ) , 6BB2F21/6BB2R22 primers , 0 . 5 µl EvaGreen ( Biotium , Inc . , Hayward , CA , USA ) , and 1 µl genomic DNA in 10 µl of reaction mixture . The cycling conditions for PCR were as follows: initial denaturation of 95°C for 1 min , followed by 50 cycles of 95°C for 10 s and 60°C for 5 s . This reaction produced 43 bp PCR products ( Figure S3 ) . The PCR and dissociation analysis were performed using a PikoReal Real Time PCR System ( Thermo Fisher Scientific Inc . , Waltham , MA , USA ) . Total RNA was isolated from the remaining bodies of the mosquitoes and cDNA was synthesized as described above . Real time quantitative PCR was performed for CYP9M6 and CYP6BB2 to quantify the levels of transcription of these genes , as described above . Forty-two females ( 24 females for each crossing ) were used for the analysis . Full-length cDNAs encoding CYP9M6 and CYP6BB2 from the SP strain were cloned into the pPSC8 protein expression vector ( Wako Chemical Pure Industries , Ltd . , Osaka , Japan ) using unique restriction sites positioned in the vector regions ( XbaI and PstI for CYP9M6 and XbaI and KpnI for CYP6BB2 ) . Because we identified two variants for CYP9M6 in the SP strain ( CYP9M6v1 and CYP9m6v2 ) , these two genes were amplified using primers 9M6F85/9M6R86 for CYP9M6v1 and 9M6F85/9M6R87 for CYP9M6v2 . For amplification of CYP6BB2 cDNA , 6BB2F19 and 6BB2R20 primers were used and cloned into the pPSC8 vector ( Table S1 and Figure S1 ) . A 200-µl aliquot of Sf900II medium ( Life Technologies Co . , CA , USA ) containing 2 µg of the vectors ( pPSC8/CYP9M6 or pPSC8/CYP6BB2 ) , 85 ng Linear AcNPV ( the baculovirus Autographa californica nuclear Polyhedrosis Virus ) , 5 µl Insect GeneJuice Transfection Reagent ( Merck KGaA , Damstadt , Germany ) was transfected to expressSF+ cells ( 1 × 106 cells in 25 cm2 flask , Wako Chemical Pure Industries , Ltd . , Osaka , Japan ) . The infected cells were incubated at 28°C for five days , centrifuged at 3000 xg for 30 min at 4°C , and the supernatant used for protein expression . Full-length cDNA encoding NADPH cytochrome P450 reductase ( PRE ) and cytochrome b5 cDNAs was also amplified with the aegREDF9/aegREDR5 and aegb5F1/aegb5R2 primer sets , respectively , and then cloned into the pFastBac1 vector ( Life Technologies Co . , Carlsbad , CA , USA ) using multiple cloning sites ( StuI and XbaI for PRE and EcoRI and XbaI for b5 ) . Recombinant constructs ( pFastBac1/PRE and pFastBac1/b5 ) were used to transform MAX Efficiency DH10Bac competent cells ( Life Technologies Co . , Carlsbad , CA , USA ) . Recombinant bacmid DNA was isolated according to the manufacturer's instructions . The Sf9 cells were infected with recombinant baculovirus in Grace's Insect Medium with Cellfectin II ( Life Technologies Co . , Carlsbad , CA , USA ) , according to the manufacturer's instructions . For preparation of cell lysates expressing P450 , 50 ml of confluent cells were infected in a 125-ml flask , and incubated at 28°C with shaking at 130 rpm . Twenty-four hours after infection , hemin ( in 50% ethanol and 0 . 1 N NaOH ) was added to the media to a final concentration of 2 µg/ml , followed by further incubation for 48 hours . The cells were then centrifuged at 3000 xg at 4°C for 30 min and the pellets were used for the in vitro permethrin metabolism study . The cell pellets prepared as described above were resuspended in 5 ml of homogenization buffer [49] and sonicated with an ultrasonic processor Sonics Vibra-Cell VCX130 ( Sonics & Materials , Inc . , Newtown , CT , USA ) for 2 min on ice water ( pulse on 10 s , pulse off 10 s , with 20% amplitude ) . The homogenates were centrifuged at 10 , 000 xg for 15 min and the supernatant was then centrifuged at 100 , 000 xg for 1 h as described above for the preparation of microsomes . The microsomal pellets were resuspended by homogenizing in 2 ml of resuspension buffer [49] , and the protein concentration was determined using Bradford's reagent ( Coomassie Plus Protein Assay; Thermo Fisher Scientific Inc . , Waltham , MA , USA ) . The in vitro permethrin metabolism study was conducted as described above using 7 mg of microsomal proteins . The experiment was replicated for tree times .
A field population of A . aegypti collected from Singapore was selected for permethrin resistance by subjecting a wild type population ( SPS0 ) to the chemical for 10 generations ( Table 1 ) . Initially , the resistance ratio ( RR ) of SPS0 was 35-fold . The RR rose rapidly over the generations and the established SP strain ( = SPS10 ) developed 1650-fold resistance after 10 generations of permethrin selection ( Figure 1A ) . The RR for each generation is summarized in Table 2 . To determine the degree of dominance of the resistance [51] , the SP strain was crossed with a susceptible reference strain ( SMK ) and bioassays were performed on the progenies . The degrees of dominance for F1 ( SP♀ × SMK♂ ) and F1 ( SMK♀ × SP♂ ) were −0 . 337 and −0 . 383 , respectively ( Figure 1B ) . The toxicity of permethrin was markedly increased in the SP strain by pretreatment with the P450 inhibitor piperonyl butoxide ( PBO ) , with the RR decreasing from 1650- to 33-fold . The synergistic ratio ( SR ) of the SP and SMK strains was 1 . 7 and 83 , respectively , indicating a significant role of P450s in permethrin resistance of SP ( Figure 1C and Table 2 ) . The low SR of SPS0 , at only 5 . 3 , suggested that the P450-mediated resistance mechanism was initiated or enhanced during the process of permethrin selection . We observed an approximately three-fold difference in the LD50s between SPS0 ( 9 . 8 ng/female ) and SP ( 30 ng/female ) in the presence of PBO . This suggested that a mechanism , other than P450s also contributed to the development of high level of permethrin resistance in SP ( Table 2 ) . The SP strain exhibited a 4 . 9-fold cross-resistance to pirimiphos methyl after permethrin selection . Although the laboratory selections were conducted during the adult stage , larvae of the SP strain also developed a high level of resistance to permethrin ( Figure 1D and Table 3 ) , from RR of 160-fold for SPS0 to 8790-fold for SP . PBO also increased the toxicity of permethrin in the larval stage . The SRs of SMK , SPS0 , and SP were 3 . 5 , 8 . 7 , and 126 , respectively . The SP strain showed a low level of resistance to organophosphates , with the RRs of larvae to temephos and pirimiphos methyl being 1 . 5- and 4 . 2-fold , respectively ( Table 3 ) . We genotyped five positions of Vssc , which potentially cause decreased sensitivity to pyrethroid insecticides . Sequencing of the partial DNA of Vssc identified three amino acid substitutions in the SPS0 population compared with the SMK strains: S989P , V1016G , and F1534C . Of these substitutions , P989 and G1016 always appeared together , indicating that these polymorphisms are located on the same haplotype . All mosquitoes in the SPS0 population possessed either the P989+G1016 or C1534 haplotypes , with a frequency of 44% and 56% , respectively . After the first selection , the frequency of P989+G1016 increased to 79% , and following the second selection , the C1534 haplotypes were no longer detected ( Table 4 ) . In SMK , all individuals were homozygous for wild-type at all five amino acid positions . The typical kdr mutation , L1014F , was not detected in any of the Singapore populations and would not be expected in this species because kdr in Aedes would require a 2 nucleotides change . We examined the rate of permethrin reduction from cuticle by measuring the disappearance of insecticide from the external surface of the mosquitoes ( Figures 2A and B ) . There was no difference in the rate of permethrin reduction from cuticle between the SP and SMK strains , rather slightly higher in SP strain , with penetration rate constants of 0 . 358 h−1 and 0 . 252 h−1 , respectively . Some previous studies suggested that synergistic effects of synergists are not due to inhibition of enzymes but due to enhancement of penetration rate of insecticides by synergists [52]–[54] . Therefore , we investigated the effects of PBO on permethrin penetration ( Figure 2A ) . Pretreatment with PBO in the SP strain did not enhance permethrin penetration but resulted in significantly delayed its penetration ( Figure 2B ) . Twelve hours after treatment , less than 20% of the insecticide had penetrated the cuticle in insects pretreated with PBO compared with 90% in insects without PBO treatment . A reduced rate of permethrin penetration was also observed even when we administered PBO to the thoracic sternum of mosquitoes . As shown in Figure 2B , pretreatment of the SP strain with acetone did not affect permethrin penetration . The marked synergistic effects of PBO on permethrin toxicity implicated P450s being involved in resistance . Therefore we further investigated trends of permethrin after penetration through the cuticle using isotopic tracer tests . It was observed that in the SMK strain , internal radioactivity reached a peak at approximately 12 h after permethrin treatment , and then gradually decreased ( Figure 2C ) . In contrast , in the SP strain , internal radioactivity reached a much earlier peak and was then rapidly eliminated . The percentage of the initial radioactivity that remained after 24 h in the SMK and SP strains was 59 . 5% and 14 . 8% , respectively . Unlike the groups without PBO pretreatment , the internal dose of radioactivity gradually increased in the groups that received PBO pretreatment and did not decrease until about 48 h after treatment . The internal radioactivity at 48 h with or without PBO in SP strain was 42 . 7% and 4 . 9% , respectively ( Figure 2C ) . Consistent with the above findings , we found that the SP strain excreted radioactivity more rapidly than the SMK strain ( Figure 2D ) . The percentage of the radioactivity that was excreted 24 h after treatment in the SMK and SP strains was 32 . 7% and 82 . 8% , respectively . This suggested that the SP strain can eliminate permethrin more efficiently . The rate constants of excretion ( hr−1 ) for SMK and SP were 0 . 022 ± 0 . 002 and 0 . 101 ± 0 . 008 , respectively . PBO significantly reduced permethrin excretion , suggesting that permethrin is excreted after being metabolized by P450s . The high performance thin layer chromatography ( HPTLC ) analysis of excrete from SP strain revealed that more than 85% of the excreted radioisotope consisted of high polar compounds and approximately 10% of the permethrin was excreted without being metabolized ( Figures 2E and F ) . Further investigation using another solvent specific for water-soluble metabolites showed that the compounds consisted of a number of metabolites ( Figure 2G ) . We carried out in vitro metabolism studies to determine whether the changes observed in vivo were related to the activity of P450s . Microsomes of the SP strain metabolized permethrin at a much higher rate than those of the SMK strain ( Figures 3A and B ) . Permethrin metabolism by microsomes was apparently due to P450 as it was NADPH-dependent and was inhibited by PBO ( Figure 3A ) . Only a very small amount of permethrin was metabolized by enzymes in the 100 , 000 xg supernatant , suggesting that carboxyl esterase have only minimal involvement in the development of resistance . The rate constant for metabolism ( min−1 ) of SMK and SP by microsomal enzymes was 0 . 0007 and 0 . 0042 , respectively . At the start of incubation , 4′-HO-permethrin was the major metabolite in SP , although its percentage did not increase . In contrast , the percentage of high polar metabolites increased over time ( Figure 3B , Table S2 ) . The metabolites located around the origin of the HPTLC plates were collected and further developed with chloroform-based solvent ( Figure 3C ) . The high-polar compounds consisted of a number of metabolites and were consistent with those in the in vivo study . In order to assess the effect of 4′-HO-permethrin on permethrin metabolism , we performed an inhibition assay by incubating different doses of unlabeled 4′-HO-permethrin with [14C]-permethrin . This showed that unlabeled 4′-HO-permethrin inhibited effective permethrin metabolism , possibly by acting as a feedback inhibitor ( Figure 3D ) . A microarray that contained 60- mer probes for all genes identified by the genome project [55] was constructed . We then compared the levels of transcription in the SP and SMK strains of the adult males and females and the fourth instar larvae . In this study , we focused on P450s and their related genes because our investigations including bioassays and in vivo and in vitro metabolism studies showed that this metabolic enzyme is the key factor in the development of resistance . The genes over expressed in the SP strain were selected using a cut-off of >3-fold relative change in expression and a P-value<0 . 01 ( Figure 4 and Table 5 ) . Nine P450 genes ( CYP9M6 , CYP9M5 , CYP6Z8 , CYP6Z7 , CYP9M4 , CYP6BB2 , CYP6F3 , CYP6F2 , and CYP4C50 ) were over expressed in both male and female insects . Out of nine , four ( CYP9M6 , CYP9M5 , CYP6Z7 , and CYP9M4 ) were over expressed in the larval sample as well . Cytochrome b5 was also over expressed in all three samples . CYP9M6 was the only P450 that had a >20-fold transcription level with a significantly low ( <0 . 01 ) P-value in all three samples . This gene had not been reported to be associated with insecticide resistance . Fifteen of the 22 P450 genes listed in Table 5 have previously been reported to be over expressed in one or more of pyrethroid-resistant strains/populations collected from Mexico [30] , [56] , Cuba [57] , Thailand [30] , Martinique , Bora Bora [58] , and the Grand Cayman [57] . Two of these , CYP9J26 and CYP9J28 have been previously shown to have the ability to metabolize permethrin [59] . Other than P450s , no glutathione transferases ( GSTs ) or carboxyl esterases , except two GSTs , was over expressed in SP ( under a cut off of >3-fold , P<0 . 01 ) . GSTD4 ( AAEL001054 ) was 11 . 9 ( P = 1 . 6 × 106 ) and 7 . 9 ( P = 0 . 0025 ) times more expressed in SP of female and male , respectively , compared to SMK . GSTD5 ( AAEL001071 ) in SP females was 5 . 5 ( P = 0 . 0020 ) times more expressed than in females of SMK . In order to design primers for real time quantitative PCR , we used eight mosquitoes individually to obtain the sequence of partial or full length DNA for seven P450 genes ( CYP6BB2 , CYP6Z7 , CYP6Z8 , CYP9M4 , CYP9M5 , CYP9M6 and CYP9M7 ) and an internal control gene , ribosomal protein S3 ( RPS3 ) . We focused on these P450 genes according to the following criteria: overexpressed in females ( >5 ) and in males ( >3 ) in SP relative to SMK in microarray analysis . CYP9M7 was also analyzed as it forms a gene cluster with CYP9M4 , CYP9M5 , and CYP9M6 within 65 kbp on the same supercontig ( 1 . 29 ) and is structurally related to these genes ( Figure 5F ) . The full length cDNA and deduced amino acid sequences of CYP6BB2 were identical between the SP and SMK strains . However , on the ninth nucleotide after the stop codon , we observed a C1533T replacement in the SP strain ( Figure S3 ) . All eight SP mosquitoes possessed two CYP9M6 genes , designated as CYP9M6v1 ( accession number AB840269 ) and CYP9M6v2 ( accession number AB840270 ) ( Figures S3 and S4 ) . The nucleotide and amino acid identities of these alleles were 97 . 6% ( 1563/1602 ) and 98 . 9% ( 527/533 ) , respectively . In the SMK strain , we found three CYP9M6 variants; two mosquitoes were homozygous for CYP9M6v3 ( accession number AB846835 ) or CYP9M6v4 ( accession number AB846836 ) , two were homozygous for CYP9M6v5 ( accession number AB846837 ) , and four were heterozygous CYP9M6 variants . The CYP9M6v5 protein appeared to be incomplete as an active enzyme because there was a stop codon at amino acid position 473 ( Figure S4 ) . We performed real-time quantitative PCR ( qPCR ) for seven P450 genes to verify the results of the microarray analyses . We focused these genes according to the criteria described above . Primers for quantitative real time PCR were designed from sequences that were consensus among eight SP and SMK mosquitoes . Five of these genes , CYP6BB2 , CYP6Z7 , CYP9M4 , CYP9M5 , and CYP9M6 were over expressed during the larval stage and in adult males and females ( Figures 5A–C ) in SP relative to SMK . As shown in Figure 5D there was a good correlation between the results of the microarrays and qPCR ( R2 = 0 . 824 ) . This confirmed that the results of the microarray were accurate . CYP9M6 showed the largest relative change in the seven genes and was over expressed 22 . 9-fold in males , 38 . 9-fold in females , and 28 . 2-fold in larvae of the SP strain relative to SMK . The relative expression level of CYP6BB2 to RPS3 was also higher ( close to CYP9M6 ) , especially in adult females ( Figure 5A ) , with the gene reported to be over expressed in five other pyrethroid-resistant strains ( Table 5 ) . The gene copy numbers of the seven P450s in the SP and SMK strains were compared by qPCR ( Figure 5E ) . RPS3 gene was used to normalize the data . While the majority of the P450 genes in the SMK strain were likely to be a single copy , some genes in SP were amplified significantly . The average copy number of CYP6Z7 , CYP9M4 , CYP9M5 , and CYP9M6 in SP were 3 . 1- , 4 . 3- , 4 . 8- , and 4 . 6-fold more than SMK , respectively . This clearly shows that the over expression of these P450 genes , at least in part , was due to gene amplification . In order to determine whether CYP9M6 and CYP6BB2 have a role in permethrin metabolism , F2 progeny ( interbred population of F1 ( SMK × SP ) ) were treated with a sub-lethal dose of [14C]-permethrin , and the excretion rate was measured 24 h after treatment in individuals . The genotypes and transcription levels of CYP9M6 and CYP6BB2 were determined for each mosquito to examine the association between these parameters and the rate of excretion ( Figures 6A and D ) . The transcription levels for CYP9M6 between RR ( homozygote of SP allele ) and SS ( homozygote of SMK allele ) were significantly different ( P<0 . 0001 ) , with a 45-fold difference between the two genotypes ( Figure 6B ) . The rate of permethrin excretion was 69% for RR and 45% for SS ( Figure 6C ) . On the other hand , although the transcription level of CYP6BB2 was significantly different between RR and SS ( P = 0 . 0298 ) , the relative change ( RR/SS ) was only 1 . 6-fold ( Figure 6E ) . The rate of permethrin excretion in RR ( 71% ) was significant than that in SS ( 49% ) ( P<0 . 0001 , Figure 6F ) . These results suggested that both CYP9M6 and CYP6BB2 were associated with rapid excretion of permethrin , although it was likely that the metabolism rate of permethrin was different between the two isozymes . In addition to CYP6BB2 and CYP9M6 , the transcription levels of other four P450 genes ( CYP6Z7 , CYP6Z8 , CYP9M4 , and CYP9M5 ) were quantified in 48 F2 progeny in order to determine their association with the rate of permethrin excretion ( Figure 7 ) . We observed a relatively high correlation between the rate of permethrin excretion and mRNA levels for CYP6BB2 and CYP9M6 , with correlation coefficients ( R2 ) of 0 . 454 and 0 . 485 , respectively ( Figures 7A and F ) . The R2 for CYP9M5 was also high ( 0 . 537 , Figure 7E ) . Of the six P450 genes , CYP6Z7 had the lowest R2 ( 0 . 236 , Figure 7B ) . Forty-eight individuals were ranked from 1 to 48 according to the level of expression of each P450 gene . The ranks of each individual against multiple P450 genes were combined to standardize the level of mRNA and we evaluated the association between permethrin excretion rate and expression level of multiple P450 genes . The standardized transcription levels of CYP9M6+CYP6BB2 had a considerably higher R2 ( 0 . 569 ) than individual values alone ( Figure 7G ) . This strongly suggests that both these P450s contribute to permethrin excretion . The combination of CYP9M6 , CYP6BB2 , and CYP9M5 further increased the R2 to 0 . 589 ( Figure 7H ) , whereas the combination of all six P450s decreased the R2 to 0 . 499 ( Figure 7I ) . These findings imply that not all of these P450 were involved in permethrin excretion . In order to determine whether CYP9M6 and CYP6BB2 are capable of permethrin metabolism , these P450s were co-expressed with A . aegypti cytochrome P450 reductase and b5 in Sf9 cells using a baculovirus . For CYP9M6 , two genes , CYP9M6v1 and CYP9M6v2 were expressed , as sequencing analysis showed that all individuals in the SP strain possessed both types of genes . Microsomes were prepared from cells infected with baculovirus , followed by an in vitro metabolism study using [14C]-permethrin as the substrate . Both CYP9M6v1 and CYP9M6v2 demonstrated relatively low but consistent permethrin metabolism , whereas CYP6BB2 exhibited strong metabolic activity for permethrin ( Figure 8 , Table S3 ) . The compound , 4′-HO-permethrin , detected in the microsome metabolism studies , was also confirmed as a major metabolite for the three P450 samples ( CYP9M6v1 , CYP9M6v2 , and CYP6BB2 ) . This suggested that these three P450s can detoxify permethrin . In addition , a large amount of polar metabolites ( E ) were detected by in vivo and in vitro studies on all three samples . Unique metabolites ( metabolites A and B ) , which were almost undetectable in the in vitro microsomal study were also detected in samples of CYP9M6v1 and CYP9M6v2 ( Figure 8 , Table S3 ) .
We selected a Singapore colony of A . aegypti ( SPS0 ) in the laboratory and established a SP ( = SPS10 ) strain , which developed an extremely high level of resistance to permethrin . Using bioassays with a synergist and in vivo and in vitro studies , we confirmed that P450s play very important roles in the development of resistance . The larvae of the SP strain also developed a high level of resistance despite the selections being conducted on adult mosquitoes . This indicates that common mechanism ( s ) causes resistance in both developmental stages . This finding is in contrast to those reported in the JPal-per strain ( Culex quinquefasciatus ) , with these insects being selected by permethrin in the larval stage and not showing a high level of resistance in the adult stage [60] . The JPal-per strain has a knockdown resistance gene allele ( L1014F ) as well plus high expression levels of CYP9M10 , known to confer resistance [61]–[63] , being approximately 250-fold higher than in the larvae of susceptible strain , and showing limited expression in the adult stage [37] . In the SP strain , CYP9M6 , CYP6BB2 , and CYP9J26 were over expressed in both the larval and adult stages and probably these were involved in the larval resistance as well . The amino acid substitutions of Vssc , V1016G and F1534C , have been shown to strongly correlate with pyrethroid resistance in A . aegypti [17] , [23]–[27] . The Singapore population ( SPS0 ) possessed both G1016 and C1534 , with a gene frequency of 44% and 56% , respectively . However , all C1534 haplotypes were disappeared after only two selections by permethrin . This clearly indicates that Vssc with a G1016 mutation has a lower sensitivity to permethrin than channel with a C1534 mutation . This finding is consistent with that reported by a recent neurophysiological study [27] . The results of our bioassays further supported this phenomenon , as we showed 3–5-fold differences in resistance ratios between SP and SPS0 insects when they were treated permethrin with PBO in the adult ( Table 2 ) and also larval stages ( Table 3 ) . Changes of Vssc genotype is also associated with development of resistance in the larval stage in SP . The reason for relatively high frequency of C1534 mutation in this country is still unclear . Further studies , including the examination of fitness costs between G1016 and C1534 , will help us better understand factors affect the equilibrium state of these two haplotypes in nature . In the bioassays , PBO exhibited a marked synergistic effect on the toxicity of permethrin in the SP strain ( Table 2 and 3 ) . On the other hand , several research groups stated that synergists may enhance the toxicity of insecticides because they accelerate the penetration of these agents [52]–[54] . We therefore examined the effects of PBO on permethrin penetration and showed that it did not enhance permethrin penetration , but instead suppressed the rate of its penetration ( Figure 2 ) . This finding is consistent with other reports on the German cockroach [64] and house fly [65]–[67] . We also found that permethrin penetration was markedly suppressed by PBO treatment , even administered on the thoracic sternum of the insect followed by permethrin treatment on the thoracic notum . This result implies that PBO was not acting as a physical barrier to permethrin penetration but rather another mechanism . We speculate that high concentration of PBO inside the mosquito body suppressed permethrin passing through the cuticle by the effect of a concentration gradient of chemicals . Therefore , caution must be exercised when we use a synergist in a bioassay to avoid underestimating the contribution of metabolic enzymes to the resistance . 4′-HO-permethrin has been well documented as a primary metabolite of permethrin in various insects including the house fly , southern house mosquito , cockroach , and cabbage looper [42] , [44] , [68] . We also showed that this compound was a major metabolite of permethrin in the SP strain in vitro and comprised approximately 40% of the metabolites after an incubation of 5 min ( Figure 3B ) . However , the rate of 4′-HO-permethrin to the total metabolites decreased over time and comprised approximately 10% after 6 h . On the other hand , the rate of high polar compounds increased over time and made up 86% of the total metabolites after 6 h . Further , 4′-HO-permethrin was barely detectable in the metabolites excreted by mosquitoes in vivo ( Figures 2E and F ) . This suggested that 4′-HO-permethrin was further metabolized to secondary metabolites that were more water soluble and excretable ( Figure 9 ) . We also found that 4′-HO-permethrin inhibited permethrin metabolism in vitro ( Figure 3D ) , indicating the presence of a negative feedback regulation ( Figure 9 ) . Insects need to convert hydroxyl-permethrin to other secondary metabolites in order to avert this negative feedback , resulting in insufficient excretion . Secondary development in the HPTLC analysis showed that the high polar metabolite consisted of a number of various compounds ( Figures 2G and 3C ) . These compounds were suspected of being conjugates of 4′-HO-permethrin with glucosides or/and with various amino acids including glycine , glutamic acid , glutamine , and serine , as reported by Shono et al . [68] . It remains unknown what enzymes play the role in the secondary metabolism of pyrethroid in insects , and it is possible that an obscure mechanism of insecticide resistance may exist . Association of gene amplification with insecticide resistance is relatively well documented [69] . In organophosphorus-resistant aphids and Culex mosquitoes , genes of carboxyl esterases are amplified 80–250 times and the capability to detoxify insecticides is increased [70]–[72] . Recent studies demonstrated that gene amplification is also associated with over expression of P450 genes in insects [40] , [62] , [63] , [73] , [74] . In A . aegypti , over expression of CYP9J29 in the CUBA-DELTA and CAYMAN strains was shown to be due to gene amplification [57] . In the current study , we found that CYP9M6 had the capability to metabolize permethrin and was over expressed in the SP strain partially due to gene amplification . Real time quantitative PCR revealed that this gene was amplified approximately 4-fold in the SP strain compared with the SMK strain . We identified two CYP9M6 genes , CYP9M6v1 and CYP9M6v2 , which were probably raised by the gene amplification event in the SP strain . In particular , two other similar genes , CYP9M4 and CYP9M5 , located within the same cluster on chromosome two [75] were also amplified to a similar extent ( Figure 5E ) . This suggested that this amplification may have occurred coincidently . Our microarray analysis revealed that 22 P450 genes were over expressed >3-fold in all samples ( males , females , or larvae ) of the resistant SP strain compared with the susceptible SMK strain ( Table 5 ) . It is noteworthy that 15 of these genes have been reported previously to be overexpressed in one or more resistant populations or strains collected from different regions of the world [28] , [30] , [56] , [57] . In this study , we found that CYP9M6 and CYP6BB2 were responsible for resistance . In addition , CYP9J26 and CYP9J28 , which have the ability to metabolize permethrin [59] , [76] , were also over expressed in the SP strain . It remains to be determined as to how many P450s contribute to the resistance in SP mosquitoes . However , it is not likely that all genes over expressed in the SP strain are involved in the development of resistance , as it is generally known that transcription of genes forming clusters are occasionally coordinately controlled by the same regulatory factor [77] . In association analysis using F2 progeny , the expression level of CYP9M5 and CYP9M6 showed a high correlation coefficient ( R2 = 0 . 90 ) , whereas no correlation was observed between the expression levels of CYP9M6 and CYP6BB2 ( R2 = 0 . 10 , data not shown ) . This suggested that CYP9M5 and CYP9M6 are controlled partially by a common mechanism . A further example of this mechanism is the finding that administration of phenobarbital to insects induces a multiple number of P450 genes [78] , [79] . Therefore , over expressed genes do not always provide an advantage for insect survival . Heterologous expression and confirmation of the metabolic activity of each P450 isoform is necessary to gain a greater understanding of these processes . To the best of our knowledge , this is the first report showing a correlation between the expression level of each P450 and the rate of insecticide metabolism of individual insects . This analysis may provide a good tool for evaluating the rate of contribution of each P450 isoform to the development of resistance . In this analysis , both CYP9M6 and CYP6BB2 showed moderate correlation between the levels of transcription and the individual rate of permethrin metabolism . Furthermore , although no correlation was observed between the expression level of CYP9M6 and CYP6BB2 ( R2 = 0 . 10 ) , there was a prominent relationship between the rate of permethrin metabolism and standardized-combined expression level of CYP9M6 and CYP6BB2 ( Figure 7G ) . These results strongly suggest that these two P450s are auxiliary parts of effective permethrin excretion . Although CYP9M6 showed the greatest relative change in gene expression in the comparison between the SP and susceptible strains , we still hesitate to conclude that this enzyme plays a major role in permethrin metabolism . First , homozygotes of CYP6BB2 had a rate of permethrin excretion as high as that observed in homozygotes of CYP9M6 in Figures 6C and F . Second , in the heterologous expression study , under our experimental conditions , CYP6BB2 showed a much higher activity of permethrin metabolism than CYP9M6 . It is therefore possible that CYP9M6 may compensate for its low metabolic activity by its high mRNA level . Although both CYP9M4 and CYP9M5 were also over expressed in the SP strain , we did not investigate the permethrin metabolic activity of these P450s because their relative expression level was below 1/10th of that of CYP9M6 . However , it is possible they may confer resistance depending on their unit activity of permethrin metabolism . Further work will be needed to clarify if these P450 isoforms are involved in the resistance . One of the goals of the study was to establish a more accurate molecular diagnosis method for monitoring field populations of A . aegypti by identifying metabolic enzymes that confer resistance . Findings of CYP9M6 and CYP6BB2 are good progress for this purpose , however , we also found that the high level of resistance in the SP strain was not due to one major mechanism but was actually the consequence of multiple P450 isozymes ( at least 4 ) and reduced sensitivity of Vssc . This implies that we may need to recognize the reality that the development of molecular diagnoses targeting metabolic enzymes will be more complicated and challenging than that of targeting only Vssc [80] . We are currently focusing on elucidating the contribution degree of these P450s on pyrethroid resistance in the field populations of A . aegypti collected from different regions . | Aedes aegypti inhabits tropical and subtropical regions worldwide and is the major vector of dengue and yellow fevers , and a secondary vector of chikungunya fever . Dengue fever is epidemic in more than 110 countries and causes up to 100 million infections annually . As there is no efficient vaccine or medicine currently available , vector control remains the primary solution for reducing the number of cases of this disease and relies heavily on the use of insecticides . Intensive and long-term use of insecticides has resulted in the worldwide emergence of mosquitoes with resistance to these chemicals . Here we newly identified two P450s , which have the ability to detoxify pyrethroid insecticide and are over produced in the resistant A . aegypti strain . Our study showed there were at least four P450 isozymes associated with resistance and target site insensitivity . These findings may lead to the development of more accurate monitoring systems in the field and also assist to identify new target sites for insecticides that are effective against resistant insects . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"biochemistry",
"entomology",
"enzyme",
"metabolism",
"enzymes",
"biology",
"and",
"life",
"sciences",
"enzymology",
"zoology",
"enzyme",
"chemistry",
"genetic",
"toxicology",
"toxicology",
"neurotoxicology"
] | 2014 | Mechanisms of Pyrethroid Resistance in the Dengue Mosquito Vector, Aedes aegypti: Target Site Insensitivity, Penetration, and Metabolism |
Wing pattern evolution in Heliconius butterflies provides some of the most striking examples of adaptation by natural selection . The genes controlling pattern variation are classic examples of Mendelian loci of large effect , where allelic variation causes large and discrete phenotypic changes and is responsible for both convergent and highly divergent wing pattern evolution across the genus . We characterize nucleotide variation , genotype-by-phenotype associations , linkage disequilibrium ( LD ) , and candidate gene expression patterns across two unlinked genomic intervals that control yellow and red wing pattern variation among mimetic forms of Heliconius erato . Despite very strong natural selection on color pattern , we see neither a strong reduction in genetic diversity nor evidence for extended LD across either patterning interval . This observation highlights the extent that recombination can erase the signature of selection in natural populations and is consistent with the hypothesis that either the adaptive radiation or the alleles controlling it are quite old . However , across both patterning intervals we identified SNPs clustered in several coding regions that were strongly associated with color pattern phenotype . Interestingly , coding regions with associated SNPs were widely separated , suggesting that color pattern alleles may be composed of multiple functional sites , conforming to previous descriptions of these loci as “supergenes . ” Examination of gene expression levels of genes flanking these regions in both H . erato and its co-mimic , H . melpomene , implicate a gene with high sequence similarity to a kinesin as playing a key role in modulating pattern and provides convincing evidence for parallel changes in gene regulation across co-mimetic lineages . The complex genetic architecture at these color pattern loci stands in marked contrast to the single casual mutations often identified in genetic studies of adaptation , but may be more indicative of the type of genetic changes responsible for much of the adaptive variation found in natural populations .
Understanding how adaptive phenotypes arise is vital for understanding the origins of biodiversity and for predicting how organisms will respond to novel selective pressures [1] . Nonetheless , there are still only a handful of examples where the molecular elements underlying adaptive variation in nature have been identified [2]–[6] . This situation is changing as new technologies make it possible to leverage nature's diversity and focus research directly on taxa that are both ecologically tractable and possess characteristics ( life history switches , behavioral modifications , or phenotypic differences ) with a priori evidence of their adaptive role [7]–[10] . The data that will emerge from these studies promise fresh insights into the genetic architecture and origins of functional variation and an exciting new understanding of the interplay between genes , development , and natural selection . Heliconius butterflies offer exceptional opportunities for genomic level studies designed to understand how adaptive morphological diversity is generated in nature [11]–[13] . The group is renowned as one of the great insect radiations and provides textbook examples of adaptation by natural selection , mimicry , and speciation [14] , [15] . The vivid wing patterns of Heliconius are adaptations that warn potential predators of the butterflies' unpalatability and also play a key role in speciation [16]–[18] . Perhaps the greatest strength of Heliconius for understanding the origins of functional variation lies is the wealth of parallel and convergent adaptation in the group- a pattern best exemplified by the parallel mimetic radiations of H . erato and H . melpomene [19]–[23] . The two species are distantly related and never hybridize [24] , [25]; yet , they possess nearly identical wing patterns and have undergone nearly perfectly congruent radiations into over 25 distinctively different color pattern races [21] . The convergent and divergent color pattern changes within and between Heliconius species provide “natural” replicates of the evolutionary process where independent lineages have produced similar phenotypes due to natural selection . Indeed , within both the H . erato and H . melpomene radiations , there are multiple disjunct populations that share identical , yet possibly independently evolved , wing patterns [26] , [27] ( for an alternative , shifting balance view , see [22] , [28] ) . Moreover , recent comparative research has demonstrated that the diversity of color patterns found within H . erato , H . melpomene and in other Heliconius species , is modulated by a small number of apparently homologous genomic intervals [29]–[31] , which provides a powerful evolutionary framework for examining the origins of functional variation and allows insights into the repeatability of evolution . The patchwork of differently patterned races in H . erato and H . melpomene is stitched together by dozens of narrow hybrid zones [20]–[22] , allowing detailed analysis of the forces that generate and maintain adaptive variation in this group [32] . Here , and in our companion paper [33] , we exploit concordant hybrid zones to explore patterns of nucleotide diversity and linkage disequilibrium ( LD ) across two of the three interacting genomic regions that control most of the adaptive differences in wing color patterns . The transition between the “postman” , H . e . favorinus and H . m . amaryllis , and “rayed” , H . e . emma and H . m . agalope , races of the two co-mimics in eastern Peru is one of the best described hybrid zones in Heliconius and occurs over a distance of slightly more than 10 km ( Figure 1 and [33] ) . Strong natural selection maintains this sharp phenotypic boundary in both species and per locus selective coefficients on color pattern loci are estimated to be greater than 0 . 2 both using field release experiments and by fitting the observed cline in allelic frequencies at each of the color pattern loci to a theoretical cline maintained by frequency dependent selection [34] , [35] . Despite strong natural selection , there are no strong pre- or post-mating barriers to hybridization between races of either H . erato or H . melpomene and in the center of the hybrid zone there is frequent admixture between divergent color pattern races . Our study focuses on two H . erato patterning loci , D and Cr . These two loci map to different linkage groups and interact to control major differences in the wing color patterns of H . erato races . The chromosomal regions tightly linked with the D and Cr loci in H . erato were recently identified [36]–[38] and map to homologous regions of the genome that control similar color pattern changes in H . erato's co-mimic , H . melpomene [29] , [31] . Variation in D in H . erato and D/B in H . melpomene cause analogous changes in the distribution of red pigments on the fore- and hindwings ( see[30] , [31] , [39] ) . Similarly , Cr ( H . erato ) and the Yb-complex ( H . melpomene ) cause similar shifts in the distribution of melanic scales revealing underlying white and yellow pattern elements ( see [29] ) . This region also contains the H . numata P locus , a close relative of H . melpomene . However , the P locus causes dramatically different pattern changes among sympatric races of H . numata highlighting the extraordinary ‘jack-of-all-trades’ flexibility of these genomic regions [29] . Wing pattern variation across Heliconius hybrid zones serves as a “natural” laboratory for genome level research into processes that generate and maintain adaptive variation . One of the most extensively studied Heliconius hybrid zones is found in Eastern Peru , where Mallet and coworkers estimated the strength of natural selection on the three unlinked color pattern loci that control phenotypic differences between “rayed” and “postman” races of H . erato [34] , [35] , [40] . We have taken the next step and used this same Peruvian H . erato hybrid zone to make four major advancements: ( 1 ) we have identified and sequenced narrow genomic intervals containing two of the three interacting loci that cause major adaptive shifts in wing patterns; ( 2 ) we have documented a rapid decay of LD in natural populations across a sharp phenotypic transition both within genes and across these intervals; ( 3 ) we have identified several genes strongly associated with the transition in warningly-colored wing patterns; and ( 4 ) we have examined expression levels in these and adjacent genes during wing development . These data , in combination with data presented in the companion paper [33] , refine our understanding of the molecular nature of color pattern loci and suggest that multiple functional sites underlie adaptive morphological variation in Heliconius .
Building on earlier work , including the initial BAC tile path of H . melpomene D/B locus [31] , we sequenced 10 H . erato BACs representing over 1 Mb of genomic sequence around the D locus ( Figure 2 ) . Across the D BAC tile path , we surveyed over 1200 individuals from our H . erato x H . himera F2 and backcross mapping families at several molecular markers , and identified an approximately 380 kb interval between the markers Gn12 and THAP that had no recombination events between color pattern phenotype and genotype ( shaded region on Figure 2 ) . The lack of recombinants across this zero recombinant window stood in marked contrast to the pattern observed at both the 5′- and 3′-end of our tile path . At both ends of the region , the number of individuals showing a recombinant event between a genetic marker and color pattern phenotype was similar to the expected 276 kb/cM based on previous mapping work [39] , but then dropped off rapidly in the centre of the region . The drop off was particularly marked on the 5′end of the interval , where the number of recombinant events fell from 35 individuals at GN47 to 0 individuals at Dna-J over a span of approximately 200 kb . We also identified the genomic interval containing the Cr locus , although in this case , we do not yet have a BAC tile path across the entire interval . The 5′-end of Cr interval is marked by the locus GerTra , where we identified a single recombinant among nearly 500 H . erato cyrbia x H . himera F2 and backcross individuals . At the 3′-end , we observed 3 Cr recombinants at HEAT , which is about 600 kb from GerTra based on comparisons to the Bombyx mori genome ( Figure 2 ) . We sequenced three new BAC clones yielding approximately 420kb of sequence at the 5′-end of the Cr interval . Across our physical sequence of the Cr interval , we found no recombinant individuals at markers 3′ of GerTra ( B9 , recQ , Invertase , LRR , and GN 71 ) a span of approximately 340 kb ( Figure 2 ) . Thus , as with the D locus interval , there were fewer recombination events than expected based on previous estimates of the relationship between physical and recombination distance . We estimated genetic diversity from 76 individuals collected from five locations along a 30 km transect , representing three distinct populations , phenotypically pure H . e . favorinus ( n = 20 ) , admixed individuals ( n = 42 ) , and largely pure H . e . emma ( n = 14 ) ( Figure 1 ) . In total , we assayed variation across 12 , 660 bp from 25 coding regions including 13 regions from the D interval , 10 from the Cr interval , and 3 unlinked to each other or any color pattern locus in H . erato ( Table S1 ) . There were 1542 polymorphic sites among the sampled individuals . Most of these ( 1110 ) positions had minor allele frequencies of less than 5 percent . Of the remaining 432 polymorphic sites , ten had more than two variant bases . The mean nucleotide diversity ( π , average number of pair-wise differences between sequences ) among all sampled gene regions in H . erato was 0 . 022±0 . 017 . In general , there were no strong differences in nucleotide diversity among loci tightly linked to color pattern genes relative to loci unlinked to color pattern ( Table 1 ) . Nucleotide diversity was also very similar among the three sampled H . erato populations , except for a few gene regions at the Cr locus in the admixed population that showed slightly elevated estimates of nucleotide diversity ( Table 1 ) . Over half the coding regions sampled in this study had patterns of nucleotide diversity not consistent with simulations of neutral evolution , in at least one of the three populations sampled . Near the D locus , many coding regions had negative Tajima's D values that were significantly different from neutral expectation ( Table 1 ) . However , there seemed to be little pattern to these departures from neutrality . For example , the coding regions at the D locus most strongly associated with color pattern variation ( see below ) all showed patterns consistent with the neutral model . In contrast , at the Cr locus , the two coding regions with associated SNPs accounted for about half of the significant deviations from neutrality in genes across this region ( Table S1 ) . We also observed significant deviations from neutrality at loci unlinked to color pattern variation . In particular , the Heliconius wingless homologue deviated in all three populations examined ( Table S1 ) . Overall nucleotide diversity was generally greater in the H . erato ( mean π = 0 . 022±0 . 017 ) than in H . melpomene ( mean π = 0 . 012±0 . 019 , [33] ) but the differences were much less than previously reported for nuclear introns [27] . Moreover , in H . melpomene , as in H . erato , there were no striking differences in diversity between loci within and outside of color pattern intervals , nor consistent departures from neutrality within color pattern intervals . Linkage disequilibrium among SNPs decayed precipitously with physical distance across both the D and Cr intervals ( Figure 3 and Figure S2 ) . This observation was true for phenotypically pure populations collected at either side of the sharp phenotypic transition ( Figure S1 ) , for “admixed” populations in the center of the transition zone ( Figure S1 ) , and even for the population as a whole ( Figure 3 ) . The only sites with high estimates of r2 ( >0 . 5 ) were found within the same coding regions . All other estimates of r2 were near zero ( Figure 3 ) , including values between D and Cr interval SNPs ( Figure 3 ) . The lack of strong LD in populations across this phenotypic boundary was perhaps best exemplified by the LD patterns within loci - for all loci , including those that fell within our zero recombinant windows , short-range LD decayed to r2 values near zero within 300–500 bp . Although broadly similar , the pattern of LD differed from what was observed in H . melpomene ( see [33] Figure S2 ) , where LD generally extended farther and there was some evidence for significant haplotype structure and long-distance LD among sites . None of the SNPs in this study had a fixed association with color pattern , suggesting that , while the site is strongly associated with color pattern , they are not the functional variants themselves . However , the obvious implication is that they are near the functional site , which could be in cis-regulatory regions that act by causing differences in gene expression . To test this possibility , we compared overall transcription levels between the two races during the early stages of wing development ( 5th larval instar and 1 , 3 , and 5 days after pupation ) , on genes at the D locus that had SNPs strongly associated with wing pattern phenotype either in H . erato or H . melpomene [33] . All genes , with the exception of Slu7 , showed significant differences in expression across wing developmental stages ( ANOVA: p<0 . 0001 to 0 . 0066; Bayesian Model Averaging: Pr ( β ≠ 0 ) = 100 for each gene ) ( Figure 5 ) . Kinesin , however , was the only candidate gene to show significant differences in expression between H . e . emma and H . e . favorinus ( overall race effect p = 0 . 0001 ) . Expression of this gene was roughly 8× higher in H . e . emma in 5th instar larvae ( p = 0 . 0028 , t-test ) and three days after pupation ( p = 0 . 0014 , t-test ) , than in H . e . favorinus . As with the ANOVA , statistical testing using Bayesian Model Averaging assigned strong probabilities to racial differences only with Kinesin ( Pr ( β≠0 ) >92 . 5 ) , although a small race effect is predicted for GPCR ( Pr ( β≠0 ) >54 . 7; higher in H . e . favorinus ) .
The genetic patterns that we observed are inconsistent with the evolution of novel wing patterns in H . erato via a very recent strong selective sweep on a new mutation or recent genetic bottleneck as have been proposed [41] . A selective sweep on a new adaptive variant , which quickly fixes beneficial alleles , is expected to generate a temporary genomic signature marked by a reduction of nucleotide variation and an increase in LD around selected sites as a result of genetic hitchhiking [42] . Empirically , these patterns have been observed around loci important in domestication ( e . g . rice [43] and dogs [44] , [45] ) , plant cultivation ( sunflowers [46] and maize [47] ) , drug resistance ( Plasmodium , [48] ) , and the colonization of new environments in the last 10 , 000 years ( sticklebacks , [49]–[51] ) . In all cases , selection has been strong , directional , and very recent . The genetic patterns across regions responsible for phenotypic variation in H . erato and H . melpomene serves as a cautionary note and may be more typical of the functional variation found in nature . In H . erato , per locus selection coefficients are high [34] , [35]; yet , we see neither a strong reduction in genetic diversity nor extended LD across color pattern intervals . There are loci with nucleotide diversity patterns that deviate significantly from the neutral expectations , but not in a manner consistent with a recent , strong selective sweep acting on a new mutation . In all three loci in the D interval with the strongest association with color pattern , the patterns of nucleotide variation were largely consistent with neutrality ( Table 1 ) . Thus , recombination has essentially reduced the signature of selection to very narrow regions tightly linked to the sites controlling the adaptive color pattern variation . This pattern is consistent with the hypothesis that pattern diversification in H . erato is quite ancient , dating perhaps into the Pliocene ( see [27] ) . Interestingly , we see a very similar pattern in H . melpomene , which likely radiated much more recently [27] . Alternatively , the patterns in both H . erato and H . melpomene could also be the result of a recent “soft sweep” , where selection acts on pre-existing variation [52] , [53] . Thus , the allelic variants modulating particular color pattern elements are themselves old but the combination of patterning loci that characterize specific wing pattern phenotypes might have evolved much more recently [54] , [55] . Under either scenario , however , the observed patterns in both H . erato and H . melpomene highlight the extent with which recombination can erase the signature of strong selection in natural populations [56] . The rapid decay of LD across both H . erato color pattern intervals marks a history of considerable recombination . Narrow hybrid zones between differently adapted populations are common in nature [32] . Hybrid individuals are frequently less fit than parental genotypes and these zones are typically envisioned as “population sinks” that are maintained by the movements of individuals from outside [32] , [57] , [58] . As a result , hybrid zones tend to show LD even among unlinked loci [59]–[62] . Instead of a population sink , the narrow transition zone between H . e . favorinus and H . e . emma can be more appropriately viewed as a population sieve- where population sizes have remained large , where recombination breaks down associations among alleles even at tightly linked loci , and gene flow allows most of the genome to be freely exchanged between the divergent races . Mallet observed similarly low population differentiation across this cline at 14 unlinked allozyme loci ( average FST = 0 . 038 , unpublished data ) . Indeed , there is very little evidence for extended LD around loci that are responsible for adaptive differences in wing pattern and only slight genetic divergence between H . e . emma and H . e . favorinus at most of 25 coding regions examined within the two color pattern intervals ( Figure 4 ) . The only exceptions are regions tightly linked to the sites controlling the color variation , and , even here , LD decays rapidly with physical distance and estimates of FST become moderate , albeit higher than at unlinked loci ( see Table 2 and Table S2 ) . The decay in LD in H . erato occurs faster than in H . melpomene , where there is evidence for strong LD ( r2>>0 . 5 ) extending hundreds of kilobases across the B and Yb color pattern intervals [33] . Nonetheless , in both co-mimics , LD decays much more rapidly than has been reported near adaptive loci in sympatric host races of the pea aphid , Acrythosiphon pisum [63] and sympatric populations of lake whitefish , Coregonus sp . [64] . In the pea aphid study in particular , Via & West [63] showed that strong LD and strong genetic differentiation around the genomic intervals that underlie variation in ecologically important traits extends tens of centimorgans , presumably representing several Mbp at least . This is probably due to lower rates of cross-mating and geneflow , coupled with the largely non-recombinogenic reproduction in the pea aphid throughout most of the year . Our results are more similar to those found between M and S forms of Anopheles gambiae , where a few tens of genes around the centromeres and telomeres are the only regions with strong divergence [65] . Although in this case , the evolutionary or ecological forces driving these differences are largely unknown . The observation that LD in H . erato populations around ecologically important traits decays at a rate more similar to Drosophila than pea aphids or mammals [63] , [66]–[68] has important practical ramifications . Foremost , it means that naturally occurring Heliconius hybrid zones can be used to localize genomic regions responsible for adaptive differences in wing coloration at an extremely fine scale . On average there are informative polymorphic sites ( with a minor allele frequency greater than 5% ) every 30 bp within coding regions in our data on H . erato . Given this , along with the observed pattern of LD , surveying polymorphism every 200–500 bp should be sufficient to capture haplotype structure across the Peruvian hybrid zone and to finely localize genomic regions responsible for pattern variation in H . erato . Even with our current coarse sampling , we were able to greatly narrow the candidate D and Cr intervals and focus research on a small set of candidate genes . Across the D interval , there are strong genotype-by-phenotype associations and high levels of genetic differentiation between phenotypically pure populations in three dispersed coding regions: Dna-J , GPCR , and VanGogh . Although several genes near these association peaks have strong sequence similarity to Drosophila genes with known molecular or biological functions that make them candidates for color pattern genes , only one , kinesin , showed strong expression differences between H . e . emma and H . e . favorinus ( Figure 5 ) during early wing development . Kinesin proteins are known to play a role in pattern specification at a cellular level in Drosophila [69] and are involved in vertebrate [70] and invertebrate pigmentation [71] . We expect patterning loci to act as “switches” between different morphological trajectories and for the genes involved to show distinctive spatial/temporal shifts in expression patterns similar to what we observed in kinesin . Although future expression and functional validation is needed , we observed similar expression shifts in the H . melpomene kinesin [33] , which further implicates this gene as playing a causal role in pattern variation in Heliconius and provides convincing evidence that parallel changes in gene regulation underlies the independent origins of these co-mimetic lineages . Across the Cr interval , the two coding regions with the strongest associations , consist of a gene with strong homology to the Drosophila gene Unkempt , and another predicted gene with a leucine-rich repeat ( LRR ) . These loci are approximately 80 kb apart . The H . erato Unkempt codes for the type of protein potentially involved in pattern generation . It contains a zinc-finger binding motif and is potentially a signaling molecule that can regulate the downstream expression of other genes . Indeed , the Drosophila Unkempt is involved in a number of developmental processes including wing and bristle morphogenesis [72] . The role of Unkempt in bristle morphogenesis is particularly intriguing , as the overlapping scales that color a butterfly wing are thought to have evolved from wing bristles [73] . Moreover , the different color scales in Heliconius have unique morphologies and are pigmented at different times during wing development [74] , suggesting that pattern variation may be tied directly to scale maturation . If Unkempt is shown by additional research to be modulating pattern variation , it could provide yet another example of a conserved developmental gene co-opted to produce novel variation [75]–[77] . Alternatively , it may turn out that the gene that underlies pattern variation in Heliconius is either Lepidoptera-specific or has diverged significantly in both form and function from other insect lineages . LRR has no strong ortholog in Drosophila , the honeybee ( Apis mellifera ) , or the flour beetle ( Tribolium castaneum ) . It is most similar to the Drosophila gene , Sur-8 , which is inferred to have RAS GTPase binding activity . This suggests it may be involved in signal transduction . This gene also showed the highest differentiation among H . melpomene races and between H . melpomene and H . cydno [33] , further implicating this gene and the surrounding regions . Three unlinked genomic intervals , D , Cr and Sd , interact to generate the phenotypic differences between H . e . favorinus and H . e . emma [40] . Yet , the overall effect on phenotype of variation across each of these loci is not identical and the much lower levels of population differentiation in the Cr interval relative to the D interval is likely due a combination of differences in dominance and selection on the two loci . In H . erato , there is a strong dominance hierarchy among the colored scale cells , where red scale cells ( containing xanthommatin and dihydro-xanthommatin ) are dominant to black ( containing melanin ) scale cells and both are dominant to yellow ( containing 3-hydroxy-L-kynurenine ) scale cells . For H . e . emma and H . e . favorinus this means that the D locus is effectively codominant , whereas the Cr allele for the emma lack of hindwing bar is dominant to presence of yellow hindwing bar in favorinus [40] . Additionally , the analysis of cline width , together with the overall percentage of wing surface affected suggests that the selection on the codominant D locus is much higher ( s≈0 . 33 ) than selection on the dominant Cr locus ( s≈0 . 15 ) [23] , [34] . Thus , the power of natural selection to remove poorly adapted alleles at Cr is reduced , especially on the H . e . emma side of the zone where the recessive yellow bar alleles are rare [34] . In H . melpomene the Yb and B locus are homologous to the H . erato Cr and D loci , respectively , and are under similar selective constraints at this Peruvian hybrid zone . Similarly , a reduction in the power of natural selection on the Cr would likely result in a similar reduction of selection on Yb , which may explain why genetic differentiation between the H . melpomene Peruvian races is , like H . erato , much lower at genes near the Yb relative to the B locus ( see [33] , Table 1 ) . Given the history of recombination implied by our data , we expect only sites extraordinarily tightly linked to the causative polymorphisms to yield strong associations . Collectively the association results across the D and Cr intervals highlight the power of using these naturally occurring hybrid zones to select candidate loci for future focused studies . Similar and possibly independently derived transitions between “postman” and “rayed” races of H . erato and H . melpomene that occur in Bolivia , Ecuador , Colombia , Suriname , and French Guiana , provide additional replicates to test the repeatability of evolution [19]–[21] , [78] , [79] . The color pattern genes of Heliconius are classical examples of large effect loci where allelic variation causes major phenotypic shifts in the distribution of melanin and ommochrome pigments across large sections of both the fore- and hindwing . We are accustomed to thinking of the generation of phenotypic variation in terms of single causal mutations . This paradigm has shaped our ideas about how variation is produced by DNA sequences , and , although consistent with some of the handful of examples [2] , [5] , [80] , there are reasons to imagine this is not the whole story , or even a dominant trend [75] , [77] , [81] , [82] . In this light , the varying pattern of LD across the D and Cr intervals and the observation that different polymorphic sites are associated with pattern phenotype in different sets of individuals seems inconsistent with a single causal functional site . Our coarse sampling provides only a preliminary snapshot of haplotype structure across these intervals and genetic hitching , drift and ancestry can create complex genotype-by-phenotype signatures [83]–[86] . Nonetheless , given the rapid breakdown of LD observed , we would expect to see a much narrower window of association if variation was explained by a single causal site . However , the pattern we observe is expected if there are multiple functional sites dispersed across these intervals . Although LD was generally higher , a similar pattern was evident in H . melpomene [33] . These emerging genetic signatures are consistent with early ideas that these patterning loci were “supergenes” composed of a cluster of tightly linked loci [21] . For example , in H . erato the D locus was originally described as three unique loci , D , R , and Y [21] and there has been one H . erato individual collected in the Peruvian hybrid zone which had a DR/y recombinant phenotype [40] . In H . melpomene both the B and Yb loci , have roughly equivalent phenotypic effects as the D and Cr loci in H . erato , and have been clearly shown to be parts of tightly linked clusters of loci that control the end wing pattern phenotype . It is possible that these “clusters” are a series of enhancer elements that influence target gene expression and the terminal phenotype in an overall switch-like fashion [87] . Selection to maintain these clusters may explain the reduced recombination rate we observed across color pattern intervals in the pedigree-based linkage mapping of the D and Cr intervals and the large haplotype blocks across the Yb and B intervals in the Peruvian H . melpomene races [33] . However , in H . erato thousands of generations of hybridization in the middle of the hybrid zone may have allowed recombination to break apart functional sites , creating the mosaic of different haplotypes we observed across these intervals . Collectively , these two companion studies serve as natural replicates of how convergence on a similar adaptive trait can be independently derived and provide compelling evidence that similar genetic changes can underlie the evolution of Müllerian mimicry . Our initial sampling of genetic variation across the color pattern loci provides important insights into the complex haplotype structure that potentially underlies the major phenotypic shifts in wing color patterns . These data suggest that finer genetic dissection of these hybrid zones and other parallel transitions will allow direct tests of the number and type of changes that underlie adaptive color pattern variation in Heliconius . These studies will help pinpoint functionally important polymorphisms and determine if a single functional site or multiple sites underlies adaptive color pattern variation and if these sites are changes in coding regions , in cis-regulatory regions , or both . Ultimately , linking the genetic changes underlying phenotypic variation with the development pathways involved in patterning the wing promises a whole new understanding of how morphological variation is created through development and modified by natural selection within the context of an adaptive radiation .
We screened the H . erato BAC libraries , to identify BAC clones that spanned the D and Cr color pattern intervals . For the Cr locus , probes were designed from the previously published H . erato BAC clone 38A20 ( AC193804 ) and H . melpomene BAC clones 11J7 ( CU367882 ) , 7G12 , ( CT955980 ) and 41C10 ( CR974474 ) [29] , [88] . Across the D locus , probes were designed from the H . erato BAC clone 25K04 ( AC216670 ) and H . melpomene BAC clones 7G5 ( CU462858 ) 27I5 ( CU467807 ) , and 28L23 ( CU467808 ) that have been previously shown to be located near the D locus [31] . BAC library probing , fingerprinting of positively identified clones , and the sequencing and assembly of BAC clones that most extended our tiling coverage was done using the methods described in [88] . BAC clone sequences were aligned using SLAGAN [89] to create contiguous H . erato consensus sequences across the D and Cr color pattern loci . SLAGAN was also used to align these H . erato consensus sequences with available H . melpomene BAC sequences and Bombyx mori genome sequence to confirm the order , orientation and locations of gaps among the H . erato sequences . The consensus H . erato sequences were then annotated using Kaikogaas ( http://kaikogaas . dna . affrc . go . jp ) , an automated annotation package that implements several gene prediction methods to identify potential coding regions . Locations of predicted coding regions and conserved domains are shown in Figure S2 . Primers for probes were designed from potential coding regions using the methods described above , in Butterfly Crosses and Fine-Mapping . Primer sequences and PCR conditions for probes are available in Table S1 . Individuals used in this study were collected from five locations transecting 32 km across a H . erato hybrid zone in Eastern Peru near Tarapoto . In total we sampled 76 individuals , 20 from phenotypically “pure” populations of H . e . favorinus in Tarapoto and Rio Pansillo , 14 individuals from a primarily phenotypically “pure” population of H . e . emma in Davidcillo , and 42 from admixed populations in Pongo de Cainarache and Barranquitas located near the center of the hybrid zone . Due to dominance and strong epistasis between the three loci , when individuals have a DemmaDemma genotype , the CremmaCremma and CremmaCrfav genotypes are phenotypically indistinguishable . Therefore some individuals were assigned a Cremma- dominant genotype ( see [34] ) , indicating the genotype of the second Cr allele is unknown . Individual's names , sex , race , color pattern genotype and collection location are recorded in Table S4 . Nucleotide variation was sampled across two candidate intervals controlling major changes in warning color patterns , as well as three other autosomal genes unlinked to color pattern . We sampled twelve coding regions from D interval , ten from the Cr interval , and three coding regions from genes on unlinked chromosomes ( Table S1 ) . Names of coding regions are based on sequence homology to annotated genes in other organisms , or if no sequence homology was found numbered gene names were assigned . On average , 520 bp fragments were sampled every 47 kb across a candidate color pattern interval . Primer design for PCR amplification and sequencing was done using Primer3 [90] . Primers for the three unlinked loci were developed by Pringle et al . [91] , and have been shown to map to different linkage groups in H . melpomene . Primer sequences and PCR conditions for each locus can be found in Table S1 . Genomic DNA extraction , PCR product purification and sequencing were completed using the same methods as described above . For some individuals , Abhydrolase was cloned from purified PCR product using TOPO cloning kit ( Invitrogen ) and 4–10 clones were sequenced . Ambiguous bases in the genomic sequences were cleaned and trimmed manually using Sequencher ( Gene Codes Corporation ) . A site was determined to be heterozygous if the lower quality nucleotide had a peak height at least 50% of the higher quality nucleotide . Sequences were initially aligned using Sequencher and the resulting alignments were then manually adjusted . Population genetic estimates of nucleotide diversity , population differentiation and tests of neutrality were conducted using SITES and HKA [92] . Nucleotide diversity was estimated as π ( average number of pair-wise differences per base pair ) for all samples i ) within the H . e . favorinus population ii ) within the H . e . emma population and iii ) within the admixed population . This was done for each of the 25 sampled coding regions independently and by concatenating all coding regions sampled on the same chromosome ( Table S1 ) . Tajima's D [93] was also calculated for all coding regions independently and by concatenating them , to examine for departures from the neutral model of evolution . For each coding region 10 , 000 coalescent simulations based on locus specific estimates of theta were used to determine if the observed patterns of nucleotide diversity and locus specific estimates of Tajima's D significantly departed from neutral expectations using the program HKA as described in [94] . FST was estimated between the two phenotypically pure populations and the admixed population for each coding region and using SITES ( Table 2 ) and FDIST2 ( Table S2 ) [95] . To determine the extent of LD across the candidate color pattern intervals in Heliconius , we computed composite LD estimates for 432 SNPs from the 25 coding regions we sampled . Of the 1542 polymorphic sites identified in this study , 442 sites had a minor allele frequency greater than 0 . 05 and were considered informative for LD analyses . Multi-allelic sites that had a minor allele with a frequency less than 0 . 05 were condensed to bi-allelic SNPs by merging the minor allele genotypes . Ten polymorphic sites had 2 or 3 minor alleles with a frequency greater than or equal to 0 . 05 that were not condensed to bi-allelic SNPs and were not included in the LD analyses . LD between the remaining 432 SNPs was executed using the commonly used composite estimate of LD method described by Weir [96] , which does not assume HWE or that haplotypes are known . LD among the 432 SNPs was estimated independently for i ) all samples ii ) within the H . e . favorinus population iii ) within the H . e . emma population and iv ) within the admixed population . LD between the 432 SNPs using all sampled individuals was visualized with GOLD [97] , by plotting the composite r2 estimates between all pair wise SNP combinations . To visualize the difference in mean r2 between the three populations , a sliding window average of r2 across 50 neighboring SNPs was calculated independently for each population and plotted by distance . We determined if any SNP was associated with a color pattern phenotype using chi-squared linear trend test [96] . This test assumes a linear relationship between the phenotype and genotype and applies a chi-square goodness-of-fit test to determine if the genotype at a SNP is significantly associated with a particular wing color pattern . For the association tests we used bi-allelic and multi-allelic SNPs with minor allele frequencies equal to or greater than 0 . 05 . Color pattern phenotypes at the D and Cr loci were scored as 0 . 0 representing H . e . favorinus phenotypes , 0 . 5 representing hybrid phenotypes and 1 . 0 H . e . emma phenotypes . Individuals with H . e . emma D phenotypes and H . e . emma Cr phenotypes were assigned 1 . 0 for the D phenotype score and 0 . 5 for the Cr phenotype score , due to the effects of dominance previously mentioned; varying the Cr value for from 0 . 5 to 1 . 0 for these individuals had a negligible effect on the association test results ( data not shown ) . We used quantitative PCR involving SYBR Green technology to detect transcript levels of kinesin , Slu7 , GPCR , Dna-J , and VanGogh in butterfly forewing tissues . Samples of whole forewings were dissected from December , 2008 - February , 2009 from reared H . e . emma and H . e . favorinus stocks founded from multiple individuals collected within 30 km of one another in Peru . We staged individuals indoors at 25°C starting in early 5th larval instar . Chosen larval wings were at mid-5th instar , stage 2 . 25–2 . 75 based on the work of Reed and colleagues [98] . Pupal stages were based on the time after the pupal molting event , including Day 1 ( 24hrs ) , Day 3 ( 72hrs ) , and Day 5 ( 120hrs ) . We sampled three individuals of each stage and race , resulting in 2 races × 4 stages × 3 biological replicates = 24 specimens . All specimens were processed randomly from dissection through amplification stages . We extracted total RNA from the tissues using an electric tissue homogenizer and the RNAqueous Total RNA Isolation Kit ( Ambion ) . This procedure was followed by a TURBO DNA-free ( Applied Biosystems ) treatment to remove genomic DNA contaminants . Extracted products were run through the Agilent Bioanalyzer to ensure the RNA was of high quality . For cDNA systhesis , 0 . 4 µg of each sample was added to the standard 20µl reaction procedure outlined in the High-Capacity cDNA Reverse Transcription Kit ( Applied Biosystems ) . Resulting products were diluted with an additional 50µl . For each gene , we performed quantitative PCR on all 24 samples in triplicate to correct for technical error . We used EF-1α as a standard to normalize the expression of the test genes . Primers for amplification of cDNA were designed using recommended criteria and range from 98 – 175 bp in length ( see Table S3 ) . We ran primer sets through an initial qPCR optimization to test for optimal primer concentrations and ran DNA-free controls to test for primer-dimers . qPCR reactions were run using 1µl of 5µM primers ( 0 . 5µl for GPCR ) , 5µl SYBR Green Mix , 1 µl template , and water to 10 µl . Reactions were run in 384-well plates in the Applied Biosystems 7900HT Fast Real-Time PCR machine under standard mode and absolute quantification for 40 cycles of 95°C for 15 sec , 60°C for 60 sec . Each cycle was followed by a dissociation step to measure the dissociation temperature of the sample , which tests for primer-dimer and differences in sequences among samples . A standard curve was generated for each gene using a 10−3 to 10−7 dilution series drawn from a PCR amplified product using the same primers . To normalize Ct values from the qPCR run , we first calculated the mean of each of the three technical replicates . We then calculated initial concentrations for each sample for each gene given the equation of the standard curve for that gene . These initial concentrations were divided by the inferred concentration of EF-1α for that sample , thus correcting for inconsistencies in initial cDNA sample concentrations . These relative values were then log2 transformed for presentation and analysis . Log2 transformation is necessary to normalize the variances of the samples and represents expression differences in more biologically realistic fold differences . Significance values were obtained from a two-way ANOVA using stage , race , and race*stage as effects . Effects of race within each stage were further dissected for each gene using series of t-tests and an FDR of 0 . 05 ( threshold at p = 0 . 0028 ) to correct for multiple testing . In addition to a general ANOVA and to compare our results to the companion paper [33] , we used a combination of generalized linear regression models ( GLMs ) and Bayesian Model Averaging ( BMA ) on the non-log transformed data to model the effect of race , developmental stage , and their interactions , on gene expression . These statistics were performed using the ‘bic . glm’ function in the ‘BMA’ package [99] implemented in R ( R Development Core Team 2008 ) . | Identifying the genetic changes responsible for beneficial variation is essential for understanding how organisms adapt . Here , we use a combination of mapping , population genetic analysis , and gene expression studies to identify the genomic regions responsible for phenotypic evolution in the Neotropical butterfly Heliconius erato . H . erato , together with its co-mimic H . melpomene , have undergone parallel and concordant radiations in their warningly colored wing patterns across Central and South America . The “genes” underlying the H . erato color pattern radiation are classic examples of Mendelian loci of large effect and are under strong natural selection . Nonetheless , we do not see a clear molecular signal of recent natural selection , suggesting that the H . erato color pattern radiation , or the alleles that underlie it , may be quite old . Moreover , rather than being single locus , the genetic patterns suggest that multiple , widely dispersed loci may underlie pattern variation in H . erato . One of these loci , a kinesin gene , shows parallel expression differences between races during wing pattern formation in both H . erato and H . melpomene , suggesting that it plays an important role in pattern variation . High rates of recombination within naturally occurring H . erato hybrid zones mean that finer genetic dissection will allow us to localize causative sites and better understand the history and molecular basis of this extraordinary adaptive radiation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"evolutionary",
"biology/evolutionary",
"and",
"comparative",
"genetics",
"evolutionary",
"biology/genomics",
"evolutionary",
"biology/pattern",
"formation",
"evolutionary",
"biology",
"genetics",
"and",
"genomics/population",
"genetics"
] | 2010 | Genomic Hotspots for Adaptation: The Population Genetics of Müllerian Mimicry in Heliconius erato |
orb is a founding member of the CPEB family of translational regulators and is required at multiple steps during Drosophila oogenesis . Previous studies showed that orb is required during mid-oogenesis for the translation of the posterior/germline determinant oskar mRNA and the dorsal-ventral determinant gurken mRNA . Here , we report that orb also functions upstream of these axes determinants in the polarization of the microtubule network ( MT ) . Prior to oskar and gurken translational activation , the oocyte MT network is repolarized . The MT organizing center at the oocyte posterior is disassembled , and a new MT network is established at the oocyte anterior . Repolarization depends upon cross-regulatory interactions between anterior ( apical ) and posterior ( basal ) Par proteins . We show that repolarization of the oocyte also requires orb and that orb is needed for the proper functioning of the Par proteins . orb interacts genetically with aPKC and cdc42 and in egg chambers compromised for orb activity , Par-1 and aPKC protein and aPKC mRNA are mislocalized . Moreover , like cdc42- , the defects in Par protein localization appear to be connected to abnormalities in the cortical actin cytoskeleton . These abnormalities also disrupt the localization of the spectraplakin Shot and the microtubule minus-end binding protein Patronin . These two proteins play a critical role in the repolarization of the MT network .
Specification of the anterior-posterior ( AP ) and dorsal-ventral ( DV ) axes of the Drosophila embryo depends upon determinants that are localized within the egg during oogenesis [1–4] . For example , expression of the TGF-α cell signaling molecule Gurken ( Grk ) at the anterior corner of the oocyte during mid-to-late oogenesis establishes the DV axis of the egg and subsequently the embryo by signaling to the overlying somatic follicle cells [5–8] . Factors important in determining the AP axis of the embryo are also localized during this same period . Specification of the posterior axis is mediated by oskar ( osk ) [9 , 10] . osk mRNA is targeted to the posterior cortex of the oocyte , where it is translated and functions in the assembly of the pole plasm and the anchoring of the mRNA encoding the posterior determinant nanos [11 , 12] . The anterior axis is specified by the Bicoid transcription factor , and its mRNA is localized to the anterior cortex of the oocyte [13–15] . The proper localization of these determinants within the oocyte during mid-to-late oogenesis depends upon the disassembly of the existing microtubule cytoskeleton ( MT ) during stage 7 of oogenesis and its subsequent repolarization [7 , 8] . The polarity of the MT network in the period prior to stage 7 is established early in oogenesis when the oocyte is initially specified [16] . A microtubule organizing center ( MTOC ) is assembled at the oocyte cortex just posterior to the oocyte nucleus and it directs the elaboration of the MT network by anchoring the minus-ends of MTs . As a consequence of this polarization of the oocyte , mRNAs encoding determinants critical for the early stages ( stage 1–7 ) of egg chamber development accumulate at the posterior cortex . One of these is gurken ( grk ) mRNA . Grk protein translated from this localized message signals to the somatic follicle cells covering the posterior of the egg chamber to specify posterior follicle cell fate ( PFC ) [7 , 8] . Subsequently , during stage 7 , an unknown signal ( s ) emanating from the somatic PFCs triggers the repolarization of MT network in the germline . This signal induces the disassembly of the posterior MTOC and the network of MTs extending from the MTOC towards the anterior of the oocyte [7 , 8 , 17–20] . At the same time , de novo MT assembly is nucleated along the anterior and lateral cortex of the oocyte by a centrosome independent mechanism . This mechanism deploys the tubulin minus-end binding protein Patronin and the actin-MT linker Short Stop ( Shot ) [21] . Accompanying the repolarization of the MT cytoskeleton , the oocyte nucleus migrates from the posterior end of the oocyte to the anterior corner [22] . grk mRNA also relocates so that it is positioned between the oocyte nucleus and the oocyte cortex . Grk protein expressed from the localized message signals dorsal follicle cell fate and this defines the DV axis of the egg chamber and embryo [6 , 23] . In addition to Patronin and Shot , the other factors implicated in oocyte repolarization are the Drosophila homologs of the partioning-defective ( Par ) group genes , par-1 , cdc42 and bazooka ( baz/par-3 ) [24–26] . These three genes together with par-6 and aPKC are also required for the initial polarization of the stage 1 egg chamber [24–30] . These proteins generate cellular asymmetries by inhibitory cross-regulatory interactions that impede association with the cell cortex [25 , 31–33] . During MT repolarization , Par-1 becomes enriched along the posterior cortex of the oocyte [34–36] . There is a complementary distribution of Baz , Par-6 , aPKC and Cdc42: they are enriched along anterior and anterior-lateral cortex , but not the posterior [24 , 25 , 37 , 38] . The available evidence indicates that the asymmetry in the oocyte generated by the activation of the Par polarity network is upstream of the localization of Shot and Patronin along the anterior and lateral cortex , and thus the Patronin dependent de novo assembly of MTs [21] . In addition to being critical for properly localizing grk , osk and bcd mRNAs , the reorganization of the cytoskeleton also alters the distribution of other mRNAs encoding oocyte-specific proteins . One of these mRNAs is orb , which encodes one of the two fly cytoplasmic polyadenylation element RNA-binding ( CPEB ) proteins [39 , 40] . During early stages of oogenesis , orb mRNA is localized at the posterior of the oocyte . After repolarization orb mRNA disappears from the posterior and becomes concentrated along the anterior-lateral margin of the oocyte [41] . While the rearrangement of orb mRNA within the oocyte is clearly downstream of the steps involved in repolarizing the oocyte MT network , the orb gene plays a central role in the initial formation and subsequent development of the oocyte and thus could be an active participant in determining oocyte polarity . In ovaries , orb expression is restricted to the germline and is required at multiple steps during oogenesis [39 , 40 , 42] . In wild type ovaries , a cystoblast , generated by an asymmetric division of a stem cell , undergoes four mitotic divisions with incomplete cytokinesis to produce a 16-cell cyst [1] . In the orb null allele , orb343 , the last of these mitotic divisions is not completed and the cyst degenerates [40] . While the strong loss-of-function allele , orb303 , forms a 16-cell cyst , the oocyte is not properly specified and egg chambers contain only nurse cells [40] . Unlike orb343 and orb303 , the Orb protein expressed by the hypomorphic orb allele , orbmel , is wild type . Instead , orbmel transcripts are incorrectly spliced generating an mRNA lacking sequences from the 5’UTR [42] . The removal of these 5’ sequences alters Orb expression as oogenesis proceeds . Prior to stage 7 the level and localization of Orb in the oocyte is similar to that observed in wild type . However , beginning around stage 7 , the amount of Orb drops dramatically and most chambers have little residual protein . As a consequence of this reduction in Orb protein , orbmel females produce eggs that give rise to embryos with a range of phenotypic abnormalities including D-V and A-P patterning defects [42] . These patterning defects arise from a failure in the localization and/or translation of two Orb regulatory targets , grk and osk mRNAs , during mid-to-late oogenesis [43–46] . grk and osk transcripts are not , however , the only mRNAs that could be subject to orb regulation during oogenesis . Several recent studies have identified many other mRNAs that are Orb associated in vivo [47 , 48] . Included in this group of potential orb regulatory targets are mRNAs encoding the Par proteins , aPKC , Baz , Par-6 and Cdc42 . Moreover , there is evidence connecting the other fly CPEB protein , Orb2 , to the functioning of one of the Par family proteins , aPKC , in cell polarization in the embryonic CNS , in testes and in tissue culture cells [49–51] . These observations prompted us to ask whether orb impacts the process of repolarization of the oocyte during mid-stages of oogenesis , and conversely whether the Par proteins , and in particular , aPKC , have any effect on orb activity .
In wild type , osk mRNA is localized in a tight crescent at the posterior pole of the oocyte after repolarization ( Fig 1A ) [11 , 52] . While osk mRNA localization to the posterior is independent of Osk , Osk protein is required to ensure that osk mRNA is properly anchored to the posterior cortex [52] . In osk protein null mutants , osk mRNA is localized at the posterior , but localization is not properly maintained ( Fig 1B ) . While orb is required for osk mRNA translation , it also plays a role in the proper localization of osk message [43 , 44] . As shown in Fig 1C , in orbmel/orb303 chambers , the tight localization of osk mRNA at the posterior pole is lost . Instead , osk mRNA puncta are distributed in a halo around the posterior pole while there is a diffuse pattern of mRNA along the anterior margin of the oocyte . As previously reported , even more extreme defects in osk mRNA localization are evident when orbmel is combined with the null allele orb343 ( Fig 1D ) [42 , 43] . In this allelic combination there is little if any osk mRNA at the posterior . The osk mRNA localization defects in the hypomorphic orb mutant combinations resemble those in staufen mutants . staufen encodes an RNA-binding protein that co-localizes with osk mRNA throughout oogenesis , and in staufenD3/Df , osk mRNA is partially localized to the posterior and also accumulates at the anterior ( Fig 1E ) [11 , 52 , 53] . Thus , one explanation for the defects in osk mRNA localization during mid-oogenesis is that orb is also required to transport osk mRNA [53] . To test this possibility we compared the distribution of Orb protein with that of osk mRNA . Prior to stage 7 , both osk mRNA and Orb protein are localized at the posterior . When the MT network commences repolarization during stage 7 , osk mRNA transiently accumulates in a cloud near the middle of the oocyte ( Fig 2A ) [54] . If Orb is directly involved in osk mRNA transport , it would be expected to co-localize with osk mRNA in this cloud . However , it does not . Instead , most of the Orb is concentrated in the sub-cortical region at the posterior end of the oocyte and along the lateral margins of the oocyte ( Fig 2A ) . Only later , after osk mRNA is re-localized to the posterior pole ( and presumably being translated ) does it again overlap with the posterior cap of Orb protein ( Fig 2B ) . Another orb regulatory target is orb mRNA and its pattern of localization differs from that of osk [55] . In stage 7 chambers , when osk mRNA is in the center of the oocyte , orb mRNA has a circumferential subcortical distribution around the anterior of the oocyte ( Fig 2C ) . This distribution is maintained at later stages ( Fig 2D ) . Another indication that orb is not directly involved in osk mRNA transport comes from the effects of grk mutations . In grk2B/2E12 ovaries , posterior follicle cell ( PFC ) specification is defective and the oocyte fails to initiate repolarization at stage 7 [7 , 8] . As a consequence , osk mRNA ( S1 Fig ) , the transport protein Staufen , and MT plus ends become enriched in the center of the oocyte , while bicoid mRNA localizes to both the anterior and posterior of the oocyte [7 , 8] . In this grk mutant combination Orb protein and also orb mRNA accumulate around the circumference of the oocyte , far from osk mRNA ( S1 Fig ) . An alternative explanation for the mislocalization of osk mRNA in orbmel/orb303 and orbmel/orb343 egg chambers is that the cytoskeleton is not properly reorganized during repolarization in the absence of normal orb function . This possibility was suggested by the studies of Martin et al . ( [56] ) , who showed that in hypomorphic orb mutant alleles the oocyte MT network is disrupted and there is premature oocyte cytoplasmic streaming . Several approaches were used to confirm and extend their findings . Repolarization of the oocyte during stage 7 is triggered by signals emanating from the somatic posterior follicle cells ( PFCs ) . The production of the repolarization signal depends upon the proper specification of the PFCs and this process is orchestrated by the expression of the Grk ligand at the oocyte posterior [7 , 8] . Since grk mRNA is a known orb regulatory target , one explanation for the repolarization defects is that the PFCs are not properly specified when orb activity is compromised . To test this possibility we examined the expression of an EGFR dependent enhancer trap , kekkon-lacZ , that is activated in follicle cells by grk signaling [68–72] . As illustrated for two stage 7 egg chambers in S4A and S4B Fig we found that kekkon-lacZ expression in PFCs in orbmel/orb343 egg chambers resembles that in control egg chambers . This result confirms previous studies which showed that anterior follicle cell fate is not duplicated in orb343/mel egg chambers [7] . While kekkon-lacZ expression is unaffected in orbmel/orb343 prior to repolarization , abnormalities are evident at later stages . As shown for a stage 10 orbmel/orb343 chamber in S4C and S4D Fig , expression of kekkon-lacZ in dorsal follicle cells is severely reduced compared to the control . This is expected since grk signaling to the dorsal follicle cells is known to be disrupted in orbmel/orb343 ovaries [42 , 45 , 46] . Other observations are also consistent with the idea that the defects in MT organization in orb are downstream of both the grk dependent specification of PFCs and of the subsequent repolarization signal from the PFCs to the oocyte . For example , in grk mutants , bicoid mRNA is localized not only along the anterior-lateral margin , but also at the posterior pole[7 , 8] . In contrast , when orb activity is compromised , localization of bcd mRNAs to the posterior pole is not observed ( S5 Fig ) [42] . The reason for this difference is that in grk mutants the PFCs fail to signal the disassembly of the MTOC at the posterior of the oocyte , whereas the posterior MTOC is disassembled in orb mutants . One explanation for the failure to repolarize the MT cytoskeleton is that orb activity impacts either directly or indirectly the functioning of the Par proteins . In fact , precedence for an orb-Par connection comes from experiments showing that one of the targets for the other fly CPEB protein , orb2 , in spermatid cyst polarization and in asymmetric cell division in the embryo is the message encoding the apical Par protein aPKC [49 , 50] . To explore this idea further we took advantage of the fact orb is weakly haploinsufficient for D-V polarity [55] . About 5% of the eggs laid by orb343/+ are ventralized due to defects in translating grk mRNA at the dorsal anterior corner of the oocyte ( Fig 4A ) . The frequency of D-V polarity defects can be enhanced by reducing the activity of other genes that are important for orb function in grk signaling . We used three different aPKC mutants , a strong allele , k06403 , and two hypomorphic alleles , ex48 and ex55 , to test for dominant genetic interactions with orb [74 , 75] . While the frequency of D-V polarity defects in eggs produced by mothers heterozygous for these three aPKC alleles is similar to WT ( S1 Table ) , these mutations substantially enhanced the frequency of D-V polarity defects when trans to orb343/+ . The weak hypomorphic alleles increase the frequency of ventralized eggs four to five fold ( 20% and 25% ) , while the frequency is increased nearly nine fold ( 44% ) by the null allele ( Fig 4A ) . To extend this analysis , we also asked whether there are genetic interactions between orb343 and the cdc42 gene , which , like aPKC , plays an important role in establishing apical cell polarity [26 , 76] . In orb343/ cdc421 trans-heterozygotes there was modest increase ( three-fold ) in the frequency of D-V polarity defects ( S1 Table ) , while in orb343/cdc424 trans-heterozygotes the frequency of D-V polarity defects increased by nearly fifteen fold ( Fig 4A ) . Consistent with the idea that the effects on grk signaling are related , at least indirectly , to the functioning of the Par proteins in the process of repolarization , we also observed genetic interactions between aPKCk06403 and cdc424 . Whereas background levels ( ~1% ) of D-V polarity defects are evident in eggs produced by either aPKCk06403 and cdc424 heterozygotes , over 35% of the eggs laid by trans-heterozygous mothers had D-V polarity defects ( Fig 4A ) . We also examined eggs produced by females triply heterozygous for orb343 , aPKCk06403 and cdc424 . In this triple heterozygote about 90% of the eggs have D-V polarity defects ( Fig 4A ) . As would be predicted , accumulation of Grk protein at the dorsal anterior corner of the oocyte is clearly reduced ( S6A–S6C Fig ) . Interestingly , the oogenesis defects are not restricted to grk translation . S6D–S6F Fig also shows that the localization of osk mRNA at the posterior pole is also reduced compared to control egg chambers . Further evidence that orb might work in conjunction with aPKC in the process of repolarization comes from analysis of oocyte nucleus positioning in backgrounds simultaneously compromised for both genes . As described above , oocyte nucleus mispositioning is observed in ~7% of the orbmel/orb343 egg chambers . The frequency of a mispositioned oocyte nucleus increases to nearly 25% of the chambers when orbmel/orb343 females are also heterozygous for aPKC k06403 ( Fig 4B ) . A similar enhancement is observed when aPKC k06403 is combined with orbmel/orb343 HD19G . In orbmel/orb343 HD19G chambers about 20% have a mispositioned oocyte nucleus , while the frequency of oocytes with a mispositioned nucleus increases to nearly 50% when the orbmel/orb343 HD19G females are also heterozygous for aPKC k06403 ( Fig 4B ) . Importantly , aPKC on its own is not haploinsufficient for proper oocyte nucleus migration ( Fig 4B ) . One plausible explanation for the genetic interactions is that one of the orb functions in repolarization is to regulate aPKC mRNA . To explore this possibility , we examined the effects of compromising orb on the pattern of accumulation of aPKC mRNA . While aPKC mRNA is present in both somatic and germline cells in wild type ovaries , the highest concentrations of mRNA in the germarium and in stage 1–7 egg chambers are found in the oocyte ( S7A and S7B Fig ) . In stage 9 and older chambers , aPKC mRNA is no longer enriched in the oocyte relative to levels in the nurse cells; however , within the oocyte a fraction of the mRNA localizes along the oocyte cortex with the highest levels of aPKC mRNA towards the anterior of the oocyte and lower levels towards the posterior ( arrows in Fig 5A and 5B and S7C Fig ) . Orb protein also localizes along the lateral cortex of the oocyte in wild type egg chambers ( see Figs 2 and 6 ) , while in chambers compromised for orb , Orb association with the cortex is substantially reduced ( S3 Fig ) . Consistent with a role for Orb in anchoring aPKC mRNA during mid-oogenesis , we find that the anterior and lateral cortex associated aPKC mRNA is either partially ( Fig 5C ) or largely ( Fig 5D & 5E ) lost when orb activity is depleted by RNAi . As noted in the introduction , aPKC mRNA is one of several thousand mRNAs that are associated with ectopically expressed Orb2 and Orb in tissue culture cells [48] . To determine if aPKC mRNA is bound by Orb in ovary extracts , we used immunoprecipitation to isolate Orb associated RNAs . After reverse transcription using an oligo dT primer , we used quantitative PCR to assay for specific mRNA species . For the positive control , we used primers for orb-RA 3’UTR which contains four canonical cytoplasmic polyadenylation elements ( CPEs: UUUUAU or UUUUAAU ) . Previous studies have shown that Orb binds to the orb mRNA 3’UTR and positively autoregulates its own expression [55] . There are twelve predicted aPKC mRNA species with six different predicted 3’UTRs . Four of the six predicted 3’UTRs have canonical CPE sequences . One of these , aPKC-RA , has a 3’UTR with three canonical CPEs while the remaining aPKC mRNAs ( RD , RF RJ , RK , RL and RM ) have overlapping UTRs with 2 canonical CPEs . Fig 5F shows that in ovary extracts both types of aPKC 3’UTRs are enriched in Orb immunoprecipitates . We next examined the pattern of accumulation of aPKC protein . In wild type stage 10–11 oocytes , aPKC protein is localized to the anterior-lateral cortex where it appears to be in close association with the cortical actin network ( Fig 6A ) [37] . Except for this cortically localized protein , there is little aPKC elsewhere in the oocyte . Orb is localized just interior to the cortical actin-aPKC layer ( Fig 6A ) . aPKC is also localized along the apical surface of the somatic follicle cells facing the germline , and in confocal images the somatic and oocyte aPKC proteins typically appear as a set of parallel tracks along the anterior-lateral cortex ( Fig 6A ) . The pattern of aPKC localization in the oocyte is altered when orb activity is compromised . Instead of a tightly organized track coincident with cortical actin , aPKC protein distribution becomes irregular and patchy ( Fig 6B ) . In some regions , there are small gaps ( Fig 6B: lower panel: blue arrowhead ) while in other regions aPKC is missing altogether ( Fig 6B , lower panel: red arrowhead ) . In other cases , the aPKC protein extends from the cortex into the interior of the oocyte ( Fig 6B , middle panel: red arrowhead ) . The effects of reducing orb activity on aPKC localization within the oocyte , taken together with the genetic interactions between orb , aPKC and cdc42 indicate that orb is required for the proper functioning of anterior Par proteins . It seemed possible that the posterior Par proteins might also be dependent on orb . To test this idea , we examined the localization of a Par-1-GFP fusion protein that is expressed in the germline . In control stage 8–11 oocytes , the Par-1-GFP fusion protein localizes along the oocyte cortex and tends to be enriched towards the posterior of the oocyte . Fig 6C and S8 Fig show that like aPKC , Par-1-GFP localization depends upon orb , and is disrupted when orb activity is compromised . The extent of disruption is correlated with the severity of the reduction in orb activity . In orbmel/orb303 , a small percentage of the chambers have an obvious , but not complete loss of Par-1-GFP association with the oocyte cortex ( Fig 6C ) . Even more extensive alterations are observed in orbmel/orb343 and orbmel/orb343 HD19G chambers . In these genetic backgrounds , more than half of the egg chambers show either a reduction ( S8B Fig ) or complete loss of cortical Par1-GFP ( Fig 6C and S8C Fig ) . In addition to its functions in Par dependent polarity , the apical Par protein Cdc42 can also activate effectors of the actin cytoskeleton ( Cip4 , WASp and Arp23 ) . Studies by Leibfried et al . ( [26] ) have shown that one of the important Cdc42 targets during oocyte repolarization is the actin cytoskeleton . When cdc42 activity is compromised , the organization of cortical actin is disrupted . While the apical Par proteins aPKC and Baz are not thought to have a direct role in modeling the actin cytoskeleton , they are required for Cdc42 localization . As a consequence , aPKC and baz mutants have equivalent defects in the anterior lateral cortical domain . For these reasons , we wondered whether orb function might also impact the organization of the cortical actin cytoskeleton during repolarization . To address this question we examined the cortical actin cytoskeleton in orbmel/orb343 and in orb RNAi egg chambers . In the experiment in Fig 7 , we labeled follicle cells membranes with Cadherin 99C ( Cad99C ) antibodies , while the actin cytoskeleton was labeled with phalloidin [77] . In wild type , actin is enriched along anterior oocyte margin and the anterior lateral cortex ( Fig 7A and 7A’ ) [26] . The tight association of actin along the oocyte cortex seen in wild type chambers is disrupted when orb activity is compromised in either orbmel/orb343 oocytes ( Fig 7B and 7B’ ) or when orb RNAi is expressed during midstages by a maternal α-tubulin driver ( #7062 ) ( Fig 7C and 7C’ ) . In some regions , the actin matrix is displaced from the cortex ( Fig 7B , arrow ) . In other regions , there are “flares” of actin filaments that extend out from the cortical actin matrix into the ooplasm ( Fig 7C’ , arrow ) . The matrix can also unravel forming small bubbles ( Fig 7C’ , arrowhead ) or even disappear completely ( Fig 7B’ ) . These defects could be due to a failure to properly crosslink the cortical actin bundles . aPKC association with the oocyte cortex is thought to depend upon the integrity of the cortical actin cytoskeleton [26] . This raises the possibility that the defects in aPKC localization in orb mutants might be connected to abnormalities in the cortical actin cytoskeleton . The results shown in Fig 6B indicate that this is likely to be the case . In regions where the cortical actin matrix is disrupted , aPKC association with the cortex is reduced or lost . There seems to be a similar connection between the severity of the defects in Par-1 localization and the extent of the abnormalities in the cortical actin cytoskeleton ( see S8 Fig ) . The orbmel/orb343 chambers that have most extensive perturbations in the cortical actin cytoskeleton have more pronounced defects in Par-1 localization ( S8C Fig ) than in chambers in which the cytoskeleton defects are less severe ( S8B Fig ) . The repolarization of the MT network during mid-oogenesis depends upon the MT binding protein Patronin and its association with the actin-MT linker Shot . Since the integrity of the cortical actin network is disrupted when orb activity is compromised , we wondered whether Patronin and Shot association with the oocyte cortex is also affected . To investigate this possibility , we compared the localization of Shot-YFP expressed from a BAC transgene and YFP-Patronin expressed from a germline specific mattub promoter ( Fig 8 ) ( [21] ) in wild type egg chambers and in chambers in which orb activity was knocked down by RNAi . In wild type stage 9–11 oocytes Shot and Patronin are found associated with the oocyte cortex ( Fig 8 , S9 and S10 Figs ) [21] . In the oocyte , Shot and Patronin are localized in a punctate pattern just underneath the cortical actin network ( S9 and S10 Figs ) . Shot-YFP ( S9 Fig ) and Patronin-YFP ( S10 Fig: expressed as an endogenously tagged protein ) also localize to the apical surface of the follicle cells , and these two proteins appear as a parallel track along the lateral surface of the oocyte with the cortical actin network in between . Both Shot-YFP and YFP-Patronin are enriched along the anterior and anterior-lateral cortex , while they are absent from the posterior cortex ( Fig 8A–8C; arrowheads ) . When orb activity is knockdown by RNAi , the association of Shot-YFP and YFP-Patronin with the anterior-lateral cortex of the oocyte is disrupted and much of the protein is instead distributed in the ooplasm ( Fig 8B and 8D ) . Similar , though not quite as severe alterations in the cortical association of Shot-YFP and Patronin-YFP are observed in stage 9–11 orb343/orbmel egg chambers ( S9B and S9B’ Fig and S10B–S10B” Fig ) . Par proteins establish and maintain polarity within a cell by both positive and negative cross-regulatory interactions . For this reason it seemed possible that aPKC and orb function in the oocyte might be mutually interdependent . To explore this possibility we used the mid-oogenesis GAL4 driver maternal α-tubulin ( 7063 ) to express aPKC RNAi ( 35140 ) . In this background , we observed that the oocyte nucleus is mispositioned in 69% of the stage 9–11 egg chambers when aPKC activity is depleted . Accompanying the oocyte nucleus position defects , Gurken protein is mislocalized with the oocyte nucleus ( Fig 9A and 9B ) . Additionally , there are alterations in the pattern of Orb protein accumulation . Instead of being distributed subcortically along the entire surface of the oocyte , high levels of Orb accumulate at the anterior oocyte-nurse cell margin ( Fig 9C and 9D ) . There is also a reduction in the posterior cap of osk mRNA compared to wild type ( Fig 9E and 9F ) . Similar effects on the positioning of the oocyte nucleus and the localization of polarity markers ( Staufen and Vasa ) have been reported for cdc42 [26] . Moreover , like orb and cdc42 , the cortical actin network is also perturbed in the aPKC knockdown ( Fig 9H and 9H’ ) . The alterations in the pattern of Orb protein accumulation in the RNAi knockdown experiments prompted us to ask whether aPKC impacts orb autoregulation . Orb promotes its own expression through sequences in the orb mRNA 3’UTR . When the orb 3’UTR is linked to coding sequences for E . coli β-galactosidase in the HD19 transgene , expression of β-galactosidase becomes dependent upon orb activity [55] . S11A Fig shows that β-galactosidase expression from the HD19 ( hsp83: lacZ-orb 3’UTR ) transgene is also dependent upon aPKC activity . In the aPKC mutant combination , aPKCk06403/aPKCex48 , β-galactosidase expression is reduced about two-fold compared to the control ( S11B Fig ) . Like CPEB proteins in other species , orb activity is regulated by phosphorylation [78] . In wild type ovaries , there are multiple phosphorylated isoforms . On standard SDS polyacrylamide gels these different Orb isoforms typically resolve into a closely spaced doublet with the more heavily phosphorylated isoforms migrating more slowly ( S11C Fig ) . In S11C Fig , we compared the relative yield of the upper ( more phosphorylated ) and lower ( less phosphorylated ) bands in wild type and aPKCk06403/aPKCex48 mutant ovaries . In the aPKCk06403/aPKCex48 the ratio of upper to lower bands is reduced compared to wild type ( S11D Fig ) .
Previous studies have implicated orb in the translational regulation of osk and grk in the stages following the repolarization of the MT network [43–45] . In addition , the proper localization of these mRNAs also depends orb activity [42–45] . This observation led to the idea that in addition to controlling translation , orb might also have a role in transport and/or anchoring of these mRNAs once they were properly localized . While our results argue against a direct role in transport , they support the idea that the mislocalization of osk and grk mRNAs when orb activity is compromised during mid-oogenesis arises at least in part because orb is required for the proper organization of both MTs and the cortical actin cytoskeleton . The reorganization of the MT network after stage 7 is a multistep process . It begins with a signal from the PFCs that induces the disassembly of the MTOC that is located just posterior to the oocyte nucleus . The production of this somatic signal depends upon the proper specification of the PFCs , and PFC specification requires the expression of Grk protein at the posterior pole of oocyte earlier in oogenesis [7 , 8] . Translation of grk mRNA at the posterior pole during stages 1–7 depends upon orb , and consequently it functions upstream of PFC specification . However , in our experiments orb activity prior to stage 7 is not limiting , and sufficient amounts of Grk are expressed to properly specify PFCs [7] . Thus , the defects that we observe in oocyte repolarization when orb activity is compromised during mid-oogenesis are downstream of both the grk signal to the posterior follicle cells and the signal from the PFCs to the germline that induces MTOC disassembly . Three other findings are consistent with this conclusion . First , when PFCs are not properly specified , the posterior MTOC fails to disassemble [7 , 8 , 18] . By contrast , when orb activity is compromised during mid-oogenesis the MTOC dissembles as in wild type . Second , the formation of a non-centrosomal cortical based MT network is initiated along the anterior/lateral margin of the oocyte even in the absence of the PFC signal . This is not true in our experiments; the anterior/lateral MT network is not properly established . Third , in the absence of the PFC signal , Staufen protein , Kinesin-β-gal and osk mRNA concentrate in the center of the oocyte , while bcd mRNA is found not only at the anterior but also at the posterior end of the oocyte . In contrast , in orb mutants , osk and also bcd mRNA accumulate at the anterior of the oocyte , while Kinesin-β-gal is unlocalized . As the posterior MTOC is disassembled , a MT network emanating from the anterior and anterior-lateral cortex of the oocyte is established . The initiation of this non-centrosomal cortical based MT network is mediated by the spectraplakin , Shot , and the minus-end MT binding protein , Patronin [21] . Shot associates with the actin rich anterior and anterior lateral cortex and recruits Patronin . Patronin then nucleates the assembly of the MT network . Nashchekin et al . ( [21] ) have shown that proper polarization of the MT network by Shot and Patronin depends upon the Par protein Par-1 . By an unknown mechanism , Par-1 blocks Shot association with the actin rich cortex . Since Par-1 is enriched around the posterior cortex of the oocyte , this restricts the de novo assembly of MTs to the anterior and anterior-lateral cortex . While Par-1 is required to exclude Shot from the posterior cortex , the de novo assembly of MTs requires Shot association with the anterior and anterior-lateral cortex . This presumably does not happen when aPKC , cdc42 and/or baz are compromised in the germline because the anterior and anterior-lateral cortical actin network is disrupted . Our results place orb upstream of Shot and Patronin and suggest that the defects in oocyte MT repolarization likely arise for several reasons . One would be defects in the localization and functioning of the Par gene products . When orb activity is compromised , the association of the Par protein Par-1 with the posterior and aPKC with the anterior-lateral cortex is disrupted . In the absence of proper cortical association , the cross-regulatory interactions between the anterior and posterior Par proteins would be expected to be ineffective . Also consistent with a role for orb in the functioning of the Par proteins in MT repolarization are genetic interactions between orb and genes encoding the anterior Par proteins , aPKC and cdc42 . orb is weakly haploinsufficient for the grk signaling pathway , and about 5% of the eggs laid by orb343/+ females , are ventralized . This weak haploinsufficiency is enhanced when the orb343 mutation is trans to mutations in either aPKC or cdc42 . For the aPKC null allele , aPKC k06403 , the frequency of ventralized eggs increases to nearly 50% , while about 70% of the eggs laid by females trans-heteozygous for orb343 and cdc424 are ventralized . Moreover , while females heterozygous for either aPKC k06403 or cdc424 alone do not lay ventralized eggs , nearly 40% of the eggs laid by females trans-heterozygous for these two mutations are ventralized . As we found for orb , the localization of the oocyte nucleus to the dorsal anterior corner of the oocyte depends upon cdc42 and aPKC . Leibfried et al . ( [26] ) found that the oocyte nucleus is mispositioned in egg chambers homozygous for cdc424 , while we have shown here that the oocyte nucleus is mispositioned when aPKC is knocked down by RNAi . Moreover , the frequency of mispositioned nuclei in orbmel/orb343 is enhanced when the females are also heterozygous for mutations in aPKC . At least some of the effects of orb on the Par proteins could be direct . Thus , aPKC mRNAs contain CPEs in their 3’UTRs and we have found that aPKC mRNA is bound by Orb protein in ovary extracts . Moreover , the distribution of aPKC mRNA within the oocyte is altered when orb activity is compromised . Interestingly , mRNAs encoding the three other anterior Par proteins , cdc42 , baz , and par-6 also have CPE motifs in their 3’UTRs and are bound by ectopically expressed Orb in tissue culture cells [48] . Thus , the localization and translation of these Par mRNAs could be regulated by or dependent upon orb . In addition , there appears to be a reciprocal relationship between orb and anterior Par proteins . This is suggested by the synergistic genetic interactions between orb and the Par genes encoding aPKC and cdc42 . It also fits with our finding that orb autoregulatory activity and the phosphorylation status of Orb are impacted by aPKC depletion . There are also likely to be indirect effects on the functioning of the Par proteins that in turn perturb the organization of the MT network . For example , Leibfried et al . ( [26] ) have shown that there is a mutually interdependent relationship between the Par proteins and the actin cytoskeleton . They found that Cdc42 localization along the anterior and anterior-lateral cortex of the oocyte depends upon the integrity of the cortical actin network . Conversely , the assembly of the cortical actin network requires cdc42 , aPKC and baz . In fact , one of the more striking phenotypes in orb mutant oocytes is the disorganization of the cortical actin network . As was observed for cdc42 [26] , the disruptions in the actin network are accompanied by the mislocalization of aPKC . Given the interdependence of the Par proteins and the actin network the disruption of the actin cytoskeleton in orb mutants could be due to the misexpression of the Par proteins . However , the Par proteins need not be the only or even the key targets for orb regulation of the actin cytoskeleton . For example , the formation of the cortical actin network during mid-oogenesis depends upon two actin nucleators , capu and spir [61–64] . Mutations in these two genes have a number of phenotypes in common with orb . The actin cytoskeleton is fragmented and this in turn leads to a failure to properly organize the MT network and localize osk and grk mRNAs . Moreover , as has been reported for orb [56] , premature cytoplasmic streaming is observed in capu and spir mutant egg chambers . Like the Par proteins , the mRNAs encoding capu and spir are bound by ectopically expressed Orb in tissue culture cells , and thus could be targets for orb regulation . On other the hand , there are some notable differences . In contrast to orb , aPKC and cdc42 , capu and spir eggs are dorsalized not ventralized . Additionally , Par-1 localization to the posterior and lateral cortex does not appear to depend upon capu or spir [62] whereas it is disrupted in orb mutant chambers . Moreover , the effects of orb on the actin cytoskeleton need not be limited to these proteins . The mRNA encoding the actin effectors Cip4 and WASp have CPEs in their 3’UTRs and are bound by ectopically expressed Orb in tissue culture cells [48] . Defects in the expression of these proteins would interfere with the remodeling of the anterior/anterior-lateral cortical actin cytoskeleton and consequently disrupt Par dependent MT polarization . Finally , orb could also act downstream of the Par proteins . Like cip4 and WASp mRNAs , the mRNAs encoding the MT assembly factors , shot and patronin , have CPEs in their 3’UTRs and are bound by ectopically expressed Orb in tissue culture cells [48] . Insufficient levels of these factors would be expected to slow or block the de novo assembly of MTs along the anterior-lateral cortex . Thus , a plausible idea is that the defects in the repolarization of the MT network when orb is depleted during mid-oogenesis are likely the consequence of the cumulative effects of misregulating mRNAs encoding not only Par proteins but also proteins involved in organizing the actin cytoskeleton and assembling MTs . Because the MT and actin cytoskeleton regulators have interdependent functions , even small perturbations in the abundance of multiple players could lead to wide ranging disruptions in cytoskeletal organization . That mRNA localization/translational regulation might impact the reorganization of the egg chamber after stage 7 at multiple levels is supported by recent studies on egalitarian ( egl ) . Sanghavi et al . ( [67] ) report that knocking down egl just before the MT network in the egg chamber is repolarized induces many of the same phenotypic abnormalities and disruptions in cytoskeletal organization that we have observed when orb activity is compromised during mid-oogenesis . Egl together with the Bicaudal-D ( BicD ) protein loads mRNAs onto a Dynein motors [79–81] . This mRNA cargo complex is responsible for localizing mRNAs in somatic cells and in developing egg chambers . Like Orb , the Egl-BicD cargo complex interacts with many different mRNA species including orb . For this reason , loss of egl activity is likely to have a global impact on mRNA transport and consequently the localized production of a diverse array of factors needed for the reorganization of the oocyte cytoskeleton during mid-oogenesis .
Endogenously tagged Patronin-YFP , YFP-Patronin expressed from a maternal tubulin promoter and Shot-YFP ( [21] ) were gifts from Daniel St Johnston and Dmitry Nashchekin; osk54 , osk84 , stauD3 , stauDf , KZ32 ( Kinesin-β-gal ) are gifts from Elizabeth Gavis; grk2B , grk2E12 , BB142 ( kekkon-lacZ ) are gifts from Trudi Schupbach; aPKC mutant alleles aPKCk06403 , aPKCex55 , aPKCex48 and mattub-GFP-Par-1-N1S are gifts from Yu-Chiun Wang and Eric Wieschaus; cdc421 and cdc424 , aPKC RNAi 35140 , maternal alphaTubulin67C Gal4 ( 7062 and 7063 ) from Bloomington Stock Center . Eggs were collected by placing flies of the appropriate genotype into cups and were kept at 18 degrees and given fresh apple juice and yeast paste plates daily . The eggshell phenotypes were scored starting on day 3 . osk FISH probes were a gift from Shawn Little at University of Pennsylvania [54] . orb FISH probes were ordered from Biosearch Technologies , and orb probes and aPKC-com FISH probes ( from Xu et al . [49] ) were coupled to Atto NHS-Ester 565 or 633 ( Sigma ) and purified using HPLC . Antibodies used were as follows: mouse anti-Orb ( 4H8 , 6H4 ) used 1:30 each , mouse anti-Gurken ( 1D12 ) used 1:20 , mouse anti-β-gal ( 401A ) used 1:10 , mouse anti-Bic-D ( 1B11 , 4C2 ) used 1:20 each from the Developmental Studies Hybridoma Bank; rabbit anti-Cadherin99C used 1:1000 was a gift of Dorothea Godt; mouse monoclonal anti-α-tubulin-FITC ( clone DM1A ) from Sigma; rabbit anti-aPKC ( clone c-20 , sc-216 ) used 1:1000 from Santa Cruz Biotechnology . Wheat germ agglutinin ( Alexa Fluor 633 , Molecular Probes ) , Phalloidin ( Alexa Fluor 546 or 633 , Molecular Probes ) and DAPI ( Molecular Probes ) were used . Secondary antibodies used were goat anti-mouse IgG Alexa 488 , 546 or 647 , goat anti-rabbit Alexa 488 , 546 or 647 ( Molecular Probes ) . Samples were mounted using aqua polymount ( Polysciences ) on slides and visualized on a Leica SP5 or Nikon A1 confocal microscope . Cytoplasmic movements were imaged in live oocytes in halocarbon oil on a Nikon A1 inverted confocal microscope . An image was collected every 5 seconds for at least 2 minutes to visualize cytoplasmic streaming . Mouse anti-Orb ( 4H8 and 6H4 ) or mouse anti- β-gal ( 401A ) were coupled to A/G agarose beads ( Santa Cruz Biotechnology ) by incubating overnight at 4 degrees . 250 females were dissected in ice cold 1xPBS and ovaries were transferred to dry ice while dissecting . RNAsin ( Promega ) was added to ovaries and they were crushed using a plastic pestle to make a paste . The ovary paste was centrifuged at 3000 rpm for 5 minutes at 4 degrees twice , and the supernatant was saved . Half of the supernatant was added to the Orb antibody coupled with beads , and the other half was added to the control antibody beads . CoIP buffer ( [55] ) and RNasin ( Promega ) was added to IPs , which were left to rotate for 3 hours at 4 degrees . The beads were pelleted by centrifugation and washed with coIP buffer 5 times . RNA was released from the beads by adding 10 mM HEPES 1% SDS solution and β-mercaptoethanol , and left in a 65 degree water bath for 15 minutes . Phenol followed by phenol chloroform was used for extraction , and the water phase was ethanol precipitated with glycogen added as a carrier . The pellet was dried and then DNAse ( Promega ) treated . The RNA samples were incubated with oligodT ( IDT ) at 65 degrees for 10 minutes . AMV reverse transcriptase ( Promega ) reactions were set up , and for each IP a control reaction was set up without reverse transcriptase . The samples went through the following program for reverse transcription on a PCR machine: 55 degrees for 1 min , 48 degrees for 30 min , 55 degrees for 15 min , 95 degrees for 5 min , hold at 4 . For quantitative PCR , Power CybrGreen PCR master mix ( Life Technologies ) was used . Each qPCR reaction was done in triplicate and the average CT was used . The control samples without reverse transcriptase were also run to confirm the DNase treatment worked . The amplification of target 3’UTRs from the Orb IP were compared to the amplification from the control IP and normalized to a control ( RPL32 ) to calculate ΔΔCT . For the Western blots to measure levels of β-gal expression ovaries were dissected in PBS and frozen on dry ice . Frozen tissue was crushed with a pestle in SDS buffer with urea , boiled and spun down . The extracts were loaded on a 10% SDS-Page gel . Proteins were transferred to a PVDF membrane and the membrane was cut to blot for β-gal and BEAF . For the phosphorylated Orb isoforms , ovaries from two female flies were dissected in 100 ul of 1X PBS . The ovaries were immediately transferred to 40 ul of 2XSDS buffer ( 100 mM Tris-Cl; 4% SDS , 200 mM DTT and 0 . 2% bromphenol blue ) and boiled . A second set of ovaries were dissected , transferred to the same tube and boiled . The samples were then loaded onto a 7 . 5% SDS polyacrylamide gel . Image J was used to measure the Orb protein upper:lower band ratio . | The specification of polarity axes in the Drosophila egg and embryo depends on the proper organization of the microtubule ( MT ) and actin cytoskeleton during mid-oogenesis . During this period , the MT organizing center at the posterior of the oocyte is disassembled and a MT network is established at the anterior and anterior-lateral cortex of the oocyte . We show that the CPEB translation factor orb plays a critical role in the reorganization of the MT network . orb appears to function at several levels during MT reorganization . orb interacts genetically with genes encoding Par proteins , aPKC and cdc42 , and disrupts the localization of Par-1 and aPKC within the oocyte . orb also plays an important role in organizing the cortical actin cytoskeleton . The defects in the actin cytoskeleton disrupt the cortical association of Shot and Patronin , which are responsible for nucleating the assembly of the anterior MT network . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cell",
"physiology",
"medicine",
"and",
"health",
"sciences",
"reproductive",
"system",
"rna",
"interference",
"messenger",
"rna",
"reproductive",
"physiology",
"cell",
"polarity",
"germ",
"cells",
"oocytes",
"epigenetics",
"cellular",
"structures",
"and",
"organelles"... | 2019 | The CPEB translational regulator, Orb, functions together with Par proteins to polarize the Drosophila oocyte |
Predators of all kinds , be they lions hunting in the Serengeti or fishermen searching for their catch , display various collective strategies . A common strategy is to share information about the location of prey . However , depending on the spatial characteristics and mobility of predators and prey , information sharing can either improve or hinder individual success . Here , our goal is to investigate the interacting effects of space and information sharing on predation efficiency , represented by the expected rate at which prey are found and consumed . We derive a feeding functional response that accounts for both spatio-temporal heterogeneity and communication , and validate this mathematical analysis with a computational agent-based model . This agent-based model has an explicit yet minimal representation of space , as well as information sharing about the location of prey . The analytical model simplifies predator behavior into a few discrete states and one essential trade-off , between the individual benefit of acquiring information and the cost of creating spatial and temporal correlation between predators . Despite the absence of an explicit spatial dimension in these equations , they quantitatively predict the predator consumption rates measured in the agent-based simulations across the explored parameter space . Together , the mathematical analysis and agent-based simulations identify the conditions for when there is a benefit to sharing information , and also when there is a cost .
Predators in numerous systems share information with one another to optimize their individual and collective gains [1] . These information-sharing cliques tend to form predatory packs or groups [2] . So too in social-ecological systems such as fisheries , where fishermen are sometimes known to share information with one another about the location of target species aggregations , and sometimes not [3 , 4] . Critically , the benefits and costs of sharing information depend on the spatial characteristics of the system . For example , fishermen that share information tend to target highly ephemeral and migratory species like salmon [5] . In contrast , sessile and slow moving species tend to be harvested by secretive and even territorial fishermen [6] . Further , information sharing need not be voluntary , and there are numerous examples of predators copying the location of others [e . g . 5] . Hence , quantifying the relationship between the benefits and costs of information sharing and the spatial characteristics of the environments in which predators search for and capture prey , is important if we are to deepen our understanding of group formation and cooperation in social and natural systems alike . An important consequence of space and information sharing is the potential inaccuracy of ecosystem models used to design management policies [4 , 7 , 8] . For instance , many population and ecosystem models used to inform policy on land and in the sea currently ignore or have overly simple parametric representations of predator and prey social and spatial behavior . In these models , it is commonly assumed that per-capita predator consumption depends only on prey abundance . This is reflected in the mathematical functions used to describe predator consumption: Type I ( linear ) , Type II ( saturating ) and Type III ( sigmoidal ) functional responses [9 , 10 , 11] . This is due to the ecological legacy of these functions , having been well explored empirically and mathematically [12 , 13 , 14] . In some limiting cases , these simple functions can accurately represent aggregate feeding rates seen in nature [15] . However , all these feeding functions assume that the rate at which predators encounter ( if not always consume ) prey is linearly proportional to the density of the prey [2 , 16] . In more complex spatial settings and when information sharing occurs , encounter rates are non-linear and as a consequence , these models will produce inaccurate individual and group feeding rates . This is acknowledged by ecosystem modelers themselves [17] , and yet we remain limited in our ability to model predator group behavior in such contexts . Agent-based models of predators searching for prey [e . g . 13 , 18 , 19] have been used to describe encounter rates , as they emerge from more realistic predator movement rules . Indeed , multiple models exist for predator search patterns , such as random [e . g . 20 , 21] , Lévy [e . g . 22] and correlated random walks [e . g . 23] to name a few . These studies deepen our understanding of the role space plays in the search process , but are limited in terms of accounting for both consumption of prey once found , and the impact of information sharing on the spatial distribution of predators on a given landscape . Together , both factors define the balance between the benefits and costs of information sharing: information sharing can reduce the time it takes to find prey , but it also comes at a cost , as prey is shared too . Here , our goal is to assess the interacting effects of spatio-temporal heterogeneity and information sharing on predator consumption rates . In order to do so , we have developed a general mathematical model of predator foraging , accounting for space implicitly using a few key parameters . These parameters are timescales for search , consumption and prey mobility , which can be measured or computed independently for a given spatial setting , and from which we derive the benefits and costs of exploration , exploitation and information sharing . We note that these results do not address the strategic choices of predators or the evolution of cooperation [24] . Instead , we provide an extended form of the feeding functional response , incorporating the level of information sharing and spatial characteristics of both prey and predator . This allows us to compute the payoff , in terms of foraging efficiency , of a given information sharing behavior at the group level and for the individual . These payoffs are the foundation on which future evolutionary analyses could be performed , building off works on cooperation that use minimal representations of space [e . g . 25 , 26 , 27] . To complement the mathematical theory , we also developed a spatially explicit agent-based model ( ABM ) of predators and their prey , which we used to validate the analytical model and derive its abstract parameters from more intuitive individual processes . Even though the dynamics in the analytical model are encoded in a few behavioral states and a single spatial correlation metric , they quantitatively predict predator consumption rates measured in the ABM across the entire parameter space that we explored . Furthermore , and critically , very different simulation settings result in equivalent predation efficiencies , as long as they are characterized by the same key timescales . This means that the ( spatial ) assumptions of the ABM have limited impact , and that the mathematical analysis is far more general , being valid for a large range of predator-prey / consumer-resource systems , as modeled by other ( perhaps more complex ) ABMs or measured with empirical data .
The first step to developing a mathematical feeding function that accounts for both space and information sharing is to view predator and prey interactions in terms of the rates at which predator behavioral states change . For example , let s be the fraction of a predator population that is moving in search of its prey , which is organized in distinct patches , and h be the fraction of predators currently “harvesting” or consuming prey . In the simple case of independent predators , there are no other states and s = 1 − h . Equivalently , since the population is for now assumed to be homogeneous , s and h can be seen as the fraction of time spent searching and consuming by a single predator . Furthermore , let us define τs the expected time taken for an individual predator to find a prey-patch ( otherwise known as the first-passage time ) , and τh the patch handling time or expected time to consume all prey in a patch . The rate at which a searching predator finds a prey patch is then 1/τs , and the rate at which a feeding predator returns to searching is 1/τh . This simple behavioral state change model is depicted in Fig 1A . At steady state , the fluxes between behavioral states are equal: s 1 τ s = h 1 τ h . ( 1 ) Furthermore , the consumption rate H ( in units of prey per time ) averaged over a foraging period will be proportional to h , the fraction of time spent consuming rather than searching . Let us denote by H* the maximal value attained when s = 0 , a case where prey is abundant everywhere and can be found instantly . Using s + h = 1 , we can write the functional response as H H * = h h + s = 1 1 + s h = 1 1 + τ s τ h . ( 2 ) Clearly , the feeding rate of a predator is inversely proportional to τs/τh , how long it takes to find prey patches relative to how long it takes to consume them . If we further assume that the rate at which predators encounter prey patches ( 1/τs ) is proportional to the density of prey , and there is no influence of predator group behavior , the expression on the far right takes the form of a classical Type II functional response . Alternatively , if the encounter rate is convex in prey density , rather than linear , then a Type III function is found instead . In order to explore information sharing amongst predators , we modify the simple behavioral state model introduced above ( Fig 1A ) . First , let us redefine the two behavioral states s and h above as searching alone and consuming a patch found by oneself . Likewise , τs and τh become the typical search time and consumption time for a predator alone , not accounting for information sharing and group dynamics yet . Furthermore , we add τl the typical time during which a prey patch can be exploited before it moves . This introduces an element of landscape stochasticity or change , representing the dynamic nature of prey distributions found in a number ecological systems on land , in lakes and in the sea [28] . These three quantities—τs , τh and τl—play a key role in our analysis , as they can be independently measured or computed from the spatial characteristics of the individual predator and prey . Thus , they interface between the generic model presented here and any specific description of how predators and prey move and interact . In the next section , we develop an agent-based model to demonstrate how these key timescales can be computed from measurable quantities , albeit in a simplified abstraction of reality . We further expand on these key timescales in the discussion , describing their real world analogues for a range of natural and human systems . In addition , there are two more parameters that control the benefits and costs of information sharing: N the number of predators foraging in the same area , and λ their propensity for sharing information , which we assume here to be identical for all predators . The consequences of relaxing this last constraint is explored in the Supplementary Online Information ( SOI ) section “Agent Level Mathematics” . The process of searching for prey patches remains solitary , but predators can now broadcast the location of a prey patch when they find one . This reflects group predator behavior found in social insects such as honey bees [29] as well as scavenger mammals such as hyenas [30] . We must then consider two new behavioral states: m which is the fraction of predators moving toward a patch whose location has been broadcast , and b the fraction of “bound” predators having reached the patch and are consuming prey with their caller . Clearly , the foraging efficiency is now the total fraction of time spent consuming prey whether one has found a patch or been called to it: HH*=h+b ( 3 ) Finally , we define the rates of change between these behavioral states: the rate at which lone searchers find and start consuming prey Ws → h , the rate at which lone searchers get a call from and start moving towards another predator Ws → m , the rate at which predators reach their caller Wm → b , and the rate at which all these predators revert back to searching alone: Wh → s , Wm → s , Wb → s . This new behavioral state model is depicted in Fig 1B and its associated parameters are listed in Table 1 . What remains is to derive formulae for the different rates of behavioral state change . Starting with the rate at which lone searchers encounter a new prey patch: Ws→h=1τs ( 4 ) Next is the rate at which lone searchers receive a call and start to move toward a patch: Ws→m=NλsWs→h . ( 5 ) This rate is proportional to N×s the number of searchers , and to Ws → h the rate at which one of these searchers will discover a patch and switch from state s to h . As mentioned above , the factor λ quantifies the strength of social interactions , here conceptualized as the probability that a predator broadcasts information upon discovery of a prey patch . Next , the rate at which predators reach their caller is the inverse of the expected travel time between two predators τd: Wm→b=1τd , ( 6 ) where , as a first approximation , τd can be taken as a constant , for example fixed to unity ( meaning that other timescales are measured in units of τd ) . In the SOI section “Distance between Predators” , we discuss a more elaborate derivation of τd which is necessary only for precise quantitative agreement with the ABM simulations , or with empirical data . Finally , we define WL , the rate at which a predator is interrupted while exploiting a patch , either because the latter is depleted or has moved . This rate WL accounts for the reversion from all other states to the lone searching state: Wh→s=Wm→s=Wb→s=WL . ( 7 ) To compute it , we need two timescales . First , τl the expected time between a patch being discovered and it moving away . Second , τh/np the expected time it takes to deplete the patch given a number np of predators consuming together ( we assume here a simple inverse proportionality , but this is easily adapted to represent interference or cooperation between predators on a patch , without qualitatively affecting our results ) . These two timescales are combined in a Poisson process approximation: assuming that the probability of the patch not moving in the time interval [0 , t] is exp ( −t/τl ) , and its probability of not being depleted yet is exp ( −tnp/τh ) , then the probability of both conditions being verified is the product exp ( −t/τl−tnp/τh ) , but it is also exp ( −WL t ) by definition . Hence , we can write: WL=1τl+npτh . ( 8 ) In the previous paragraph , we introduced an important new quantity: np , the number of predators consuming at the same patch simultaneously . Deriving np is in fact the most intricate part of this analysis , as this quantity encodes the main relevant spatial and temporal correlations in the system , and thus represents non-mean-field dynamics in this otherwise space-less model . One derivation is given in the SOI section “Refined Occupancy Approximation” , but a basic intuition can be obtained from the following approximations , which hold only in simple limits: np≈{ Nλ→1 , ( 1+bh ) λ≪1 . ( 9 ) If λ = 1 , we expect all the predators to be fully correlated in space and time , and form a pack that always consumes the same prey . In such a situation np ≈ N . As λ becomes smaller however , time correlations become negligible and it is possible to use the time-averaged values b and h in np ≈ 1 + b/h , where b/h gives an estimate of the number of predators consuming at the same patch in addition to its finder ( assuming a patch is not independently found by multiple searchers ) . Given these formulae for the rates of change of the predator behavioral state occupation probabilities , it is possible to describe the evolution of these states in the predator population over time . In the long time limit , it converges toward a stationary point , where all the out- and in-fluxes between states are balanced: s ( Ws→h+Ws→m ) =h Wh→s+m Wm→s+b Wb→s ( 10 ) h W h → s = s W s → h ( 11 ) m W m → s + W m → b = s W s → m ( 12 ) b W b → s = m W m → b ( 13 ) This system can be reduced to a single equation over s , the fraction of predators searching alone: 1τs ( Nλs2+s ) = ( 1−s ) WL , ( 14 ) where WL is a function WL ( τl , τh , s ) and all other terms are constant . Finally , the consumption rate for a predator in the group can itself be expressed using only s: H ( τs , τh , τl , N , λ ) H*=h+b= ( 1τs+λNs1+τdWL ( τl , τh , s ) ) 1WL ( τl , τh , s ) s ( 15 ) The term 1/τs describes the baseline ( single predator ) success rate: it decreases with τs , which can be thought of as the “cost” or difficulty of exploration , and increases with prey abundance ( although it is a non-monotonic function of prey patchiness at a constant level of coverage , as detailed in the SOI section “Population Level Mathematics” ) . To this individual baseline , the second term within the large parenthesis describes the benefit of information sharing , and vanishes when λ = 0 . The whole expression is modulated by 1/WL , the expected time during which a patch is available for consumption , and thus the effective “value” of each patch . This term increases with τl and τh , that is for less mobile or richer patches . It is also a strictly decreasing function of λ the amount of information sharing , as a result of faster depletion , which is part of the cost of sharing . However , for small τl ( very mobile prey ) this variation is negligible , since mobility overtakes depletion as the main cause for returning to the searching-alone state . In other words , prey mobility discounts the cost of information sharing . In conclusion , solving Eq ( 14 ) allows to us compute H in Eq ( 15 ) , the consumption rate of any predator in the group given its environment , encoded by the key timescales τs , τh and τl , the number of predators N and their propensity toward information sharing λ . It is easily solvable numerically , and thus is a fast alternative to agent-based simulations . However , due to the complexity of the full expression of np , the expected number of predators consuming prey at the same patch , it does not have an explicit solution . In our SOI section “Solvable Limits” , we discuss certain limiting cases where simplifying assumptions can be made to obtain explicit results . While these can contribute some insight into parts of the model , we show that none of these partial solutions can reflect the full phenomenology of this model or our ABM simulations below . Finally , in SOI section “Agent Level Mathematics” we show that this model is naturally extended to compute individual consumption rates in the case of heterogeneous agents , especially when they differ by their communication strategy λ .
In addition to the general behavioral state model described above , we developed a spatially explicit , computational agent-based model ( ABM ) of predators and their prey . For a better qualitative understanding of how the ABM works , we refer the reader to the schematic in Fig 2 , to our Supplementary Online Information where several movies show different simulation experiments ( SOI: Movies ) , and to Table 2 where the parameters of the ABM are listed . This ABM allowed us to relate the abstract timescales of the mathematical analysis above ( Table 1 ) to more concrete processes and properties that might be observed and measured in real-world systems ( see Fig 3 ) . The ABM operates on a 2D landscape with periodic boundaries . On this landscape are a given number Ns of circular prey patches , a fish school for example , with a defined radius σ . Within these patches are a number of uniformly distributed prey items , Ni , that could for instance represent individual fish within a school . The radius and number of prey items per patch is the same across patches . Prey mobility is important for our purposes only inasmuch as it limits the time that a predator can spend exploiting the same patch; we could let patches move so that predators could randomly lose their track , but for simplicity we simply let patches disappear with probability per unit time rl . However , we wish to conserve the total fraction of the landscape occupied by patches , and therefore we introduce a new patch at a random location every time one vanishes . Thus , the landscape change timescale is simply: τl=1rl . ( 16 ) A number N of predators is also found on this landscape , searching for prey patches . When a patch is found , they consume prey items within at rate rq . The expected patch handling timescale is: τh=Nirq , ( 17 ) We note that rq = H* the maximal consumption rate discussed since Eq ( 2 ) , since the average consumption rate over a simulation run is clearly at most rq . For increased realism , rq could vary during consumption: for instance , a patch might be quickly depleted at first , then become harder to exploit . In that case , it could be advantageous for predators to leave a patch after a time τh that does not correspond to full depletion ( as above ) but instead optimizes their consumption rate . As this complicates both the simulations and the analysis , we leave this possibility to further inquiry . In many systems , predators search for prey by alternating between fast travel between prey habitats , and intensive search ( slow movement ) at these habitats [31] . This is exhibited throughout nature , where numerous predator species switch between strict transiting/migration and feeding behaviors , and in social-ecological systems too , for example fishermen searching for ephemeral target-catch , such mobile pelagic species such as anchovy [32] . Multiple models exist to capture these predator search patterns , such as random [e . g . 20 , 21] , Lévy [e . g . 22] and correlated random walks [e . g . 23] . The qualitative feature shared by all these various models is an alternation between long-range movement and focused search within a local area . The simplest and most tractable model that retains this property is intermittent search [19]: predators move ballistically—in a straight line—at a constant speed v , with random direction changes made with probability per unit time ( turning rate ) rp . In-between two ballistic segments representing long-range movement , a predator spends one time-step exploring its surroundings to find prey , as represented by a sensing radius σs . If there is no prey patch within that radius , it moves ballistically in the new direction . These three parameters , together with prey patch number Ns and radius σ , determine the search ( first-passage ) time τs . To compute it , we build off work developed by Benichou et al . 2011 [19] , who mathematically analyzed this type of search for only one prey patch and one predator . We provide a simple extension to account for multiple prey patches , which is presented in the SOI section “Intermittent Search” . Predators can share information with one another about the location of prey patches: if two predators have a social tie , and one has found a prey patch but the other has not , information is shared between the predators with probability equal to λ the strength of the tie . In this situation , the predator receiving information moves towards the other predator . The social network is recorded in the form of a symmetric matrix of tie strengths . For simplicity , we mainly discuss two network structures: a fully connected graph where all ties have strength λ ∈ [0 , 1] , or a binary matrix where 1’s identify a social tie and 0’s the absence of a social tie . However , the model extends to arbitrary networks . Here , the predators’ communication range is assumed to encompass the whole landscape , and N is therefore the number of predators within communication distance of each other , whether randomly , due to coordination , or to extraneous social factors such as pack size . If the behavioral state model described in the previous section is correct , the output of the ABM—namely the expected catch rate of a predator—should depend on its parameters only through the timescales which we derived , as well as λ and N . This we indeed show in SOI section “Validation of Analytical Results” , meaning that the parameter space has a much lower effective dimension than it first appears to have . While τl and τh have simple expressions in terms of the ABM parameters , τs is more complicated , reflecting the fact that our implementation puts an emphasis on details of the search process . Among the parameters influencing τs , the most obvious one is the sensing radius σs , which has an unambiguous influence on search difficulty , and affects nothing else . Thus , we generally explore the parameter space of the ABM by using σs as a control variable , fixing the total prey coverage Ns πσ2 and predator velocity v . Furthermore , in all simulations we use the turning rate rp that minimizes τs for a single predator . We acknowledge that in nature , there is no guarantee that predators will have such optimal spatial behavior , and that the turning rate rp will likely evolve together with the social structure of the predator population ( encoded here by N and λ ) . However , we do not explore here the consequences of changing rp , and simply use the turning rate that minimizes τs as a way to be consistent across simulation experiments , and to speed them up .
We implemented the computational ABM using the Julia language ( www . julialang . org ) , running several simulation experiments . Initially , these were used to vet the mathematical model , as presented in the SOI section “Validation of Analytical Results” and ultimately we found strong qualitative and quantitative agreement between the results of the ABM and the mathematical model . This success was especially important for subsequent simulation experiments , as it showed that the parameter space of the ABM could be efficiently explored along only three dimensions: those present in the mathematical model , i . e . first-passage time τs , patch handling time τh and landscape change timescale τl . First , we performed a number of two-predator simulations , with the objective of measuring how the dimensionless ratio H/H* varies with predator information sharing . H* can be thought of as the maximal consumption rate obtained if prey-patch encounters are instantaneous and as a consequence predators consume prey constantly . Hence H/H* provides an estimate of the “efficiency” of the predators given their level of information sharing . We also mapped out the full functional response of the predators by measuring their consumption rates when systematically varying the landscape change timescale τl , through changes in the rate at which prey patches move rl; the patch handling time τh , through changes in the rate at which predators consume prey rq; and the first-passage time τs by varying σs , the predator sensing radius . This numerical implementation the ABM was also used to explore situations that involved more predators . However , these simulations were extremely computational demanding . Hence , in order to proceed and investigate situations with large numbers of predators , we used the mathematical model instead . We were then able to compute consumption rates for arbitrarily large numbers of predators . This allowed us to study the role of group size in maximizing individual consumption rate . This can be done in a number of ways , for example by varying N to find , as a function of spatial parameters , the optimal number of predators foraging within communication range of each other . Here , we chose to fix N and find the optimal group size within that population size . In doing so , we answered the question of whether predators that are already foraging in the same space should do so collaboratively . In these simulation experiments , we chose the total number of predators N = 30 , and arranged predators into groups of various size , defined using random partitions . Predators within a group were assumed to share all information . Using an agent-level extension of the analytical model ( see SOI section “Agent-level Mathematics” ) , we then computed the expected consumption rate of any one individual predator , as a function of the size of the group it belonged to , over a range of environments defined by the three key timescales τh , τl and τs . Doing this numerous times , changing the distribution of group sizes at each iteration , allowed us to calculate optimal group sizes for any environment . This produced results qualitatively comparable to other simulation experiments , for example when dividing the entire population into equally sized groups , or when having social ties between all agents and varying λ from 0 to 1 .
The ABM was first used to explore equilibrium encounter and consumption rates with only two predators in the system , for a range of parameter combinations . The parameter space of the ABM has three essential dimensions—τs , τl and τh—which can be effectively explored by holding one constant , and systematically exploring the other two . To give a clearer intuitive picture of model results , we present below two choices for the parameters to hold constant , although they can formally be made equivalent . First , we explored prey-handling and first-passage times normalized by a constant landscape mobility , τh/τl and τs/τl respectively . Unsurprisingly , we find that encounter rates ( Fig . AA in S1 text ) diminish with increasing first-passage times , but they can also show some sensitivity to τh . Normalized consumption rates H/H*—the foraging efficiency—are largest when first-passage times are small , in other words searching for prey is easy , and handling times are high , in other words exploitation is not interrupted by depletion ( Fig . AB in S1 text ) . Second , we explored various prey handling and landscape change timescales normalized by a constant first-passage time ( representing a constant difficulty to find prey ) , τh/τs and τl/τs respectively ( Fig 2 ) . Encounter rates ( Fig 2A ) are constant unless the landscape timescale τl becomes smaller than τs: this highlights the fact that for high enough prey mobility , encounters do not depend on the predator’s search process , as even a static predator is likely to encounter prey as they move . As for consumption rates ( Fig 2B ) , the symmetry between τh and τl is made apparent: both timescales limit how long a patch can be exploited , and therefore only the smaller of the two plays a significant role . This dependence of the consumption rate on predator and prey mobility and group behavior—a generalization of functional response accounting for more than prey density—is well reproduced by the analytical model as seen in Fig . C . Throughout the entire parameter space that we explored , information sharing always increased predator encounter rates . The costs of information sharing hinge entirely on having to share prey , which only affects consumption rates . Indeed , information sharing has a highly variable effect on consumption rates , depending on the environment . In the case where the landscape change timescale τl is varying with the handling time τh ( both normalized by the a constant first-passage time τs:Fig 4A ) , the value of information is greatest when handling times are long and the landscape change time scale is short . Conversely , the value of information diminishes as handling times decrease and the landscape change time scale increases . Importantly , there is a clear demarcation—a line where the value of information sharing is zero—between environments where it is beneficial to share information and when it is not . In the case where the first-passage time τs varies with the handling time τh ( both normalized by a constant landscape change timescale τl ) , we again see the presence of a distinct line of zero value ( Fig 4B ) . Here , there is value to information when handling times and first-passage times are long . and the value of information is least when the handling and first-passage time is short . In both parameter spaces ( Fig 4A & 4B ) our results are intuitive , and highlight that predators only value information if its benefits are high ( search is long ) and its cost is discounted by prey mobility: if targets move away before the predators can deplete them together , then there is no penalty to exploiting the same patch . The optimal group size experiments produced expected consumption rates for individual predators , as a function of group size . Depending on the environment , as defined by the different key timescales , this relationship can be concave , where there is a single optimal number of social ties to have , or convex , where there may be two equally good group sizes to be in . Consider Fig 5A , where three different possible group size curves are shown . In black is an environment in which the curve is convex , and both being alone or operating as one large group are more attractive than anything in-between . As a consequence , despite individual search being the global optimum , fully collective action is also a local optimum . In contrast , the orange curve is concave , revealing an environment in which it is always best to be in a group of intermediate size . Interestingly , other shapes are present , such as the sinuous grey curve , in which there is one global maximum as well as one global minimum , a group size that always performs the poorest . We calculated the curvature of these relationships ( defined as the average of y″/ ( 1 + y′2 ) 1 . 5 , where y is the consumption rate as a function of group size ) over the τh , τl and τs spaces ( Fig 5B and 5C ) . In both spaces , there is a clear demarcation between environments in which there is a single optimal group size ( negative curvature , blue regions ) and those where individual and fully collective action both lead to consumption rates that are far greater than found at intermediate levels of information sharing ( positive curvature , red regions ) . The line separating these regions in the parameter space is close to the one found in the previous figure ( Fig 4 ) between regions of positive and negative value of information in the two-predator simulations . This is striking because it implies that even when information sharing is suboptimal ( for example , in the top-left corner of Fig 4A ) , predators could exhibit high sharing behavior , and not be able to cross over to optimal individual search as intermediate situations are less attractive . Optimal group sizes ( leading to maximum expected consumption rates ) were calculated from these curves and are shown for the τh , τl and τs spaces in Fig 6A and 6B . For the τl and τh space , both normalized by a constant τs , we find that the optimal group size is highest when handling times are high and landscape change timescales are low ( Fig 6A ) . Optimal group sizes then get smaller as handling times decreases and the landscape change timescale increases . However , we find that when handling times are very short and landscape change times are very long , there is possible bistability , as evidenced by the presence of extremely large optimal group sizes ( Fig 6A , red blob in the top-left ) . We observe similar qualitative features in the τs and τh space , both normalized by a constant τl ( Fig 6B ) . In general , optimal group sizes are largest which handling times are large , and are relatively insensitive to changes in the search timescale . The jump to one large super group occurs at intermediate handling times ( Fig 6B , the red blob to the left ) , and at intermediate to large first-passage times . These plateaus of high information sharing occur when the relationship between consumption rate and group size is at its most convex ( reflected in Fig 5B & 5C ) and it is almost equally good to search either alone or as one large group .
In summary , we have developed a mathematical model of predators searching for and consuming prey , accounting for spatio-temporal heterogeneity and information sharing . The result is a generalized functional response that accounts not only for the density of prey , but also its patchiness and mobility , as well as the number of predators and their behavior in terms of information sharing . We have identified that these factors shape the consumption rate through three key timescales: τl the timescale over which the prey landscape changes , τs the timescale over which prey patches are found when searching alone , and τh the timescale of exploitation of a patch by one predator . These three timescales control the dynamics of a spatially implicit model representing the behavioral states of the predators . The last critical part of this model was np , the expected number of predators present simultaneously on a patch: it is only through this quantity that spatial and temporal correlations had any significant impact on the output of the model , namely the predators’ consumption rate . Therefore , while the behavioral state abstraction of spatial dynamics is qualitatively robust , for any quantitative agreement to hold , there has to be a satisfactory approximation for np . In addition to this mathematical analysis , we developed a computational agent-based model of predators and their prey , accounting for space explicitly and providing a more concrete set of processes . We used both the ABM and the mathematical model to explore a wide swath of parameter space and identified the payoff , either positive or negative , of information sharing for a range of environments . Starting with two predators only , we found that information sharing always improves encounter rates , but reduces consumption rates if prey have low mobility . For a larger number of predators , we found that there is an optimal number ( or intensity ) of social ties that maximizes consumption rates , going from full collaboration to individual search as prey mobility decreases . However , with low-mobility prey , we found that hunting alone and hunting as one super-group can both be better than intermediate levels of information sharing . Due to the positive curvature of the relationship between individual consumption rates and group-size in that case , selection for optimal group size could exhibit bistability , with full information sharing occurring in regions where individual search was equally or more efficient . A key step in our analysis was to connect the abstract parameters of the mathematical model to those of the ABM . This allowed us to show that the mathematical predictions made only from the key timescales , matched those of the spatially explicit simulations . We acknowledge that the ABM is still an abstraction of reality , but this choice was necessary in order to compute the key timescales that are central to our analysis . Indeed , we did so using the ABM parameters representing prey density , patchiness and mobility , and predator search patterns , and showed that these estimates were enough to guarantee qualitative and quantitative agreement between the numerical and mathematical results . Within the range of situations covered by our simulation model , we thus found that many of the spatially relevant details had limited impact , as very distinct configurations would lead to identical values for the key timescales and as a consequence predation efficiency . This highlights the complementarity between the two models: the ABM helps give meaning to the few variables entering the mathematical model , while the latter helps predict the consequences of changing simulation parameters or even rules . Certainly , many changes to the ABM would translate to different expressions for the key timescales—for instance , τs would be affected by more realistic predator or prey motion . However , these more complex simulations would not change how the key timescales then predict predation efficiency . Hence , the approach developed here is general and can be readily extended to more complex ABMs , better representing specific biological , ecological and social systems . The key time-scales could also be derived from empirically measured quantities . For instance , the timescale over which a predator consumes a prey patch τh—the handling time—is a well known quantity in ecology and can be measured for natural and social-ecological systems alike . It is the reciprocal of the rate at which predators consume prey once encountered , which is typically thought of as the time taken to catch , consume and digest prey [33 , 34] . Admittedly difficult to measure in the field , there are numerous laboratory estimates of these values for many species [35] . In social-ecological systems , this quantity is possibly easier to measure . For example , for a fishery this value is the expected time it would take for a fishing vessel to either catch an entire fish school or fill its hull [36] . Similarly , the timescale over which prey patches are encountered by a lone predator , the landscape change timescale and the typical distance between predators and their travel speed are all measurable quantities in both natural and social-ecological systems [13 , 14] . While the general mathematical model operates at the level of the population , it can be used to derive an agent-level analytical model , allowing us to compare and contrast results directly with the numerical ABM . The derivation of the agent-level mathematics is shown in the SOI section “Agent Level Mathematics” , and these equations provide further intuition about the impacts of information sharing . Indeed , these agent-level equations allow us to address questions of behavioral adaptation or evolution . In exploratory analyses that we performed on predators adapting their social networks , the numerical ABM was far too slow in estimating Evolutionarily Stable Strategies ( ESS ) . In contrast the agent-level mathematics allowed us to compute these ESSs orders of magnitude faster . However , questions of social foraging and evolved cooperation are beyond our scope here , as , in addition to the agent-level mathematics , these analyses require assumptions about how behaviors are selected for [24 , 27] . We do not take this next step , and instead simply present our mathematical formulae for the pay-offs of information sharing in different spatial environments . In all our analyses we have modeled scramble competition , the sharing of prey by predators at the same patch , which is just one way in which predators interact [37] . In nature , there are many other possible interactions . For example , two predators consuming prey from the same patch could interfere with one another , diminishing the rate at which prey is consumed [16] . The converse can happen , where per-predator prey consumption rates increase with predator density , for example when it requires multiple predators to catch prey , such as lions and gazelle [2] . So too for humans , for example squid fishers work together by shining lights from their boats into the water [38] . Squid are attracted to the light , and as a consequence , there is a positive relationship between the number of fishers ( lights ) and consumption rates . These different forms of within-patch interactions by predators can be factored into our general mathematical model in the τh and np terms . For example , if predators interfere with one another as they capture and consume prey , then τh will increase . This will have an impact on np , the expected number of predators consuming prey from a patch , given a certain level of information sharing . We have also focused on how consumption rates can be maximized through information sharing . But for many systems , the variance in consumption through time is important too . For example , in social-ecological systems , subsistence hunters and fishers are less concerned with maximizing the total amount of money or food they gain . Rather , they are often concerned with avoiding a prolonged state of poverty or hunger [39] . One might assume , then , that information sharing would be a boon to these kinds of predators . However , as we have shown , large levels of information sharing can lead to spatial and temporal correlation in predators ( i . e . roaming around as one pack ) . It is precisely under these social conditions , for certain environments , that predators experience high variance in consumption rate . As a consequence , if minimizing the variance in consumption rates is the objective , then full sharing will not necessarily be selected for . Indeed , human-predators avoid spatial and temporal correlations by , instead of simply sharing information , developing profit- and/or risk-sharing institutions [40] to minimize the variance in consumption . In our modeling framework we have also assumed that the abundance of prey , either in terms of the number of patches or the number of prey per patch , is stationary and independent of the predators . This reflects systems where the scale at which the predators operate is smaller than that of the prey population . In other words , it is as if our domain is embedded in a larger area describing the dynamics of the prey . However , there are many systems where the scale of the predator population is similar to those of the prey , and as a consequence , predators can have a large impact on coupled demographics [e . g . 41] . This can be accounted for in our general mathematical model , specifically Eq 15 . For example , consider a situation where resource depletion diminishes the number of prey patches , while the number of prey per patch remains constant . In such a case , the time between prey patch encounters τs will increase . Following Eq 15 , this naturally leads to a decrease in consumption rate . Similarly for the situation when the number of patches remains constant , but the number of prey per patch diminishes , then the patch handling time τh decreases . Again , following Eq 15 , this leads to an increase in the rate at which predators revert to the searching alone state WL , and hence an overall decrease in consumption rate . However , as prey abundance change , so do the incentives to share information . In this case , there are coupled dynamics between the behavior of the predator and the abundance of the prey , with levels of information sharing continually changing as prey abundances change themselves . However , as we have not addressed the ( selective ) mechanisms behind behavioral change , we do not go any further in studying such a coupled system . Throughout this paper we have described information sharing in the context of predators and prey in ecological systems , and sometimes in the context of social-ecological systems such as fisheries . However , our results should hold for any system where one actor may benefit from the findings of another . This could be the extraction of natural resources such as oil and minerals by firms , or it could even describe purely social systems , such as dating or finance . The key to linking our results to these other systems , is to identify the analogues to the main dimensions which control the value of information: τh , τl and τs . Given these quantities , it is possible to hypothesize about the benefits and costs of working together . It is important to note also that all our results could be posed entirely in terms of systems where predators might actively try to conceal their private information . Ultimately , whether information is shared or not , understanding the feedbacks between predator-prey spatial dynamics , and their social preferences , is essential to improving our management of social and ecological systems [8] . | When should we work together and when should we work alone ? This question is central to our efforts to understand social and ecological systems alike , from lions hunting in the Serengeti to fishermen searching for their catch . Here , we develop a mathematical modeling framework to identify the essential spatial factors controlling the benefits and costs of sharing information . Our approach marries computation with mathematical analysis , and our results highlight that it is only under certain spatial conditions that information sharing is a useful cooperative strategy . Notably , we find conditions for which fully collective and fully individual search are both attractive . | [
"Abstract",
"Introduction",
"Behavioral",
"Modeling",
"of",
"Predators",
"Agent-Based",
"Model",
"Simulation",
"Experiments",
"Results",
"Discussion"
] | [
"ecology",
"and",
"environmental",
"sciences",
"predator-prey",
"dynamics",
"population",
"dynamics",
"mathematical",
"models",
"animal",
"signaling",
"and",
"communication",
"simulation",
"and",
"modeling",
"systems",
"science",
"mathematics",
"animal",
"behavior",
"popu... | 2016 | The Spatial Dynamics of Predators and the Benefits and Costs of Sharing Information |
As a major component of ideal plant architecture , leaf angle especially flag leaf angle ( FLA ) makes a large contribution to grain yield in rice . We utilized a worldwide germplasm collection to elucidate the genetic basis of FLA that would be helpful for molecular design breeding in rice . Genome-wide association studies ( GWAS ) identified a total of 40 and 32 QTLs for FLA in Wuhan and Hainan , respectively . Eight QTLs were commonly detected in both conditions . Of these , 2 and 3 QTLs were identified in the indica and japonica subpopulations , respectively . In addition , the candidates of 5 FLA QTLs were verified by haplotype-level association analysis . These results indicate diverse genetic bases for FLA between the indica and japonica subpopulations . Three candidates , OsbHLH153 , OsbHLH173 and OsbHLH174 , quickly responded to BR and IAA involved in plant architecture except for OsbHLH173 , whose expression level was too low to be detected; their overexpression in plants increased rice leaf angle . Together with previous studies , it was concluded that all 6 members in bHLH subfamily 16 had the conserved function in regulating FLA in rice . A comparison with our previous GWAS for tiller angle ( TA ) showed only one QTL had pleiotropic effects on FLA and TA , which explained low similarity of the genetic basis between FLA and TA . An ideal plant architecture is expected to be efficiently developed by combining favorable alleles for FLA from indica with favorable alleles for TA from japonica by inter-subspecies hybridization .
Leaf angle is the inclination between leaf and stem , which is an important agronomic trait that attracts attention . Erect leaves maximize carbon gain by optimizing the interception of photosynthetically active radiation for canopy photosynthesis and by mitigating heat stress induced by excess infrared radiation [1–3] . Crops with erect leaves can be grown in an increasing plant density without compensation by the photosynthesis rate , which consequently increases grain yield . Therefore , leaf erectness as one of the components of ideal plant architecture has been a breeding target for several decades [4–7] . In addition , the more upright leaves also improve the accumulation of leaf nitrogen for grain filling in rice [8] . In addition to breeders , scientists in plant developmental biology have paid much attention to the mechanism of leaf angle formation . The molecular mechanisms in leaf angle have differed among various reports , but there is a common opinion that phytohormone synergism is the key regulator of leaf angle . Endogenous hormones , especially brassinosteroids ( BRs ) , play important roles in controlling rice leaf angle by promoting the growth of cells on the adaxial side of the lamina joint [9 , 10] . Most of the rice leaf angle-related genes with regard to BR biosynthesis and BR signaling or that are otherwise BR related have been identified , such as OsDWARF4 , D2/CYP90D2 , OsBRI1 and OsBZR1 [11–14] . OsIAA1 and OsARF19 control the leaf angle by responding to auxin and BR hormones [15 , 16] . OsSPY and D1/RGA1 are two genes in the GA signaling pathway that regulate leaf angle in a BR-GA crosstalk manner in rice [17 , 18] . Increased Leaf Angle 1 ( ILA1 ) , different from the abovementioned leaf angle-related genes , regulates mechanical tissue formation in the rice leaf lamina joint [19] . CYCU4;1 promotes the proliferation of sclerenchyma cells on the abaxial side of the lamina joints to affect rice leaf erectness through the BR signaling pathway [20] . oslg1 is the T-DNA insertion mutant of OsLIGULELESS1 ( OsLG1 ) , with erect leaves from the loss of the lamina joint structure [21] . Therefore , there are many findings on the development of leaf angle of rice that are helpful for understanding its molecular mechanism . However , many of these studies are based on the reverse genetic approach . While understanding the natural variation in rice leaf angle is still limited , which is more important for breeding rice varieties with ideal plant architecture . Since the 1990s , researchers have explored several leaf angle-related quantitative trait loci ( QTLs ) with bi-parental mapping populations [22–24] . Recently , genome-wide association study ( GWAS ) has become a popular approach for QTL mapping in crops due to its strong power and high-resolution mapping [25 , 26] . In maize , GWAS demonstrated that the genetic architecture of the leaf traits , including leaf angle , is dominated by multiple minor QTLs , with little epitasis or environmental interaction [27] . GWAS focused on plant architecture traits , including FLA , were conducted in two separate indica rice collections [28 , 29] . Although several novel FLA-related loci have been detected , this alone is not comprehensive for understanding the natural variation in leaf angle in cultivated rice . The upper canopy of the rice plant , especially the flag leaf , the top leaf after heading , intercepts most of the solar radiation at stages of heading and grain filling . FLA is closely associated with the efficiency of solar utilization by the flag leaf . It is of significance to further dissect the genetic basis of FLA for improvement of plant architecture . In this study , we performed GWAS for FLA with 529 Oryza sativa accessions at the heading stage using a linear mixed model ( LMM ) . Several genome regions were associated with FLA , and 3 previously uncharacterized genes of basic helix-loop-helix ( bHLH ) transcriptional factor subfamily 16 were in or around the associated regions . Overexpression transgenic plant analysis confirmed that members of bHLH subfamily 16 have a conserved function in controlling rice leaf angle . Although both TA and FLA are the major components of plant architecture , a low correlation between TA and FLA and only one QTL with pleiotropic effects on both traits indicated their different genetic bases .
The FLA of 529 O . sativa accessions shared a similar distribution in Hainan and Wuhan , China ( Fig 1 ) . There was a large variation in FLA in the whole population: from 3 . 5° to 152 . 5° in Hainan and from 6° to 163° in Wuhan . However , the FLA of one half of the accessions in the middle ranged from 15° to 30° , and the median value of the FLA was approximately 20° . FLA showed a high heritability of 0 . 79 . FLA in the indica accessions had smaller variations with smaller mean values than those in japonica in both environments ( Fig 1 ) . The correlation coefficients of FLA between Hainan and Wuhan were 0 . 40 and 0 . 60 within the indica and japonica subpopulations that both reached a significant level ( P<0 . 05 ) , respectively . Two-way ANOVA revealed that FLA was dominantly controlled by genetic factors but also influenced by genotype-by-environment interactions ( Table 1 ) . In the indica subpopulation , the interaction between genotype and environment accounted for 28 . 4% of the variation ( Table 1 ) . We also measured TA at the heading stage for the collection grown in the two environments [30] . A significant positive correlation was observed between FLA and TA only in the indica subpopulation , but the correlation coefficients were small in both environments ( Hainan , 0 . 20; Wuhan , 0 . 34 ) . GWAS for FLA were performed using LMM approach in the whole population and in the indica and japonica subpopulations , respectively ( S1 Fig ) . We detected a total of 62 QTLs in Hainan and Wuhan ( Table 2; S1 and S2 Tables ) . They were unevenly distributed on 12 chromosomes . Chromosome 8 harbored the highest number , with 10 QTLs . Eight QTLs were commonly detected in both environments ( Table 2 ) . Of these , 3 QTLs ( qFLA1e , qFLA8f and qFLA12a ) were identified in the full population , 2 QTLs ( qFLA3a and qFLA11a ) were found in the indica subpopulations , and 3 QTLs ( qFLA3e , qFLA6b and qFLA7e ) were detected in the japonica populations . Additionally , qFLA6b and qFLA7e were also detected in the full population separately in Hainan and Wuhan ( S1 and S2 Tables ) . qFLA1g was identified in the full population in Hainan , as well as in the indica subpopulation in Wuhan , while qFLA5b was detected both in the indica subpopulation in Hainan and in the full population in Wuhan ( S1 and S2 Tables ) . We detected 22 QTLs that were unique in Hainan ( S1 Table ) : 16 , 3 and 2 QTLs were only detected in the full population , indica and japonica subpopulations , respectively , and one QTL , qFLA1c , was identified both in the full population and the japonica population . Thirty QTLs were only detected in Wuhan ( S2 Table ) : 18 , 10 and 1 QTLs were found in the full population , indica and japonica subpopulations , respectively , and one QTL , qFLA8e , was detected both in the full population and the japonica population . We compared the genomic positions of known leaf angle genes with the associated sites detected in this study . Four genes were co-localized with our associated sites ( S1 and S2 Tables ) : OsBRI1 , an orthologue of Arabidopsis BRI1 , which plays an important role in BR signaling [14] , was located in the linkage disequilibrium ( LD ) region where qFLA1d was detected in Wuhan . PGL1 and PGL2/OsBUL1 , the homologs of OsILI1 and BU1 in rice [31–33] , were located in the LD regions of qFLA3b detected in Hainan and qFLA2f detected in Wuhan , respectively . OsSPY , encoding an O-linked N-acetylglucosamine transferase [17] , was located in the region of qFLA8j detected in Wuhan . In addition , we compared the localization of the associated sites detected in this study with previously detected leaf angle QTLs with bi-parental mapping populations from the gramene website ( http://www . gramene . org ) and with the significant FLA loci detected via GWAS in previous studies [28 , 29] . A total of 11 associated sites were co-localized with 8 previously reported QTLs ( Table 2; S1 and S2 Tables ) : both qFLA1e and qFLA6b , commonly identified in both environments , were located in the regions of QLa1 and QFla6 , respectively; qFLA6c and qFLA6d , detected only in Wuhan , were located in the region of QFla6; qFLA9c and qFLA9d , only identified in Hainan , were co-localized with fla9; qFLA5b was located in the QFla5 region; and qFLA1f , qFLA2f , qFLA3d and qFLA7d , only detected in Wuhan , were also located in the previous QTL regions . These results indicated the reliability of FLA-related associations in our study . Three bHLH genes , such as OsILI1 ( OsbHLH154 ) , BU1 ( OsbHLH172 ) and PGL2/OsBUL1 ( OsbHLH170 ) , control leaf angle and belong to bHLH subfamily 16 [32–37] . Interestingly , OsbHLH173 and OsbHLH174 , two members of subfamily 16 , are in the local LD regions where the lead SNP of qFLA10c was located ( S1 Table ) . In addition , OsbHLH153 was located close to the QTL qFLA3b ( S1 Table ) . They are likely the genes underlying these QTLs . Haplotype level association analysis frequently improves the power of QTL mapping [26 , 38 , 39] . To provide more evidence for their identities , we constructed the haplotypes of all 6 members of bHLH subfamily 16 and tested the difference in FLA between all possible haplotype pairs . All these genes except BU1 were highly significantly associated with FLA in the whole population in both environments ( S3 Table ) . OsbHLH153 , OsbHLH174 and ILI1 were significantly associated in the japonica subpopulation in Hainan , but none of the 6 genes were associated in the indica subpopulation ( S3 Table ) . Here we presented the results of OsbHLH174 and OsbHLH173 as the candidate genes of qFLA10c and OsbHLH153 as the candidate gene of qFLA3b . Only one SNP ( sf1013651480 ) causing amino acid change ( S8G ) was detected in OsbHLH174 coding region ( Fig 2A ) . Most aus accessions belong to Hap1 containing Ser8 , while most indica and japonica accessions were divided into Hap2-Hap10 shared Gly8 ( Fig 2A ) . The effects of two major haplotypes in the indica subpopulation were similar , while the effects were significantly different among major haplotypes in the japonica accessions , especially between Hap6 and Hap7 ( Fig 2B ) . Ser and Gly are both uncharged hydrophilic amino acids , and both have similar biochemical characters . Therefore , sf1013651480 was not a functional nucleotide polymorphism site ( FNP ) , and the nucleotide change in promoter may lead to the significant difference for FLA . Accordingly , only a SNP ( sf1013702588 ) causing a non-synonymous mutation ( S3G ) was detected in OsbHLH173 , and most aus and indica accessions carried Ser at this site ( S2A Fig ) . Although significant differences in FLA were detected among haplotypes in japonica ( S2B Fig ) , sf1013702588 was not associated with leaf angle . This result suggested the genetic variation of promoter controlling FLA . For OsbHLH153 , one SNP ( sf0303844743 ) caused a non-synonymous mutation ( G12V ) ( S3A Fig ) . Most aus , indica and japonica accessions carried Gly at site 12 ( Haplotypes 1–5 and 7 ) . A small proportion of japonica accessions carried Val at site 12 ( haplotypes 6 and 8 ) . Unlike Gly , Val is an uncharged hydrophobic amino acid . Within japonica rice , FLA of Hap5 with G12 was significantly smaller than that of Hap6 with V12 in Hainan ( S3B Fig ) . In addition , some SNPs in the promoter region were also associated with FLA such as sf0303844574 , sf0303844528 , sf0303844032 , sf0303843859 and sf0303843700 ( S3A Fig ) . Therefore , both polymorphisms in coding and promotor regions of OsbHLH153 caused the variation of leaf angle . To verify the contribution of promotor variation , we firstly examined the expression pattern of these bHLH genes with the rice chip DB “CREP” ( http://crep . ncpgr . cn/crep-cgi/home . pl ) , and found they preferably expressed in young and growing tissues but not in mature tissues ( S4B Fig ) . Then we sampled the flag leaf from 296 accessions for RNA sequencing . Accordingly , it was hard to detect the expressions of all the six members of bHLH subfamily 16 ( S6 Table ) . Therefore , it was failed to perform the associations between expression amounts of these three bHLH genes and FLA . To test whether the expression level of these genes affects leaf angle , then we generated overexpression plants for these 3 bHLH genes . Many OsbHLH174 overexpression plants ( T0 ) showed a significantly increased leaf angle ( Fig 3A ) . We measured FLA and the top second leaf angle ( TSLA ) in two T1 overexpressing lines . All the overexpression plants showed significantly increased lamina joint bending ( Fig 3B and 3C ) . T0 overexpression plants of OsbHLH153 and OsbHLH173 showed an increased leaf angle , and some of these exhibited defective phenotypic variations , including leaf and stem twisting ( S5 Fig ) . A mutation in the promoter of Style2 . 1 , a homolog of these rice bHLHs genes , resulted its decreased expression and differentiated its function in cultivated tomatoes [40] . We overexpressed Style2 . 1 from tomato Solanum pennellii in japonica rice variety Zhonghua 11 . The overexpression plants showed an increased lamina joint ( S6 Fig ) . As many leaf angle-related genes are regulated by plant hormones , especially BRs , we treated the three-leaf seedlings with 4 kinds of hormones and checked the expression change of these 3 bHLH genes , OsbHLH153 , OsbHLH173 and OsbHLH174 , by quantitative real-time reverse transcription-polymerase chain reaction ( qRT-PCR ) . OsbHLH153 and OsbHLH174 were up-regulated immediately and reached a maximum level of over 27× compared with the control group at 4 h after treatment with 100 μM indole-3-acetic acid ( IAA ) , whereas they gradually decreased expression after treatment with 100 μM abscisic acid ( ABA ) ( Fig 4 ) . The expression of OsbHLH153 was up-regulated to a maximum level ( 3 . 7× ) at 8 h after treatment with 10 μM epibrassinolide ( eBL ) and increased 3 . 0× at 1 h after treatment with 100 μM GA4/7 compared with the control group ( Fig 4 ) . The expression of OsbHLH174 reached 3 . 1× and 2 . 2× at 4 h when treated with 10 μM eBL or 100 μM GA4/7 , respectively , compared with the control group ( Fig 4 ) . However , the expression of OsbHLH173 was too low to be detected in seedling and other tissues ( S4A Fig ) . These results suggested that these bHLH genes were regulated by plant hormones and might regulate leaf angle by these hormone pathways . qFLA1d/OsBRI1 made a large contribution to FLA variation in the full population in Wuhan ( S2 Table ) . Haplotype-level association analysis showed that OsBRI1 was strongly associated not only in the whole population in two environments but also in the indica and japonica subpopulations in Hainan ( S3 Table ) . A total of 6 major haplotypes were constructed based on all SNPs in OsBRI1 . Most indica accessions belong to Hap2 or Hap3 , and most japonica rice carried Hap4-Hap6 ( Fig 5A ) . Within indica subpopulation , the flag leaf of accessions carried Hap2 ( D212; Asp , an amino acid with a negative charge of polarity ) was more erect than those carried Hap3 ( G212; Gly ) ( Fig 5A and 5B ) . Within japonica , FLA of the accessions carried OsBRI1-Hap6 ( L623; Leu , a nonpolar amino acid ) was larger than those carried OsBRI1-Hap4 ( S623; Ser , an uncharged hydrophilic amino acid ) and OsBRI1-Hap5 ( S623 ) ( Fig 5A and 5C ) . SNPs sf0129929653 and sf0129928420 might be FNPs separately controlling leaf angle in the indica and japonica subpopulations . Three members of OsbHLH 16 subfamily genes acted downstream of OsBRI1 in BR signaling pathway [33 , 36 , 37] . Based on the hypothesis that OsBRI1 combined with its downstream genes controlling leaf angle with varied effects . Then we investigated effects of gene combinations between three OsbHLHs identified in this study and OsBRI1 on FLA in japonica subpopulation ( Fig 6 ) . Different combinations showed significant differences in FLA . The combinations constructed by haplotypes with small FLA such as OsBRI1-Hap4/OsbHLH153-Hap5 , OsBRI1-Hap5/OsbHLH153-Hap5 , OsBRI1-Hap4/OsbHLH173-Hap4 , OsBRI1-Hap4/OsbHLH174-Hap6 and OsBRI1-Hap4/OsbHLH174-Hap8 had small FLA; and the combinations constructed by haplotypes with large FLA such as OsBRI1-Hap6/OsbHLH153-Hap6 , OsBRI1-Hap6/OsbHLH173-Hap5 and OsBRI1-Hap6/OsbHLH174-Hap7 had large FLA . The other types usually had intermediate FLA such as OsBRI1-Hap5/OsbHLH174-Hap7 . These results indicated that selection for the combinations between these three OsbHLHs and OsBRI1 should be more important and efficient than for single gene when improving leaf angle in japonica subpopulation .
In this study , we found that indica rice has a smaller FLA with a narrower distribution than japonica rice in Hainan and Wuhan ( Fig 1 ) . A total of 17 and 8 unique associations were detected by GWAS in the indica and japonica subpopulations , respectively . However , only 2 and 3 were commonly detected in both environments in indica and japonica subpopulations , respectively ( Table 2; S1 and S2 Tables ) . Further haplotype-level association analysis of leaf angle-related genes located in or around the regions of associations in the two subpopulations indicated that significant differences in FLA between/among haplotypes were detected for all these genes in the japonica accessions in Hainan , while no significant difference was observed for these genes , except for OsBRI1 , in the indica accessions ( S3 Table , Figs 2 and 5; S2 and S3 Figs ) . So , we concluded that FLA has undergone diversifying selection . The indica subpopulation has been fixed with non-functionally differential haplotypes of the leaf angle-related genes , whereas the spread of functionally differentiated haplotypes in japonica has led to a wider variation in FLA . Therefore , there are diverse genetic bases of FLA between the two subpopulations , and some genes regulate FLA , dependent on the environment . The bHLH family , a large family of transcription factors , is found throughout the eukaryotic kingdoms . The basic region functions as a DNA-binding motif , and the HLH region allows the homodimer or heterodimer formation [34 , 35 , 41–43] . Many bHLHs have been functionally characterized with multiple functions in regulating many biological developments . bHLH subfamily 16 is composed of several atypical proteins that modulate the expression of downstream genes by forming heterodimers as non-DNA-binding bHLHs . The PREs in Arabidopsis and style2 . 1 in tomato , which belong to bHLH subfamily 16 , have been reported to regulate cell elongation in different tissues [40 , 44] . In rice , there are 6 genes in bHLH subfamily 16 . Of these , OsILI1 ( OsbHLH154 ) , BU1 ( OsbHLH172 ) and PGL2/OsBUL1 ( OsbHLH170 ) were previously confirmed to regulate leaf angle [33 , 36 , 37] . In this study , the remaining genes of subfamily 16 , OsbHLH153 , OsbHLH173 and OsbHLH174 , were associated with FLA by GWAS ( S1 and S2 Tables ) and by haplotype analysis ( Fig 2 , S2 and S3 Figs ) . There were few SNPs caused non-synonymous in these three bHLH genes . Meanwhile , the mutations in promoters were associated with the variation of FLA ( Fig 2 , S2 and S3 Figs ) . However , we failed to establish the relation between the expression levels of these candidate bHLH genes and FLA because their expressions were too low to be detected in flag leaf ( S4 Fig , S6 Table ) . It is noticed that IBH1 ( ILI1 Binding bHLH Protein 1 ) , formed a heterodimer with ILI1 and its homologs , and its activity inhibited by ILI1 , could be detected in mature organs [33 , 37] . A negative correlation was detected between expression of OsIBH1 and FLA within 64 japonica accessions ( -0 . 255 , P = 0 . 042 ) that was consistent with its negative regulation to FLA . The accessions carried haplotypes of OsbHLH153 and OsbHLH174 with smaller FLA had higher expression levels of OsIBH1 which indicated the expression of OsbHLH153 and OsbHLH174 might be associated with FLA in japonica ( S7 Fig ) . Moreover , overexpressing these three bHLHs increased the leaf angle , like the phenotypic change of overexpressing OsILI1 , BU1 and PGL2/OsBUL1 ( Fig 3 , S5 Fig ) . Overexpressed tomato Style2 . 1 in rice plants also enlarged the FLA ( S6 Fig ) . Although OsILI1 was not associated with FLA at SNP level in this study , haplotype-level association analysis showed that it was associated with FLA in japonica rice ( S3 Table ) . We conclude that the genes in bHLH subfamily 16 had a conserved function in controlling rice leaf angle together with previous studies . In addition , all the members of bHLH subfamily 16 had different expression levels in vivo , but there were similar expression patterns ( S4 Fig ) . Both FLA and TA are important components of plant architecture . The erect growth of cultivated rice showed a smaller TA compared with the prostrate growth of wild rice ( O . rufipogon ) , which was a critical domestication event . A wider phenotypic variation of FLA , from 3 . 3° to 166 . 7° ( Fig 1 ) , was observed than that of the TA , which ranged from 1 . 8° to 34 . 4° [30] in the same collection used in this study . In general , japonica rice has a compact plant architecture , which exhibits a small tiller angle [30] . However , this is completely the opposite for FLA . Specifically , indica rice has a smaller FLA than japonica rice . Therefore , a very low correlation efficient was detected between FLA and TA , indicating distinct genetic bases for FLA and TA . Previous studies also reported a low correlation coefficient [23 , 24 , 29] . Therefore , the possibility of the co-localization of QTLs for leaf angle and TA or QTLs with pleiotropic effects on both traits is limited . In fact , there are very few QTLs with pleiotropic effects on TA and FLA in rice cultivars . It has been reported that only Ta on chromosome 9 and QFla5 had pleiotropic effects on TA and FLA [23] . Here , we compared the genome regions of 62 FLA-related QTLs with 30 QTLs for TA in our previous work [30] and found that only one QTL , qFLA8f , detected in the full population was co-located with TA QTLs qTA8a and qTA8b , likely indicating a new pleiotropic QTL for TA and FLA or two linked genes in this region . More importantly , haplotype analysis showed that the haplotypes of 6 FLA genes were functionally differentiated in japonica accessions , while those in indica accessions were not functionally differentiated , except for OsBRI1 ( Figs 2 and 5 , S2 and S3 Figs , S3 Table ) . However , this is almost the opposite for TA-related genes . Specifically , the haplotypes of the TA-related genes are not functionally differentiated in japonica accessions and are fixed with functional alleles , decreasing the TA [30] . These results suggested that these genes have no pleiotropic effects on TA and FLA . Thus , it is promising that cultivars with compact plant status may be developed by combining favorable indica original alleles for leaf angle and japonica original alleles for TA without considering linkage drag . In summary , FLA is mainly determined by genetic factors in rice , and different genetic factors control the variation of FLA in the indica and japonica subpopulations . The members of bHLH subfamily 16 have the conserved function regulating rice FLA . There is a low correlation coefficient between FLA and TA , and very few QTLs with pleiotropic effects on both traits indicate their diverse genetic bases . The ideal plant architecture in rice may be efficiently developed by combining favorable alleles for both traits by indica-japonica hybridization .
A diverse worldwide collection consisting of 529 O . sativa landraces and elite accessions was sown at the experimental farm of Huazhong Agricultural University in the winter of 2013 in Hainan and in the 2014 rice growing season in Wuhan , China . The 2-year field experiment was designed with 2 replicates per year . The FLA of 5 plants in the middle for each accession 5 days after flowered was used for FLA measurement . The angle between flag leaf and stem was measured by a protractor . The average FLA across 2 replicates within one year was used for GWAS . The basic information of the 529 O . sativa accessions is available in the RiceVarMap ( http://ricevarmap . ncpgr . cn/ ) [45] . The average FLAs used for GWAS are shown in S4 Table . Two-way analyses of variance were separately used to test significant difference between environments and genotypes for the whole population and two subpopulations . The analysis was run in the program Statistica 7 . 0 ( StatSoft . Tulsa , OK , USA ) . Broad-sense heritability ( H2 ) of FLA in the whole population was calculated based on the experiments using the formula: H2=δg2/ ( δg2+δge2/n+δe2/nr ) , where δg2 , δe2 and δge2 were the estimates of genetic , genotype by environment and error variances derived from the mean square expectations of two-way analysis of variance ( ANOVA ) , respectively; n was the number of environments and r was the number of replicates . The whole genomic DNA sequences of the 529 cultivar accessions were genotyped with approximately 2 . 5×coverage genome sequencing using a bar-coded multiplex sequencing approach on an Illumina Genome Analyzer II [46] . The diverse global rice collection was classified into 9 subpopulations: indI , indII , indica intermediate , Tej , Trj , japonica intermediate , Aus , VI and intermediate [46] . Of these 529 varieties , 295 were classified into the indica subpopulation , including indI , indII and indica intermediate , and 156 were classified into the japonica subpopulation , including Tej , Trj and japonica intermediate . To control spurious associations , population structure and kinship were regarded as cofactors when performing GWAS using LMM by the FaST-LMM program [47 , 48] . Kinship was calculated as a realized relationship matrix using FaST-LMM program . Population structure was calculated as Q matrix base on the admixture model[47] . A total of 3 , 916 , 415 , 2 , 767 , 159 and 1 , 857 , 845 SNPs ( minor allele frequency ( MAF ) ≥0 . 05; the number of accessions with minor alleles ≥ 6 ) were employed for GWAS in the full population , indica and japonica subpopulations , respectively . 757 , 578 , 571 , 843 and 245 , 348 effective independent SNPs ( Me ) which were calculated using a method described by Li et al [49] were found in the full population and indica and japonica subpopulations , respectively . The suggestive P values ( 1/Me , 1 . 3×10−6 for the full population , 1 . 8×10−6 for indica and 4 . 1×10−6 for japonica ) were used as the thresholds for associations commonly detected in Hainan and Wuhan or detected only in one environment , but the candidate genes were in their LD regions . Genome-wide significance thresholds ( 0 . 05/Me ) of 6 . 6×10−8 , 8 . 7×10−8 and 2 . 0×10−7 calculated by a modified Bonferroni correction were used for the full population and the indica and japonica subpopulations , respectively , for the associations detected only in Hainan or Wuhan . To obtain independent association signals , multiple SNPs exceeding the threshold in a 5-Mb region were clustered based on an r2 of LD ≥ 0 . 25; the SNPs showing the minimum P value in a cluster were considered to be the lead SNPs [50] . LD was investigated based on standardized disequilibrium coefficients ( D’ ) and squared allele-frequency correlations ( r2 ) for the pairs of SNP loci . The extent of genome-wide LD decay in the different populations was shown in previous studies [50 , 51] . The distances in LD decay in the regions surrounding the lead SNPs identified in this study were calculated , and the method was described in our previous study [30] . The SNPs of the targeted genes in the 529 O . sativa accessions were obtained from the RiceVarMap ( http://ricevarmap . ncpgr . cn/ ) using the gene ID , while for OsbHLH153 , OsbHLH173 and OsbHLH174 , the SNPs in their 2kb promoter regions for the few SNPs in the CDS coordinates were added . The haplotypes of the individual genes carried by at least 10 accessions were used for comparison . An independent t-test and a Duncan’s test were used to compare the differences in the FLA between/among haplotypes using the SSPE program [52] . The method of haplotype-level association analysis of candidate genes was described by our previous studies [38] . Genomic DNA fragments of OsbHLH153 , OsbHLH173 and OsbHLH174 were amplified from Nipponbare DNA , and a genomic DNA fragment of Style2 . 1 was amplified from tomato ( Solanum pennellii ) DNA with gene-specific primers ( S5 Table ) using the high-fidelity LA Taq polymerase ( Takara ) . The PCR products without mutations were cloned into PU1301 with a maize ( Zea mays ) Ubiquitin promoter or into pCAMBIA1301s with the 35S promoter . The constructs were then introduced into Zhonghua 11 ( ZH11 ) by Agrobacterium tumefaciens-mediated transformation using callus induction from mature embryos as subjects [53 , 54] . At least two independent overexpression plants for each construct were used for measuring the FLA . The total DNA were extracted from fresh leaves using the CTAB method [55] . First , the GUS fragment was amplified from the transgenic plants with the primers GUS-F and GUS-R ( S5 Table ) to identify the positive transgenic plants . The positive and negative transgenic plants checked by GUS amplification were then selected to compare their expression levels of the target genes by qRT-PCR . The wild type seeds were sown and germinated on agar medium . After two weeks , the seedlings were transferred to water . On the second day that the plants were grown in water , 4 kinds of hormones were separately added in the water , ensuring the final concentrations of 10 μM eBL , 100 μM IAA , 100 μM GA and 100 μM ABA in the experimental group , with nothing added to the water in the control group . Total RNA was extracted from the whole seedling , except for the root tissue , after treatment for 0 . 5 , 1 , 4 , 8 , 12 and 24 h , respectively . We then analyzed the expression pattern by qRT-PCR . Total RNA was extracted using an RNA extraction kit ( TRIzol reagent , Invitrogen ) . RNA sequencing data of OsbHLH genes were listed in S6 Table . The expression patterns of the FLA genes were then analyzed by qRT-PCR . Measurements were obtained using the relative quantification method . Expression levels were normalized against expression of a ubiquitin ( UBQ ) gene . Error bars indicate standard deviations ( n = 3 ) . All primers used for qRT-PCR are listed in S5 Table . | Rice leaf angle is a major component of ideal plant architecture that determines plant density . Many leaf angle-related genes have been characterized based on mutants , but natural variations of these genes and potential values in genetic improvement have not been evaluated . Here we explore the genetic basis of rice FLA in rice by GWAS . Dozens of quantitative trait loci ( QTLs ) have been identified , but few FLA QTLs have been commonly detected between indica and japonica subpopulations . Alleles for small leaf angle mostly come from indica rice . Three novel OsbHLH genes have been confirmed to regulate leaf angle by overexpression . Together with previous studies , all 6 members of bHLH subfamily 16 have a conserved function in regulating leaf angle . Only one QTL has been identified with pleiotropic effects on FLA and TA in the same collection , indicated the diverse genetic basis between FLA and TA . Alleles for small TA mostly come from japonica rice . We suggest to breed compact plant type cultivars by simultaneously utilizing favorable alleles for FLA and TA via inter-subspecies hybridization . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"biotechnology",
"genome-wide",
"association",
"studies",
"plant",
"anatomy",
"ecology",
"and",
"environmental",
"sciences",
"quantitative",
"trait",
"loci",
"conservation",
"genetics",
"genetic",
"mapping",
"plant",
"science",
"rice",
"genetically",
"modified",
"plants",... | 2018 | Genome-wide association studies reveal that members of bHLH subfamily 16 share a conserved function in regulating flag leaf angle in rice (Oryza sativa) |
Inside individual cells , expression of genes is inherently stochastic and manifests as cell-to-cell variability or noise in protein copy numbers . Since proteins half-lives can be comparable to the cell-cycle length , randomness in cell-division times generates additional intercellular variability in protein levels . Moreover , as many mRNA/protein species are expressed at low-copy numbers , errors incurred in partitioning of molecules between two daughter cells are significant . We derive analytical formulas for the total noise in protein levels when the cell-cycle duration follows a general class of probability distributions . Using a novel hybrid approach the total noise is decomposed into components arising from i ) stochastic expression; ii ) partitioning errors at the time of cell division and iii ) random cell-division events . These formulas reveal that random cell-division times not only generate additional extrinsic noise , but also critically affect the mean protein copy numbers and intrinsic noise components . Counter intuitively , in some parameter regimes , noise in protein levels can decrease as cell-division times become more stochastic . Computations are extended to consider genome duplication , where transcription rate is increased at a random point in the cell cycle . We systematically investigate how the timing of genome duplication influences different protein noise components . Intriguingly , results show that noise contribution from stochastic expression is minimized at an optimal genome-duplication time . Our theoretical results motivate new experimental methods for decomposing protein noise levels from synchronized and asynchronized single-cell expression data . Characterizing the contributions of individual noise mechanisms will lead to precise estimates of gene expression parameters and techniques for altering stochasticity to change phenotype of individual cells .
The level of a protein can deviate considerably from cell-to-cell , in spite of the fact that cells are genetically-identical and are in the same extracellular environment [1–3] . This intercellular variation or noise in protein counts has been implicated in diverse processes such as corrupting functioning of gene networks [4–6] , driving probabilistic cell-fate decisions [7–12] , buffering cell populations from hostile changes in the environment [13–16] , and causing clonal cells to respond differently to the same stimulus [17–19] . An important source of noise driving random fluctuations in protein levels is stochastic gene expression due to the inherent probabilistic nature of biochemical processes [20–23] . Recent experimental studies have uncovered additional noise sources that affect protein copy numbers . For example , the time take to complete cell cycle ( i . e . , time between two successive cell-division events ) has been observed to be stochastic across organisms [24–32] . Moreover , given that many proteins/mRNAs are present inside cells at low-copy numbers , errors incurred in partitioning of molecules between two daughter cells are significant [33–35] . Finally , the time at which a particular gene of interest is duplicated can also vary between cells [36 , 37] . We investigate how such noise sources in the cell-cycle process combine with stochastic gene expression to generate intercellular variability in protein copy numbers ( Fig 1 ) . Prior studies that quantify the effects of cell division on the protein noise level have been restricted to specific cases . For example , noise computations have been done in stochastic gene expression models , where cell divisions occur at deterministic time intervals [33 , 38 , 39] . Recently , we have analyzed a deterministic model of gene expression with random cell-division events [40] . Building up on this work , we formulate a mathematical model that couples stochastic expression of a stable protein with random cell-division events that follow a general class of probability distributions . Moreover , at the time of cell division , proteins are randomly partitioned between two daughter cells based on a framework that allows the partitioning errors to be higher or lower than as predicted by binomial partitioning . For this class of models , we derive an exact analytical formula for the protein noise level as quantified by the steady-state squared Coefficient of Variation ( CV2 ) . This formula is further decomposed into individual components representing contributions from different noise sources . A systematic investigation of this formula leads to novel insights , such as identification of regimes where increasing randomness in the timing of cell-division events decreases the protein noise level . Next , we extend the above model to include genome-duplication events that increase the gene’s transcription rate [36 , 41] . To our knowledge , this is the first study integrating randomness in the genome-duplication process with stochastic gene expression . An exact formula for the protein noise level is derived for this extended model and used to investigate how the timing of duplication affects different noise components . Counter intuitively , results show that doubling of the transcription rate within the cell cycle can lead to smaller fluctuations in protein levels as compared to a constant transcription rate through out the cell cycle . Finally , we discuss how formulas obtained in this study can be used to infer parameters and characterize the gene expression process from single-cell studies .
We consider the standard model of stochastic gene expression [42 , 43] , where mRNAs are transcribed at exponentially distributed time intervals from a constitutive gene with rate kx . For the time being , we exclude genome duplication and the transcription rate is fixed throughout the cell cycle . Assuming short-lived mRNAs , each transcription event results in a burst of proteins [43–45] . The corresponding jump in protein levels is shown as x ( t ) ↦ x ( t ) + B , ( 1 ) where x ( t ) is the protein population count in the mother cell at time t , B is a random burst size drawn from a positively-valued distribution and represents the number of protein molecules synthesized in a single-mRNA lifetime . Motivated by observations in E . coli and mammalian cells , where many proteins have half-lives considerably longer than the cell-doubling time , we assume a stable protein with no active degradation [46–48] . Thus , proteins accumulate within the cell till the time of cell division , at which point they are randomly partitioned between two daughter cells . Let cell division events occur at times ts , s ∈ {1 , 2 , …} . The cell-cycle time T ≔ t s - t s - 1 , ( 2 ) follows an arbitrary positively-valued probability distribution with the following mean and squared coefficient of variation ( CV2 ) 〈 T 〉 = 〈 t s - t s - 1 〉 , C V T 2 = 〈 T 2 〉 - 〈 T 〉 2 〈 T 〉 2 , ( 3 ) where 〈 . 〉 denotes expected value through out this paper . The random change in x ( t ) during cell division is given by x ( t s ) ↦ x + ( t s ) , ( 4 ) where x ( ts ) denotes the protein levels in the mother cell just before division and x+ ( ts ) denotes the protein levels in one of the daughter cells just after division . Conditioned on x ( ts ) , x+ ( ts ) is assumed to have the following statistics 〈 x + ( t s ) | x ( t s ) 〉 = x ( t s ) 2 , x + 2 ( t s ) - 〈 x + ( t s ) 〉 2 | x ( t s ) = α x ( t s ) 4 . ( 5 ) The first equation implies symmetric partitioning , i . e . , on average each of the daughter cells inherits half the number protein molecules just before division . The second equation in Eq ( 5 ) describes the variance of x+ ( ts ) and quantifies the error in partitioning of molecules through the non-negative parameter α . For example , α = 0 represents deterministic partitioning where x+ ( ts ) = x ( ts ) /2 with probability equal to one . A more realistic model for partitioning is each molecule having an equal probability of being in the each daughter cell [49–51] . This results in a binomial distribution for x+ ( ts ) Probability { x + ( t s ) = j | x ( t s ) } = x ( t s ) ! j ! ( x ( t s ) - j ) ! 1 2 x ( t s ) , j ∈ { 0 , 1 , … , x ( t s ) } , ( 6 ) and corresponds to α = 1 in Eq ( 5 ) . Interestingly , recent studies have shown that partitioning of proteins that form clusters or multimers can result in α > 1 in Eq ( 5 ) , i . e . , partitioning errors are much higher than as predicted by the binomial distribution [33 , 39] . In contrast , if molecules push each other to opposite poles of the cell , then the partitioning errors will be smaller than as predicted by Eq ( 6 ) and α < 1 . The model with all the different noise mechanisms ( stochastic expression; random cell-division events and partitioning errors ) is illustrated in Fig 2A and referred to as the full model . We also introduce two additional hybrid models [52 , 53] , where protein production and partitioning are considered in their deterministic limit ( Fig 2B and 2C ) . Note that unlike the full model , where x ( t ) takes non-negative integer values , x ( t ) is continuous in the hybrid models . We will use these hybrid models for decomposing the protein noise level obtained from the full model into individual components representing contributions from different noise sources . In order to quantify the steady-state protein mean and noise , we need to define the stochastic process that governs the timing of cell division . Variations in the duration of cell cycle can result from a variety of factors , such as cell physiology , growth rate , cell size and expression of genes that affect cell-cycle time such as FtsZ [24–32] . Given these complexities , we take a phenomenological approach to modeling cell-cycle time , and assume it to be an independent and identically distributed random variable that is drawn from a mixture of Erlang distributions ( also known as phase-type distribution ) . The motivation for choosing this distribution is two fold: Consider a mixture of n Erlang distributions with mixing probabilities pi , i = {1 , … , n} . Recall that an Erlang distribution of order i is the distribution of the sum of i independent and identical exponential random variables . The cell-cycle time is assumed to have an Erlang distribution of order i with probability pi and can be represented by a continuous-time Markov chain with states Gij , j = {1 , … , i} , i = {1 , … , n} ( Fig 3 ) . Let Bernoulli random variables gij = 1 if the system resides in state Gij and 0 otherwise . The probability of transition Gij → Gi ( j+1 ) in the next infinitesimal time interval [t , t + dt ) is given by ikgij dt , implying that the time spent in each state Gij is exponentially distributed with mean 1/ik . To summarize , at the start of cell cycle , a state Gi1 , i = {1 , … , n} is chosen with probability pi and cell division occurs after transitioning through i exponentially distributed steps . Based on this formulation , the probability of a cell-division event occurring and a new cell cycle obtained from an Erlang distribution of size i starting in the next time interval [t , t + dt ) is given by k p i ∑ j = 1 n ( j g j j ) d t , and whenever the event occurs , the protein level changes as per Eq ( 4 ) . Finally , the mean , the squared coefficient of variation , and the skewness of the cell-cycle time in terms of the Markov chain parameters are given by 〈 T 〉 = 1 k , C V T 2 = ∑ i = 1 n p i i , Skewness = 〈 T 3 〉 - 3 〈 T 〉 ( 〈 T 2 〉 - 〈 T 〉 2 ) - 〈 T 〉 3 ( 〈 T 2 〉 - 〈 T 〉 2 ) 3 / 2 = 2 ∑ i = 1 n p i i 2 ( 7 ) [55] , where 〈T3〉 is the third-order moment of the cell-cycle time . An important property of this class of distributions is that increasing C V T 2 also makes the distribution highly skewed , because from Eq ( 7 ) both the CV and skewness are linear combinations of pi , albeit with different linear coefficients that decrease with i . Considering that ∑ i = 1 n p i = 1 , the only way to increase C V T 2 is by increasing smaller-index components and decreasing larger-index components of the distribution ( i . e . increasing pi and decreasing pj , where i < j ) . Since higher values of i are more penalized in the skewness equation , this would correspond to making the distribution more positively skewed . Hence high values of C V T 2 also means high values of skewness , thus occurrences of longer cell cycles are more probable . As we will shortly see , this property leads to mean protein levels being dependent on C V T 2 .
All the models shown in Fig 2 are identical in terms of finding 〈x ( t ) 〉 and in principle any one of them could have been used . We choose to analyze the full model illustrated in Fig 2A . Time evolution of the statistical moments of x ( t ) can be obtained from the Kolmogorov forward equations corresponding to the full model in Fig 2A combined with the cell-division process described in Fig 3 . We refer the reader to [52 , 56 , 57] for an introduction to moment dynamics for stochastic and hybrid systems . Analysis in Appendix A in S1 Text shows d 〈 x 〉 d t = k x 〈 B 〉 - k 2 ∑ j = 1 n j 〈 x g j j 〉 . ( 8 ) Note that the time-derivative of the mean protein level ( first-order moment ) is unclosed , in the sense that , it depends on the second-order moment 〈xgij〉 . Typically , approximate closure methods are used to solve moments in such cases [52 , 57–62] . However , the fact that gij is binary can be exploited to automatically close moment dynamics . In particular , since gij ∈ {0 , 1} 〈 g i j n x m 〉 = 〈 g i j x m 〉 , n ∈ { 1 , 2 , … } ( 9 ) for any non-negative integer m . Moreover , as only a single state gij can be 1 at any time 〈 g i j g r q x m 〉 = 0 , if i ≠ r or j ≠ q . ( 10 ) Using Eqs ( 9 ) and ( 10 ) , the time evolutions of 〈gij〉 and 〈xgij〉 are obtained as d 〈 g i 1 〉 d t = k p i ∑ j = 1 n ( j 〈 g j j 〉 ) − i k 〈 g i 1 〉 , ( 11 ) d 〈 g i j 〉 d t = i k 〈 g i ( j − 1 ) 〉 − i k 〈 g i j 〉 , j = { 2 , … , i } , ( 12 ) d 〈 x g i 1 〉 d t = k x 〈 B 〉 〈 g i 1 〉 + k 2 p i ∑ j = 1 n ( j 〈 x g j j 〉 ) − i k 〈 x g i 1 〉 , ( 13 ) d 〈 x g i j 〉 d t = k x 〈 B 〉 〈 g i j 〉 − i k 〈 x g i j 〉 + i k 〈 x g i ( j − 1 ) 〉 , j = { 2 , … , i } ( 14 ) and only depend on 〈gij〉 and 〈xgij〉 ( see Appendix A in S1 Text ) . Thus , Eqs ( 8 ) and ( 11 ) – ( 14 ) constitute a closed system of linear differential equations from which moments can be computed exactly . To obtain an analytical formula for the average number of proteins , we start by performing a steady-state analysis of Eq ( 8 ) that yields ∑ j = 1 n j 〈 x g j j 〉 ¯ = 2 k x 〈 B 〉 k , ( 15 ) where 〈 . 〉 ¯ denotes the expected value in the limit t → ∞ . Using Eq ( 15 ) , 〈 x g i 1 〉 ¯ is determined from Eq ( 13 ) , and then all moments 〈 x g i j 〉 ¯ are obtained recursively by performing a steady-state analysis of Eq ( 14 ) for j = {2 , … , i} . This analysis results in 〈 x g i j 〉 ¯ = k x 〈 B 〉 i k p i 1 + j i . ( 16 ) Using Eqs ( 7 ) , ( 16 ) and the fact that ∑ i = 1 n ∑ j = 1 i g i j = 1 we obtain the following expression for the mean protein level 〈 x 〉 ¯ = x ∑ i = 1 n ∑ j = 1 i g i j ¯ = ∑ i = 1 n ∑ j = 1 i 〈 x g i j 〉 ¯ = k x 〈 B 〉 〈 T 〉 3 + C V T 2 2 . ( 17 ) It is important to point that Eq ( 17 ) holds irrespective of the complexity , i . e . , the number of states Gij used in the phase-type distribution to approximate the cell-cycle time distribution . As expected , 〈 x 〉 ¯ increases linearly with the average cell-cycle time duration 〈T〉 with longer cell cycles resulting in more accumulation of proteins . Consistent with previous findings , Eq ( 17 ) shows that the mean protein level is also affected by the randomness in the cell-cycle times ( C V T 2 ) [40 , 63] . For example , 〈 x 〉 ¯ reduces by 25% as T changes from being exponentially distributed ( C V T 2 = 1 ) to periodic ( C V T 2 = 0 ) for fixed 〈T〉 . Next , we determine the noise in protein copy numbers , as quantified by the squared coefficient of variation . Recall that the full model introduced in Fig 2A has three distinct noise mechanisms . Our strategy for computing the protein noise level is to first analyze the model with a single noise source , and then consider models with two and three sources . As shown below , this approach provides a systematic dissection of the protein noise level into components representing contributions from different mechanisms . The full model introduced in Fig 2 assumes that the transcription rate ( i . e . , the protein burst arrival rate ) is constant throughout the cell cycle . This model is now extended to incorporate gene duplication during cell cycle , which increases the burst arrival ( transcription ) rate by f times ( Fig 5 ) . Note that due to gene dosage compensation , doubling the genome does not always correspond to f = 2 [72–74] . If f > 1 , then accumulation of proteins will be bilinear as illustrated in Fig 1 . As before , we again take a phenomenological approach to model the timing of gene duplication . The cell-cycle time T is divided into two intervals: time from the start of cell cycle to gene duplication ( T1 ) , and time from gene duplication to cell division ( T2 ) . T1 and T2 are independent random variables , each drawn from a mixture of Erlang distributions ( see Fig . B in S1 Text ) . The mean cell-cycle duration and its noise can be expressed as 〈 T 〉 = 〈 T 1 〉 + 〈 T 2 〉 , β = 〈 T 1 〉 〈 T 〉 , C V T 2 = β 2 C V T 1 2 + ( 1 - β ) 2 C V T 2 2 , ( 42 ) where C V X 2 denotes the squared coefficient of variation of the random variable X . An important variable in this formulation is β , which represents the average time of gene duplication normalized by the mean cell-cycle time . Thus , β values close to 0 ( 1 ) imply that the gene is duplicated early ( late ) in the cell-cycle process . Moreover , the noise in the gene-duplication time is controlled via C V T 1 2 . We refer the reader to Appendix F in S1 Text for a detailed analysis of the model in Fig 5 and only present the main results on the protein mean and noise levels . The steady-state mean protein count is given by 〈 x 〉 ¯ = k x 〈 B 〉 〈 T 1 〉 2 f ( 1 - β ) + 3 β + β C V T 1 2 2 + k x 〈 B 〉 〈 T 2 〉 3 f ( 1 - β ) + 4 β + f ( 1 - β ) C V T 2 2 2 , ( 43 ) and decreases with β , i . e . , a gene that duplicates early has on average , more number of proteins . When β = 1 , then the transcription rate is kx throughout the cell cycle and we recover the mean protein level obtained in Eq ( 17 ) . Similarly , when β = 0 the transcription rate is fkx and we obtain f times of the amount as in Eq ( 17 ) . As per our earlier observation , more randomness in the timing of genome duplication and cell division ( i . e . , higher C V T 1 2 and C V T 2 2 values ) increases 〈 x 〉 ¯ . Our analysis shows that the total protein noise level can be decomposed into three components C V 2 = C V E 2 + C V R 2 + C V P 2 ( 44 ) where C V E 2 is the extrinsic noise from random genome-duplication/cell-division events , and the sum of the contributions from partitioning errors ( C V R 2 ) and stochastic expression ( C V P 2 ) is the intrinsic noise . We refer the reader to Appendix F in S1 Text for noise formulas for any f , and only present formulas for f = 2 here . In this case , the intrinsic noise is obtained as C V R 2 + C V P 2 = 4 α ( 2 - β ) 3 ( β 2 - 4 β + 6 ) + β 2 C V T 1 2 + 2 ( 1 - β ) 2 C V T 2 2 1 〈 x 〉 ¯ ︷ C V R 2 + ( 10 - 8 β + 3 β 2 ) + 6 ( 1 - β ) 2 C V T 2 2 + 3 β 2 C V T 1 2 3 ( β 2 - 4 β + 6 ) + β 2 C V T 1 2 + 2 ( 1 - β ) 2 C V T 2 2 〈 B 2 〉 〈 B 〉 1 〈 x 〉 ¯ ︷ C V P 2 . ( 45 ) Note that for β = 0 and 1 , we recover the intrinsic noise level in Eq ( 39 ) from Eq ( 45 ) . Interestingly , for B = 1 with probability 1 and α = 1 , the intrinsic noise is always 1/Mean irrespective of the values chosen for C V T 1 2 , C V T 2 2 and β . For high precision in the timing of cell-cycle events ( CVT1 → 0 , CVT2 → 0 ) CV2≈ 4−3 ( β−2 ) 2β2︷CVE23 ( β2−4β+6 ) 2︸Extrinsic noise + 4α ( 2−β ) 3 ( β2−4β+6 ) 1〈x〉¯︷CVR2 + ( 10−8β+3β2 ) 3 ( β2−4β+6 ) 〈B2〉〈B〉1〈x〉¯︷CVP2︸Intrinsic noise , ( 46 ) where mean protein level is given by 〈 x 〉 ¯ ≈ k x 〈 B 〉 〈 T 1 〉 4 - β 2 + k x 〈 B 〉 〈 T 2 〉 3 - β . ( 47 ) We investigate how different noise components in Eq ( 46 ) vary with β as the mean protein level is held fixed by changing kx . Fig 6 shows that C V P 2 follows a U-shaped profile with the optima occurring at β = 2 - 2 ≈ 0 . 6 and the corresponding minimum value being ≈ 5% lower than its value at β = 0 . An implication of this result is that if stochastic expression is the dominant noise source , then gene duplication can result in slightly lower protein noise levels . In contrast to C V P 2 , C V R 2 has a maxima at β = 2 - 2 which is ≈ 6% higher than its value at β = 0 ( Fig 6 ) . Analysis in Appendix F5 in S1 Text reveals that C V R 2 and C V P 2 follow the same qualitative shapes as in Fig 6 for any C V T 1 2 and C V T 2 2 . Interestingly , when C V T 1 2 = C V T 2 2 , the maximum and minimum values of C V R 2 and C V P 2 always occur at β = 2 - 2 albeit with different optimal values than Fig 6 ( see Fig . C in S1 Text ) . For example , if C V T 1 2 = C V T 2 2 = 1 ( i . e . , exponentially distributed T1 and T2 ) , then the maximum value of C V R 2 is 20% higher and the minimum value of C V P 2 is 10% lower than their respective value for β = 0 . Given that the effect of changing β on C V P 2 and C V R 2 is small and antagonistic , the overall affect of genome duplication on intrinsic noise may be minimal and hard to detect experimentally . While the above analysis is for a stable protein , a natural question to ask is how would these results change for an unstable protein ? Consider an unstable protein with half-life considerably shorter than the cell-cycle duration . This rapid turnover ensures that the protein level equilibrates instantaneously after cell-division and gene-duplication events . Let γx denote the protein decay rate . Then , the mean protein level before and after genome duplication is 〈 x 〉 ¯ = k x 〈 B 〉 / γ x and 〈 x 〉 ¯ = 2 k x 〈 B 〉 / γ x , respectively . Note that in the limit of large γx there is no noise contribution from partitioning errors since errors incurred at the time of cell division would be instantaneously corrected . The extrinsic noise , which can be interpreted as the protein noise level for deterministic protein production and decay is obtained as ( for analysis on general f see Appendix G in S1 Text ) C V E 2 = ( 1 - β ) β ( 2 - β ) 2 , ( 48 ) which is similar to noise level reported in [75] . When β = 0 or 1 , the transcription rate and the protein level are constant within the cell cycle and C V E 2 = 0 . Moreover , C V E 2 is maximized at β = 2/3 with a value of 1/8 . Thus , in contrast to a stable protein , extrinsic noise in an unstable protein is strongly dependent on the timing of gene duplication . Next , consider the intrinsic noise component . Analysis in Appendix G in S1 Text shows that the noise contribution from random protein production and decay is C V P 2 = 1 2 〈 B 2 〉 〈 B 〉 + 1 1 〈 x 〉 ¯ , 〈 x 〉 ¯ = k x 〈 B 〉 ( 2 - β ) γ x . ( 49 ) While the mean protein level is strongly dependent on β , the intrinsic noise Fano factor = C V P 2 × 〈 x 〉 is independent of it . Thus , similar to what was observed for a stable protein , the intrinsic noise in an unstable protein is invariant of β for a fixed 〈 x 〉 ¯ .
In this first-of-its-kind study , we have investigated how discrete f-fold changes in the transcription rate due to gene duplication affect the intercellular variability in protein levels . Not surprisingly , the timing of genome duplication strongly affects the mean protein level—〈 x 〉 ¯ can change up to f folds depending on whether the gene duplicates early ( β = 0 ) or late ( β = 1 ) in the cell cycle . Results show that genome duplication has counter intuitive effects on the protein noise level ( Fig 6 ) . For example , if stochastic expression is the dominant source of noise , then doubling of transcription due to duplication results in lower noise , as compared to constant transcription throughout the cell cycle . This is because for the same mean protein level , there are more burst ( transcription ) events in the case of genome duplication ( f = 2 ) than constant transcription ( f = 1 ) . For example , consider deterministic timing ( C V T 1 2 = C V T 2 2 = 0 ) and gene duplication in the middle of the cell cycle ( β = 0 . 5 ) . Then , for the case β = 1 , there are on average kx〈T〉 burst events per cell cycle . For the same 〈 x 〉 ¯ , there are 1 . 05kx〈T〉 production events in the case of gene duplication ( β = 0 . 5 ) . This slight increase in the number of transcription events leads to better averaging of bursty protein synthesis and lower noise levels . Overall , the effect of β on different noise component is quite modest: as β varies , noise components deviate at maximum ≈ 20% from their values at β = 0 ( Fig 6 ) . These results are in contrast to the case of an unstable protein , where noise from the cell-cycle process is strongly dependent on β as shown in Eq ( 48 ) . The mathematical framework introduced for modeling timing of cell division can be easily used to compute noise in synchronized cells . For example , let the cell-cycle duration be an Erlang distribution with shape parameter n and rate parameter nk ( i . e . , pn = 1 in Fig 3 ) , which can be biologically interpreted as cells moving through n cell-cycle stages Gn1 , Gn2 , … , Gnn . Statistical moments conditioned on the cell-cycle stage Gnj can be obtained using 〈 x m ∣ g n j = 1 〉 ¯ = 〈 g n j x m 〉 ¯ 〈 g n j 〉 ¯ , m ∈ { 1 , 2 } . ( 50 ) Using Eq ( 50 ) and moments 〈 g n j x m 〉 ¯ obtained from Eqs ( 16 ) and ( 35 ) – ( 37 ) , yields the following conditional mean 〈 x | g n j = 1 〉 ¯ = k x 〈 B 〉 〈 T 〉 1 + j n , ( 51 ) which increases with cell-cycle stage ( i . e . , higher values of j ) . The protein noise level given that cells are in stage Gnj is given by CV2|gnj=1≔〈x2|gnj=1〉¯−〈x|gnj=1〉¯2〈x|gnj=1〉¯2=n+3j3 ( n+j ) 2︷CVE2︸Extrinsic noise+2nα3 ( n+j ) 1〈x|gnj〉¯︷CVR2+n+3j3 ( n+j ) 〈B2〉〈B〉1〈x|gnj〉¯︷CVP2︸Intrinsic noise . ( 52 ) Note that if n is large then the first term , which represents the noise contribution from the cell-cycle process , is negligible and can be dropped . Interesting , while the noise contribution from partitioning errors C V R 2 decreases with cell-cycle stage , the noise contribution from stochastic expression C V P 2 increases with j . Moreover , for B = 1 with probability 1 and α = 1 , the intrinsic noise is always 1/Mean irrespective of j . Assuming high n , the noise at cell birth ( j = 1 ) and division ( j = n ) are obtained as CV2|gn1=1=2α31〈x|gn1〉¯︷CVR2+13〈B2〉〈B〉1〈x|gn1〉¯︷CVP2︸Intrinsic noise ( 53 ) CV2|gnn=1=α31〈x|gnn〉¯︷CVR2+23〈B2〉〈B〉1〈x|gnn〉¯︷CVP2︸Intrinsic noise , ( 54 ) respectively . Thus , measurements of Eqs ( 53 ) and ( 54 ) by synchronizing cells ( or by using cell size as a proxy for cell-cycle stage ) can be used to quantify α and 〈B2〉/〈B〉 , providing a novel way to separate these noise contributions . Next , we discuss how noise in asynchronous cell can be used to quantify these parameters . Simple models of bursty expression and decay predict the distribution of protein levels to be negative binomial ( or gamma distributed in the continuous framework ) [80 , 81] . These distributions are characterized by two parameter—the burst arrival rate kx and the average burst size 〈B〉 , which can be estimated from measured protein mean and noise levels . This method has been used for estimating kx and 〈B〉 across different genes in E . coli [47 , 82] . Our detailed model that takes into account partitioning errors predicts ( ignoring gene-duplication effects ) Intrinsic noise = 4 α 3 ( 3 + C V T 2 ) 1 〈 x 〉 ¯ + 3 C V T 2 + 5 3 ( 3 + C V T 2 ) 〈 B 2 〉 〈 B 〉 1 〈 x 〉 ¯ . ( 55 ) Using C V T 2 ⪡ 1 and a geometrically distributed B [50 , 83–85] , Eq ( 55 ) reduces to Intrinsic noise = 4 α 9 1 〈 x 〉 ¯ + 5 9 1 + 2 〈 B 〉 〈 x 〉 ¯ . ( 56 ) Given measurements of intrinsic noise and the mean protein level , 〈B〉 can be estimated from Eq ( 56 ) assuming α = 1 ( i . e . , binomial partitioning ) . Once 〈B〉 is known , kx is obtained from the mean protein level given by Eq ( 17 ) . Since for many genes 〈B〉 ≈ 0 . 5–5 [47] , the contribution of the first term in Eq ( 56 ) is significant , and ignoring it could lead to overestimation of 〈B〉 . Overestimation would be even more severe if α happen to be much higher than 1 , as would be the case for proteins that form aggregates or multimers [33] . One approach to estimate both 〈B〉 and α is to measure intrinsic noise changes in response to perturbing 〈B〉 by , for example , changing the mRNA translation rate through mutations in the ribosomal-binding sites ( RBS ) . Consider a hypothetical scenario where the Fano Factor ( intrinsic noise times the mean level ) is 6 . Let mutations in the RBS reduces 〈 x 〉 ¯ by 50% , implying a 50% reduction in 〈B〉 . If the Fano factor changes from 6 to 4 due to this mutation , then 〈B〉 = 3 . 6 and 〈α〉 = 3 . 25 . Our recent work has shown that higher-order statistics of protein levels ( i . e . , skewness and kurtosis ) or transient changes in protein noise levels in response to blocking transcription provide additional information for discriminating between noise mechanisms [86 , 87] . Up till now these studies have ignored noise sources in the cell-cycle process . It remains to be seen if such methods can be used for separating the noise contributions of partitioning errors and stochastic expression to reliably estimate 〈B〉 and α . An important limitation of our modeling approach is that it does not take into account the size of growing cells . Recent experimental studies have provided important insights into the regulatory mechanisms controlling cell size [88–91] . More specifically , studies in E . coli and yeast argue for an “adder” model , where cell-cycle timing is controlled so as to add a constant volume between cell birth and division [78 , 91–93] . Assuming exponential growth , this implies that the time taken to complete cell cycle is negatively correlated with cell size at birth . More importantly , cell size also affects gene expression—in mammalian cells transcription rates linearly increase with the cell size [94] . Thus , as cells become bigger they also produce more mRNAs to ensure gene product concentrations remains more or less constant . A main direction of future work would be to explicitly include cell size with size-dependent expression and timing of cell division determined by the adder model . This formulation will for the first time , allow simultaneous investigation of stochasticity in cell size , protein molecular count and concentration . Our study ignores genetic promoter switching between active and inactive states , which has been shown to be a major source of noise in the expression of genes across organisms [95–104] . Taking into account promote switching is particularly important for genome-duplication studies , where doubling the number of gene copies could lead to more efficient averaging of promoter fluctuations . Another direction of future work will be to incorporate this additional noise source into the modeling framework and investigate its contribution as a function of gene-duplication timing . | Inside individual cells , gene products often occur at low molecular counts and are subject to considerable stochastic fluctuations ( noise ) in copy numbers over time . An important consequence of noisy expression is that the level of a protein can vary considerably even among genetically identical cells exposed to the same environment . Such non-genetic phenotypic heterogeneity is physiologically relevant and critically influences diverse cellular processes . In addition to noise sources inherent in gene product synthesis , recent experimental studies have uncovered additional noise mechanisms that critically effect expression . For example , the time within the cell cycle when a gene duplicates , and the time taken to complete cell cycle are governed by random processes . The key contribution of this work is development of novel mathematical results quantifying how cell cycle-related noise sources combine with stochastic expression to drive intercellular variability in protein molecular counts . Derived formulas lead to many counterintuitive results , such as increasing randomness in the timing of cell division can lower noise in the level of a protein . Finally , these results inform experimental strategies to systematically dissect the contributions of different noise sources in the expression of a gene of interest . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"engineering",
"and",
"technology",
"cell",
"cycle",
"and",
"cell",
"division",
"signal",
"processing",
"cell",
"processes",
"binomials",
"random",
"variables",
"noise",
"reduction",
"dna",
"transcription",
"probability",
"distribution",
"mathematics",
"algebra",
"polyn... | 2016 | Intercellular Variability in Protein Levels from Stochastic Expression and Noisy Cell Cycle Processes |
Kinetochores are multi-protein complexes that mediate the physical coupling of sister chromatids to spindle microtubule bundles ( called kinetochore ( K ) -fibres ) from respective poles . These kinetochore-attached K-fibres generate pushing and pulling forces , which combine with polar ejection forces ( PEF ) and elastic inter-sister chromatin to govern chromosome movements . Classic experiments in meiotic cells using calibrated micro-needles measured an approximate stall force for a chromosome , but methods that allow the systematic determination of forces acting on a kinetochore in living cells are lacking . Here we report the development of mathematical models that can be fitted ( reverse engineered ) to high-resolution kinetochore tracking data , thereby estimating the model parameters and allowing us to indirectly compute the ( relative ) force components ( K-fibre , spring force and PEF ) acting on individual sister kinetochores in vivo . We applied our methodology to thousands of human kinetochore pair trajectories and report distinct signatures in temporal force profiles during directional switches . We found the K-fibre force to be the dominant force throughout oscillations , and the centromeric spring the smallest although it has the strongest directional switching signature . There is also structure throughout the metaphase plate , with a steeper PEF potential well towards the periphery and a concomitant reduction in plate thickness and oscillation amplitude . This data driven reverse engineering approach is sufficiently flexible to allow fitting of more complex mechanistic models; mathematical models of kinetochore dynamics can therefore be thoroughly tested on experimental data for the first time . Future work will now be able to map out how individual proteins contribute to kinetochore-based force generation and sensing .
Chromosomes are attached to , and their movements powered by , kinetochores , multi-protein machines that assemble on each sister chromatid and form dynamic attachments to bundles of kinetochore-microtubules ( K-MTs ) called K-fibres [1] ( see Fig 1A ) . A long-standing challenge in the mitosis field is to measure the magnitude of forces that kinetochores can generate and identify the molecular components and mechanisms responsible . Nicklas and colleagues addressed this question by using calibrated micro-needles to pull on chromosomes in grasshopper spermatocytes , measuring the force needed to stall chromosome movement [2] . These classic experiments found that > 20 pN was necessary to slow , and 700 pN to stall , chromosome-to-pole movement in anaphase , while there was a much lower stall force ( 50 pN ) for chromosome movement during congression . These measured values are considerably higher than the 0 . 1 pN that was calculated ( based on Stokes law; force = viscosity × chromosome size × velocity ) to be required for moving a chromosome under normal conditions [3 , 4] . Neither of these approaches , however , are able to separate out the different forces that are acting on a kinetochore: these are known to include ( i ) K-MT polymerisation and depolymerisation dynamics , ( ii ) elastic forces from the centromeric chromatin that operates as a compliant linkage between sister kinetochores [5] , ( iii ) polar ejection forces ( PEF ) mediated by the interactions between non-kinetochore microtubules ( MTs ) and chromosome arms , ( iv ) poleward MT flux that involves the continuous displacement of K-fibres towards the pole driven by minus-end depolymerisation and MT sliding [6] . Metaphase provides a unique phase of mitosis for scrutinising these mechanisms since sister kinetochores undergo quasi-periodic oscillatory motion along the spindle axis for several minutes [7 , 8] . These movements are possible because each sister kinetochore can adopt either a poleward ( P ) moving state ( the leading sister ) by attaching to a depolymerising K-fibre or an away-from-the-pole ( AP ) state ( the trailing sister ) by attaching to a polymerising K-fibre . Switching between the AP and P states causes kinetochores to change direction a behaviour termed directional instability [7] . Major advances in understanding these chromosome oscillations have come from tracking fluorescently-labelled kinetochores in living cells [8–13] . In particular , tracking the full complement of kinetochores in three dimensions has generated systematic and comprehensive datasets describing kinetochore motion in human mitotic cells [8 , 11] and during mouse meiosis [10] . Such complete kinetochore tracking methodologies have allowed identification of many factors and mechanisms that control different aspects of sister kinetochore dynamics [8 , 11 , 14–19] . Crucially , it has emerged that in human cells kinetochore oscillations are very stochastic with a wide-range of periods and inter-trajectory variations [8] . Therefore single trajectory analysis is necessary , so that both the unique dynamics inherent to each single trajectory and the inter-kinetochore variation across whole populations can be determined . To achieve this , fitting of kinetochore mathematical models is required , firstly for accurate description of kinetochore dynamics by automated partitioning into P and AP movements , and secondly , in order to unravel the influence of multiple force generators on the dynamics . In particular , through automated analysis large numbers of trajectories can be analysed to determine the trajectory stochasticity and variability; only through model fitting can we obtain quantitative and statistical assessment of the range of dynamics present in trajectories . Reverse engineering differs from other modelling approaches in cell biology ( for a review , see [20] ) in that it involves inferring a model ( and its parameters ) from observed data , a field that is extremely active in many disciplines ( e . g . , weather prediction , computer vision , epidemiology , phylogenetics , econometrics and financial markets ) and has rapidly grown in importance in systems biology over the last decade . It is a leading method for the analysis of microarray data ( e . g . , BGX software [21] ) , inference of gene regulatory networks from time series or multiple conditions ( e . g . , GRENITS software [22] ) , and biochemical network dynamics [23 , 24] . However , there is a dearth of applications to spatial dynamics in cell biology , with only studies of immunological synapse patternation dynamics [25] , origin firing in cell division [26] , molecular motor force generation in Drosophila mitotic spindles [27] , and single molecule conformational change [28] employing the methodology to the best of our knowledge . A biophysical model is required for the reverse engineering of stochastic spatial systems , adding to the complexity of an already computationally challenging problem . However , the reverse engineering methodology , particularly in a Bayesian framework , has mature theoretical underpinnings [29] . It is exceptionally powerful and highly flexible with regard to model structure , with an extensive suite of methods for model fitting , including methods for inferring parameters of stochastic differential equations [30] . There are also well-established algorithms for model selection and hypothesis testing [31] , which allow biological hypotheses to be addressed in a principled manner , and techniques to incorporate experimental design are also available [32] . These methods are very much under utilised in spatial biological applications . A key problem however is the choice of the model to fit . There are a number of models of kinetochore oscillations in the literature ( see [33 , 34] for reviews ) , with the most biophysically-detailed model [35] demonstrating how oscillations can emerge from dynamic instability of individual K-MTs . However , this model is too complex to fit to individual trajectories since kinetochore positional data is insufficient to identify parameters pertaining to molecular kinetics or energetics . Moreover , the model does not account for additional factors which are very likely crucial in vivo , e . g . , the interactions between K-MTs in K-fibre bundles , the response of bundles to forces , and the energetic/dynamic effects of K-MT-associated proteins . Therefore , in this paper , we propose a new model that captures the high-level dynamics of the system whilst making minimal assumptions , yet can be directly reverse engineered to individual trajectories . The model is sufficiently simple to allow the determination of all parameters on kinetochore trajectory data alone . This paper is organised as follows: In section ‘A principle force model of kinetochore oscillations’ we present our new model for kinetochore oscillations and develop the reverse engineering methodology , specifically a Markov chain Monte Carlo ( MCMC ) algorithm that infers the model parameters ( and hidden states ) from a trajectory time-series . Model derivation and the model likelihood underpinning the MCMC algorithm are described in the Materials and Methods . In section ‘Reverse engineering individual trajectories’ we demonstrate on a single trajectory how the reverse engineering methodology works , demonstrating high confidence estimation of the model parameters and the switching points of the two sisters . This is extended across 1000s of trajectories in section ‘Identifying oscillatory trajectories using model comparison statistics’ , where we develop techniques that can discriminate the quality of oscillation . In section ‘Trajectory heterogeneity across the metaphase plate’ we examine the variability of trajectories across the population and within cells . In section ‘Relative force components during oscillations’ we show that the K-fibres dominate the mechanics , while there is variation in the PEF across the metaphase plate that explains plate structure and oscillation characteristics throughout the plate . Finally in ‘Discussion’ we discuss the implications of our results to the kinetochore field .
In this section we present a new model of kinetochore oscillations , the coherence-incoherence model , and an inference algorithm that infers the model parameters from a single paired-sister trajectory time-series . In order to objectively identify oscillatory trajectories , we used three methods . Firstly we used Bayes factors ( ratio of model marginal likelihoods ) , comparing the coherence-incoherence model Mcoh to a Brownian motion ( BM ) model MBM that considers both kinetochores to move as independent ( 1D ) BMs ( similarly discretised to frames , see section 2 . 1 in S1 Text , and section 2 . 2 for more complex BM models ) . We used Chen’s method [39] to estimate the marginal likelihood π ( X∣M ) from MCMC samples ( see section 2 . 3 in S1 Text ) . The data supports model M1 over M2 when the Bayes factor B[M1/M2] = π ( X|M1 ) /π ( X|M2 ) is greater than 1 . Secondly , we used an intrinsic model statistic , explained variance ( EV ) , which measures how much of the variance in the kinetochore pair trajectory is explained by the model ( see section 2 . 4 in S1 Text ) . A pure BM ( MBM ) would have a low EV , while a coupled kinetochore pair undergoing saw-tooth oscillation would have an EV of 1 . These two methods are based on explaining kinetochore displacements between frames . However , displacements can be arbitrarily reordered; hence these statistics do not account for correlations between consecutive frames as is seen in processive movement . Therefore , our third method is a directional switching statistic designed to account for correlations between displacements by testing if the number of times the direction of motion is changed is consistent with a random walk , where the number of directional switches is binomially distributed ( see section 2 . 5 in S1 Text ) . Live-cell imaging and tracking of kinetochores using HeLa-Kyoto ( K ) cells stably expressing eGFP-CENP-A / eGFP-Centrin1 is described in S1 Text and based on previous work [40 , 41] . The core of the tracking software ( MATLAB kinetochore tracking software ( KiT ) ) is available from http://mechanochemistry . org/mcainsh/software . php and will be described in a forthcoming paper .
To investigate sister kinetochore dynamics during metaphase we developed a dynamic mathematical model which incorporates the three principle forces acting on kinetochores: a constant driving force from either a polymerising or depolymerising K-fibre , the spring tension in the connecting chromatin spring ( modelled as a Hookean spring ) , and the PEF , linearised around the metaphase plate ( Fig 1A , see Materials and Methods ) , with frame to frame displacement dynamics in Eq ( 5 ) . Simulations of this coherence-incoherence model are qualitatively consistent with the data ( see simulation and data in Fig 1B and 1C ) . We defined sisters as moving coherently if both sisters are moving in the same direction , i . e . one sister is attached to a polymerising K-fibre ( + state ) and the other to a depolymerising K-fibre ( – state ) ; otherwise the sisters are in an incoherent state that can be either a +/+ ( both polymerising hereafter called contracting ) or –/– ( both depolymerising hereafter called expanding ) . Note that , by coherence we refer to coherence between K-fibres , rather than to ( in ) coherence observed within K-fibres [19 , 42] . In this model sisters switch from + to − ( and vice versa ) independently through a probabilistic waiting time ( exponentially distributed , i . e . , there is no location or history dependence ) , with a waiting time that is dependent on the sister coherence state ( Fig 1D ) . This model produces qualitatively realistic oscillations when the coherence state is longer lived than the incoherence state , i . e . , when coherence is restored quickly , either as a sustained switching ( both sisters switch direction ) or as a switching reversal ( first sister switching back to its original direction of motion ) . Crucially , the sister which switches is chosen at random , i . e . , there is no intrinsic bias in the model thereby allowing the bias inherent to the kinetochores to be estimated from the data . We developed a reverse engineering algorithm for this model within a Bayesian framework , i . e . , we used a Markov chain Monte Carlo ( MCMC ) algorithm to infer the posterior probability π ( Θ , σ t k | X t k ) for the parameters Θ ( spring constant κ; spring natural length L; polymerisation v+ , and depolymerisation v− speeds; PEF strength α and noise σ2 ) and the unobserved sister states σ t k ( − / − , − / + , + / − or + / - ) through time from the trajectory data X t k ( sister k = 1 , 2 ) . The MCMC algorithm is described in section 1 of S1 Text . The model has an a posteriori identifiability problem , i . e . , a single trajectory does not have sufficient information to determine all of the parameters; specifically only two of v+ , v− and L can be inferred from a trajectory because of an ( approximate ) symmetry in the model ( see section 1 . 3 of S1 Text and S1C Fig ) . However , this is easily resolved through determination of the spring natural length by treating cells with nocodazole which depolymerises all the spindle MTs , thus allowing the chromatin linkage between sisters to relax ( S1A Fig; see also section 3 . 1 in S1 Text ) . In absence of external forces ( K-fibre forces ) the inter-kinetochore distance can be modelled as moving in a harmonic well; thus we were able to infer the rest length of the linkage L = 775 ± 110 nm ( ± population s . d . ) . This resolved the identifiability problem and enabled the estimation of all parameters in Eq ( 8 ) for each paired sister trajectory . Unfortunately we cannot infer the forces directly; only relative forces can be inferred up to the drag coefficient , giving an effective velocity for each force component thereby allowing comparison between these components . This also means that the spring constant κ and PEF constant α are reported in s−1 . If the kinetochore/chromatid drag coefficient γ is independently determined we can use our method to infer the magnitude of the component forces , nevertheless great insight can be obtained by a comparative analysis alone . We applied our reverse engineering algorithm for the coherence-incoherence model of kinetochore sister dynamics ( Eq 5 ) to paired sister trajectories derived from 3D live cell imaging of HeLa-K eGFP-CENP-A eGFP-Centrin1 cells . Previous analysis of kinetochore dynamics in human cells using a frame rate of 7 . 5 s demonstrated a range of stochasticity in individual trajectories with irregular saw-tooth oscillations of approximate period 75 s ( 10 frames ) [8 , 17] but the temporal resolution was not sufficient to pinpoint directional switching events . In 2D imaging at a higher frame rate of 2 s kinetochore switches are clearly evident [19] but the window averaging algorithm used to assign P/AP direction in this study was not able to locate switching events . Furthermore , tracking kinetochores in 2D produces short trajectories since kinetochores can move out of the focal plane . To produce suitable data for inferring switching points we used a 3D spinning-disk confocal imaging system capable of a 2 s frame rate with 25 Z-planes ( voxel size 138 × 138 × 500 nm3 ) over 5 min and tracked the sub-pixel position of sister kinetochores ( Fig 1E; see Burroughs et al . [40] ) . We observed pseudo-periodic oscillations as previously reported ( Fig 1C ) , with a half-period of 26 s , as determined by autocorrelation analysis . The oscillation is approximately saw-tooth ( constant velocity ) , with clearly defined switching events . Parameter estimation and hidden state determination for the trajectory in Fig 1C is illustrated in Fig 2 , a trajectory with good oscillatory behaviour . Parameter posteriors appear Gaussian and all show low variance , i . e . , low parameter uncertainty ( Fig 2A–2F ) . The natural length L posterior is close to the prior; the shift is probably due to the approximate nature of the symmetry , i . e . , the twist carries some information . Switching events were confidently identified for both sisters ( Fig 2G and 2H ) . The inferred hidden state shows strong regions of coherence ( sisters moving in the same direction ) interspersed with short periods of incoherence ( sisters moving in opposing directions ) that correspond to contraction ( +/+ ) in this trajectory ( Fig 2I ) . There is high confidence in assignment of the sister polymerisation state ( Fig 2I ) indicative of a highly deterministic behaviour ( strong clear oscillations ) . This particular trajectory shows the previously reported [7] switching choreography wherein the lead sister switches first at every directional switching event . This is also evident directly from the trajectory time-series Fig 1C . This lead sister driven dynamics is responsible for the contracting incoherent state observed between coherent runs and gives the ‘standard’ choreography with the inter-sister distance relaxing at a switching event and increasing over the following half-period as the lead sister moves away [12] . It is clear from examining examples of trajectories ( S2 Fig ) that there is a high degree of variability between kinetochore trajectories . Thus , analysing a few examples is insufficient to assess kinetochore dynamics or the efficacy of a model in capturing those dynamics . We therefore applied our reverse engineering strategy to a large database of trajectories , specifically 1169 sister pairs across 81 cells; on an additional 84 trajectories the MCMC algorithm failed to converge suggesting that the signal-to-noise ratio was low . Visual observation of these trajectories confirmed they were very stochastic; convergence failure was thus a direct consequence of a lack of oscillations . Since our model is aimed at explaining oscillations we could safely ignore non-converging trajectories . To quantify levels of oscillatory behaviour we used a combination of statistics ( see Materials and Methods and section 2 of S1 Text ) : ( i ) the explained variance ( EV ) of the model , which describes how much of the trajectory variance can be explained by the model , ( ii ) Bayes factors derived from a Bayesian model selection analysis between the coherence-incoherence model and a BM model , and ( iii ) a directional correlation statistic based on the expected number of directional switches in a BM . EV varied from around 0 to 66% ( Fig 3A ) with mean 26 ± 14% ( ± distribution s . d . ) . The variability in trajectories is apparent in the range of EV shown on a per cell basis ( Fig 3B ) , indicating that all cells have a similar profile of near deterministic and highly stochastic trajectories . EV allowed us to rank trajectories by how well the model fitted—the trajectory shown in Fig 1C and reverse engineered in Fig 2 had the highest EV —but does not provide support for the model compared to any other since it is an intrinsic measure of fit . We therefore complemented the EV statistic with a comparative test using the Bayes factor B[Mcoh/MBM] between the coherence-incoherence model Mcoh and a Brownian motion ( BM ) model MBM ( we also compared against BM models with an inter-sister spring and drift with almost identical results; see section 2 . 2 of S1 Text and S3A and S3B Fig ) . Surprisingly , B[Mcoh/MBM] showed significant preference for MBM; we found log B [Mcoh/MBM] < 0 for over 95% of the trajectories ( Fig 3C and 3D; a log B[Mcoh/MBM] < 0 indicates preference for MBM ) . This is due to the lack of temporal structure in the MBM —displacements are treated as independent , identically Gaussian distributed so B[Mcoh/MBM] is predominantly a measure of whether the kinetochore displacements are Gaussian or not . Trajectories where Mcoh is preferred have very regular saw-tooth oscillations and thus an over-representation of large displacements; B[Mcoh/MBM] is thus a good discriminator of strong oscillating trajectories . Comparing EV with B[Mcoh/MBM] demonstrated a correlation as expected ( overall ρ = 0 . 14 , p = 0 . 0008; Fig 3G ) , but also revealed a group of outlier trajectories ( green shading in figure ) that had higher than average B[Mcoh/MBM] but low EV . These trajectories tended to have a few excessively large displacements indicative of tracking errors . We thus filtered these from the analysis by restricting to trajectories approximating the linear relationship between B[Mcoh/MBM] and EV ( the selection region shown as a grey bar in Fig 3G; ρ = 0 . 90 for grey region ) comprising 843 out of 1169 converged trajectories , 72% . The consistency between EV and B[Mcoh/MBM] as measures of oscillatory quality is reassuring; however as discussed above , although able to rank trajectories neither gives a classification into oscillatory and non-oscillatory trajectories , only classifying into Gaussian and non-Gaussian displacements . We therefore used a direction correlation statistic DnΔt ( see section 2 . 5 of S1 Text ) to assess whether higher EV and B[Mcoh/MBM] was indicative of correlated displacements as would be expected for an oscillating trajectory which makes repeated sustained runs ( Fig 3E and 3F ) . We found a strong correlation between D1Δt and both B[Mcoh/MBM] and EV ( r = 0 . 49 and 0 . 57 , respectively; S3C and S3D Fig ) . With subsampling this was drastically reduced , most likely because subsampling reduces noise and thus noisy trajectories , that are penalised under B[Mcoh/MBM] or EV ( in absence of subsampling ) , score higher under DnΔt for n > 1 . We show trajectories scoring D1Δt ≥ 0 . 99 in blue in Fig 3G ( i . e . , failing correlation test at 1% significance ) ; the majority of these are in the top-right quadrant indicating that trajectories towards the top-right of the selection region are indeed more oscillatory . All these measures show that all cells have varying levels of stochasticity amongst their trajectories ( Fig 3B , 3D and 3F ) . The above measures together provide strong evidence that we are able to isolate the oscillatory trajectories ( blue in Fig 3G ) and these conform to the coherence-incoherence model dynamics , i . e . , this model is able to describe the dynamics well . In the remaining sections we limit analysis to those trajectories passing the filtering criterion described above ( indicated by grey region in Fig 3G ) , incorporating both oscillatory and highly stochastic trajectories as determined by the correlation statistic DΔt . As shown in Fig 2 , the posterior distributions for the model parameters can be computed for single trajectories , typically giving high confidence parameter estimates . To summarise across the dataset , we used the posterior means of each trajectory ( Fig 4 ) ; these show that there is significant heterogeneity between trajectories in their parameter values , with ranges over an order of magnitude in many cases . The natural length is the exception because it is constrained by the informed nocodazole prior and thus the means are tightly clustered ( mean 795 ± 31 nm; Fig 4E ) about the prior mean ( 775 ± 110 nm ) . These parameter estimates can be used to unravel the forces driving kinetochore movements as follows . It is widely believed that the depolymerising K-fibre attached to the leading sister provides the dominant driving force in kinetochore motion [33 , 43] . Our data supports this view with the estimate for v− being significantly larger in magnitude than v+ ( v− = −35 ± 15 vs . v+ = 13 ± 16 nm s−1 , ± s . d . , with ∣v−∣ > v+ in 97% of trajectories , p < 10−202; binomial test ) ; recall speed is commensurate with force in our model and v± are the inferred speed components . This indicates that depolymerising forces are significantly larger than polymerising ( Fig 4A ) . Typical kinetochore speeds are around 23 ± 6 nm s−1 , similar to the average of v+ and ∣v−∣ [8] , whilst ∣v−∣ > v+ implies that the sisters will typically separate over the course of a coherent run ( standard choreography ) , increasing the inter-sister distance . This is consistent with the measured average inter-sister distance of around 1 µm , which means that the inter-kinetochore spring is typically under tension with an average extension of 205 ± 31 nm . For the spring and PEF , we found that κ ≈ 2α ( Fig 4B and 4C ) implying that the PEF is equal in magnitude to the inter-kinetochore spring force when kinetochores are twice as far from the metaphase plate as the spring is extended; thus the PEF typically dominates the spring force at displacements away from the metaphase plate of over 0 . 5 μm . This also indicates that the PEF is effective at maintaining a thin metaphase plate since for displacements of a couple of µm the PEF is comparable to the force from the K-fibre , although the linear approximation for the PEF may lose validity at displacements over this distance [37] . Fig 4F shows the estimated probabilities of staying coherent , pc ( blue ) , and staying incoherent , pic ( red ) , between consecutive frames , with mean frame lengths of 6 . 3 ± 8 . 7 and 3 . 0 ± 1 . 0 for coherent and incoherent periods . The strong skew towards 1 of pc demonstrates that coherent runs are stable; it is relatively unlikely to switch states between consecutive frames giving rises to extended runs . Naturally , this effect is less pronounced for noisier trajectories since they would be expected to make fewer coherent runs or switch more often ( S3E and S3F Fig ) . The distribution of the noise precision τ ( = s−2 ) showed a range from close to zero to 800 s2 μm−2 ( where larger numbers indicate less noise; Fig 4D ) . Our analysis of trajectory stochasticity had already shown that individual cells harbour inter-trajectory heterogeneity ( Fig 3B , 3D and 3F ) . Kinetochore pairs located nearer the centre of the cell had more oscillatory trajectories as judged by EV ( Fig 3H ) , consistent with previous observations that oscillatory trajectories tend to be more centrally located within the metaphase plate [44] . To ascertain if this heterogeneity is also reflected in the dynamic parameters we explored parameter trends with respect to distance from the centre of the metaphase plate r ( Fig 4G–4L ) . The most significant trends were with regard to the trajectory noise ( ρ = −0 . 43 , p < 10−37; Fig 4J ) and PEF strength coefficient ( ρ = 0 . 20 , p < 10−8; Fig 4I ) which almost doubles between the centre and periphery . The former correlates with stochasticity as measured by EV ( Fig 3H ) , τ and EV being highly correlated ( ρ = 0 . 57 , p < 10−73 ) . This increase in stochasticity with r also explains the increase in the probability of switching direction per frame with r ( ρ = −0 . 27 , p < 10−14 when coherent , ρ = −0 . 37 , p < 10−28 when incoherent; Fig 4L ) , kinetochore movements becoming more random and liable to switch direction per frame . There was no significant dependence on metaphase plate position for the other parameters: v± , L and κ . From the parameter estimates and inferred K-fibre state determined for each trajectory , we can use Eq ( 3 ) to calculate effective contributions to the total force on each kinetochore ( recall that in our model the parameterisation of forces and velocities are commensurate; we quote forces in μm s−1 ) . Specifically we can estimate the contributions of the 3 component forces: the PEF , the K-fibre forces and the spring force ( Fig 5A ) throughout the trajectory . In Fig 5B and 5C we show the forces acting on one sister ( sister 1 ) of the sister pair of Fig 1C; Fig 5D and 5E show the same for the second sister ( sister 2 ) . The K-fibre force on sister 1 was by far the dominant force on the kinetochores and because it corresponds to the K-fibre state , it changes sign at directional switches ( Fig 5B ) . The PEF and spring force were a minor contribution , and were in phase every period , the spring force having double the period of the oscillation as previously observed [8 , 12] . A comparison between the sisters ( Fig 5B and 5D ) clearly demonstrates the approximate anti-phase of the K-fibre state , although this is not exact and extended periods of incoherence did occur; in this example , both sisters have polymerising K-fibres during incoherence . This is reflected in the spring force profile ( Fig 5C and 5E ) since +/+ incoherence causes relaxation of the spring to near zero spring tension and apparent compression ( Fspring < 0 ) during the extended period of incoherence . The PEF force , being positional and linearised in our model , corresponds to the kinetochore position thereby reflecting the off-plate position of the sister pair ( Fig 1C ) .
In this paper we have presented a framework for the ( Bayesian ) reverse engineering of individual paired kinetochore trajectories utilising a mathematical model of kinetochore oscillations . This tool enables a whole swathe of new analyses to be performed across a range of levels , specifically: ( i ) The processing of kinetochore tracks , locating switch points and partitioning trajectories into coherent and incoherent phases , thereby facilitating interpretation of this complex , stochastic dynamical system . ( ii ) A high-throughput semi-automated analysis at the level of populations of kinetochore trajectories , generating a range of population level statistics . There is no typical trajectory since there is both a high variability in trajectory stochasticity ( here and [8] ) and heterogeneity within cells , evidenced by significant parameter trends across the metaphase plate ( Fig 4 ) that are only evident upon analysis of 1000s of trajectories . ( iii ) Force inference and identification of mechanistic signatures , thereby revealing previously inaccessible mechanistic information . Our methodology complements forward simulation analyses ( reviewed in [33 , 34 , 47] ) where models of kinetochore dynamics are developed from knowledge and assumptions pertaining to the biological processes and are demonstrated to qualitatively or semi-quantitatively reproduce observed behaviour under suitable parameter values . However , our analysis of 2 second resolution 3D live-cell tracking data has revealed new depths of complexity in kinetochore dynamics and choreography , substantially adding to the qualitative and quantitative constraints that these models must satisfy , including trail sister initiated switching , a bias towards lead sister initiated switching , and spatial parameter trends . Reproducing these probabilistic behaviours is likely to require the incorporation of new mechanistic or regulatory processes . Moreover , we have demonstrated that data-driven reverse engineering methods are able to fit biologically meaningful mechanistic models to real trajectory data . Developing more realistic mechanistic models within this framework is the next challenge; the model selection methods we have implemented here will be invaluable in determining whether mechanistic hypotheses are supported by data . The kinetochore dynamics model we have proposed here is parsimonious , only incorporating the three essential forces believed to affect kinetochore positioning , whilst simplifying the form of these forces . Firstly , we have treated each K-fibre coarsely as producing a single unified force . K-fibres are bundles of microtubules and therefore both the number of microtubules attached to the kinetochore varies [48] and the balance between the number of polymerising and depolymerising microtubules will change as they individually undergo catastrophe and rescue events [19] . How kinetochores control the dynamic instability of the attached microtubules , thereby giving rise to K-fibres that are either polymerising or depolymerising remains a mystery . Incorporating this level of complexity is not possible within the reverse engineering framework since individual kinetochore trajectory dynamics do not have enough information to determine the K-fibre composition . However , since there is typically a majority of either polymerising or depolymerising microtubules within each K-fibre [19 , 42] , the coarse-grained model is a reasonable approximation . The composition fluctuations , that are likely to produce deviations from this constant force approximation will inflate the noise in the fits of our model to data . In the future , it may be possible to combine quantification of fluorescent tubulin ( number of microtubules ) and fluorescent EB proteins ( fraction of polymerising microtubules [19] ) proximal to the kinetochore with trajectory data to incorporate these effects into the reverse engineering . Secondly , we have assumed that the PEF is linear around the metaphase plate . This is well-supported experimentally for the size of displacements typical of metaphase kinetochores [37] . Incorporating a nonlinear form for the PEF ( stronger at the poles ) is feasible , although it is likely that kinetochore trajectories that explore locations closer to the spindle poles will be needed to obtain reliable estimates of that nonlinearity . This would be true in prometaphase , where kinetochores congress to the plate from all over the mitotic spindle , and may be a crucial change necessary for the model to account for this phase . Thirdly , we have assumed that the centromeric spring is a linear ( Hookean ) spring . Recently , it has been suggested that this spring is nonlinear [49] and by incorporating this into the reverse engineering model it may be possible to determine the nature of the nonlinearity and any asymmetry in the spring . More complex models , such as visco-elastic models are likely beyond the scope of what is identifiable with trajectory data alone . As it stands the model is simple enough to allow parametrisation from individual trajectories subject to additional data on the natural length of the inter-sister spring being supplied . Our reverse engineering approach using this model allows behavioural and mechanistic ( force ) variability amongst trajectories to be quantified at a previously inaccessible level , in particular spatial trends could be detected in the force components across the metaphase plate ( Fig 4G–4L , Fig 9 ) . Our analysis of the quality of the fit indicates that kinetochore oscillations are well-explained by this simple force balance model , with explained variance ( EV ) reaching as high as 66% , similar to levels achieved on simulated data using estimated parameter values ( not shown ) . A minority of trajectories show very strong preference for this model over a simple Brownian motion model ( Fig 3C , 3D and 3G ) . This indicates that displacements at the 2 second level retain a strong signature of the underlying dynamics , although as trajectory stochasticity increases the displacements lose that signature and conform to a Gaussian distribution . Correlations between consecutive displacements are however retained for the majority of trajectories , even for highly stochastic trajectories ( Fig 3G ) , suggesting that directional switching of K-fibres drives kinetochore motion even when highly stochastic . Reverse engineering is able to provide crucial biological insight into complex , heavily regulated systems such as kinetochore dynamics . Firstly , directional switching has two choreographies , lead initiated directional switching ( LIDS ) and trail initiated directional switching ( TIDS ) , with the latter inhibited in the eGFP-CENP-A/eGFP-Centrin1 cell line examined here ( bias 4:1 , Fig 6B ) . Secondly , the spring force is weak , constituting typically less than 15% of the total force acting on a kinetochore and on average rising only to 20% at switching . Even combined with the PEF , which both oppose the pulling force on the lead kinetochore by its K-fibre , a stall in the lead kinetochore under force equilibration is extremely rare ( Fig 8C ) . This means that kinetochore switching is not a result of a tug-of-war between opposing forces ( as various models predict [35 , 50] , and as occurs in Drosophila [51] ) . In effect , both kinetochores move at 13 to 35 nm s−1 with directional switching events essentially preventing the inter-sister distance increasing too high , thereby keeping the spring tension low . This suggests that tension is utilised as a means to regulate switching , as suggested in [7 , 12 , 45 , 46] , to both prevent occurrence of a high spring tension—which could potentially cause separation of the sister chromatids—and to localise the kinetochore at the plate . Thirdly , by aligning many hundreds of switching events , we revealed a distinct switching signature which provides experimental validation of the lead-sister induced switch model [7 , 12 , 45 , 46] . Using the inferred model parameters , we estimate the relative contributions of each force component prior to and across the switching events . This clearly revealed the spring force as the minor force component throughout the profile ( Figs 7 and 8 ) ; kinetochore dynamics are dominated by K-fibre forces . The increase in spring force leading up the directional switch most likely triggers the switch [7 , 12 , 45 , 46] . However , a substantial revision of these ideas is required to unify this LIDS mechanism with the alternative TIDS choreography ( see [40] ) . Fourthly , as previously reported for Ptk1 cells [44] , the most strongly oscillating trajectories are located in the centre of the metaphase plate ( Fig 3H ) . We show that there are concomitant changes in the switching rates with distance from the centre ( Fig 4L ) , and an increase in the PEF ( Figs 4I , 9A and 9B ) correlates with increased plate thinning with distance from the centre of the metaphase plate and damping of oscillations ( Fig 9C–9E ) . This is in agreement with a laser microsurgery study which demonstrated that reducing the PEF increases the amplitude of oscillations [37] and concurs with the proposal that the PEF is the cause for the loss of oscillations towards the plate periphery in [44] . Here we quantified the inferred forces in terms of speed ( μm s−1 ) . This is because we have no reliable estimate of the effective drag coefficient γ of the kinetochore/chromatid system; non-thermal contributions to noise from mechanical fluctuations imply estimation from τ would result in significant underestimation . If it were possible to robustly determine γ experimentally , we would be able to separate the current τ into intrinsic active and external molecular stochasticity , and directly determine the forces . To obtain estimates of absolute forces we can appeal to a simple approximation using Stokes’ law [3] , assuming a chromosome radius of 0 . 5 μm ( estimated from volume measurements [52] ) and a spindle viscosity of 190 Pa s[53] . This yields mean absolute forces for the three mechanical components of F− = 62 pN , F+ = 23 pN , Fspring = 6 . 9 pN , and FPEF = 15 pN . The calculated force from a P-moving kinetochore ( F−; depolymerising ) is compatible with the 50 pN stall force measured by Nicklas in pre-anaphase cells [2] . This indicates that our reverse engineering is extracting physiologically meaningful forces . Our reverse engineering approach raises significant questions . Firstly , how is kinetochore directional switching regulated to generate the observed pseudo-periodic oscillations incorporating both LIDS and TIDS choreographies , with TIDS being inhibited relative to LIDS . Secondly , how is this switching accomplished under low spring tension , that rarely , if ever , achieves a kinetochore stall . Thirdly , what changes in the spindle explain the increase in trajectory stochasticity and the PEF with distance from the centre of the metaphase plate . Future work is needed to tie down the biochemical and mechanical processes that underpin these behaviours . We expect reverse engineering will be invaluable in this endeavour to both quantify how the behaviour changes with genetic perturbations and also for fitting increasingly complex models of the kinetochore-microtubule interface . | To achieve proper cell division , newly duplicated chromosomes must be segregated into daughter cells with high fidelity . This occurs in mitosis where during the crucial metaphase stage chromosomes are aligned on an imaginary plate , called the metaphase plate . Chromosomes are attached to a structural scaffold—the mitotic spindle , which is composed of dynamic fibres called microtubules—by protein machines called kinetochores . Observation of kinetochores during metaphase reveals they undergo a series of forward and backward movements . The mechanical system generating this oscillatory motion is not well understood . By tracking kinetochores in live cell 3D confocal microscopy and reverse engineering their trajectories we decompose the forces acting on kinetochores into the three main force generating components . Kinetochore dynamics are dominated by K-fibre forces , although changes in the minor spring force over time suggests an important role in controlling directional switching . In addition , we show that the strength of forces can vary both spatially within cells throughout the plate and between cells . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Inferring the Forces Controlling Metaphase Kinetochore Oscillations by Reverse Engineering System Dynamics |
The potential of broadly neutralizing antibodies targeting the HIV-1 envelope trimer to prevent HIV-1 transmission has opened new avenues for therapies and vaccines . However , their implementation remains challenging and would profit from a deepened mechanistic understanding of HIV-antibody interactions and the mucosal transmission process . In this study we experimentally determined stoichiometric parameters of the HIV-1 trimer-antibody interaction , confirming that binding of one antibody is sufficient for trimer neutralization . This defines numerical requirements for HIV-1 virion neutralization and thereby enables mathematical modelling of in vitro and in vivo antibody neutralization efficacy . The model we developed accurately predicts antibody efficacy in animal passive immunization studies and provides estimates for protective mucosal antibody concentrations . Furthermore , we derive estimates of the probability for a single virion to start host infection and the risks of male-to-female HIV-1 transmission per sexual intercourse . Our work thereby delivers comprehensive quantitative insights into both the molecular principles governing HIV-antibody interactions and the initial steps of mucosal HIV-1 transmission . These insights , alongside the underlying , adaptable modelling framework presented here , will be valuable for supporting in silico pre-trial planning and post-hoc evaluation of HIV-1 vaccination or antibody treatment trials .
Recent years have seen tremendous success in the isolation and characterization of broadly neutralizing antibodies ( bnAbs ) from selected HIV-1 infected patients . By binding to the HIV-1 envelope glycoprotein trimer ( Env ) , bnAbs succeed to neutralize a majority of circulating HIV-1 strains . It is assumed that the elicitation of antibodies will constitute a crucial component of a successful HIV-1 vaccination strategy , and known bnAbs are intensely explored as templates for HIV-1 vaccine development [1–5] . Indeed , it has been conclusively demonstrated in animal models that passive immunization with bnAbs can protect against virus challenge , delay viral rebound and transiently lower viremia [6–19] . Furthermore , passive immunization in human patients demonstrated an impact of bnAbs on established HIV-1 infection [20–22] , underscoring the potential relevance of bnAbs to prevent or treat HIV-1 infection . However , despite this wealth of information on the protective effects of bnAbs in vivo , key parameters of the HIV-1 nAb interaction and HIV-1 host-to-host transmission remain ill-defined . This concerns both fundamental molecular aspects of Env trimer-nAb binding and systemic factors of mucosal HIV-1 transmission . Importantly , comprehensive knowledge of the molecular and systemic parameters governing HIV-1 transmission and nAb neutralization would empower in silico modelling of nAb activity and be instrumental to guide vaccine development or nAb treatment trials [23 , 24] . We thus propose that precise numerical quantification of the parameters that steer nAb efficacy and in vivo HIV-1 transmission is needed . Moving towards this aim , we report here on a combined experimental-mathematical analysis providing comprehensive quantitative insight into mucosal HIV-1 transmission and nAb neutralization ( Fig 1 ) . Starting at the molecular level , the first question we addressed regards the number of nAbs required to neutralize each HIV-1 Env trimer ( the stoichiometry of neutralization , N ) . This number , in combination with the mean number of trimers per HIV-1 virion ( η¯ ) and the number of trimers required for virus entry ( T , Fig 1A ) defines the number of nAbs needed to neutralize single HIV-1 virions or entire virion populations [25–31] . While such stoichiometric definitions may appear academic given the well-known potential of nAbs to neutralize HIV-1 , we show here that these parameters are indeed crucial for an in-depth understanding of HIV-1 nAb neutralization and enable mathematical predictions of nAb neutralization potency . Moving to the systemic level , we next addressed uncertainties in our understanding of mucosal HIV-1 transmission . Non-human primate studies revealed the interplay between nAb potency , in vivo nAb concentrations , and the resulting susceptibility or protection against virus infection [6–19] . However , a detailed systemic understanding of the mucosal infection process and the factors resulting in nAb protection from infection , ideally down to the single-virion level , are missing . Utilizing our stoichiometric model framework we performed a post-hoc analysis of selected animal studies , and obtained precise quantitative insight into mucosal nAb neutralization and the probability for single infectious HIV-1 virions to establish a systemic host infection . Lastly , significant uncertainty is associated with the process and the probabilities of mucosal HIV-1 transmission in the human host via sexual contact . Per-exposure risk estimates of HIV-1 transmission vary widely , and uncertainty prevails regarding the concentration of nAbs in genital mucosal tissues that would provide protection from infection . This is not surprising , given the difficulties in estimating these parameters directly in the human population [32 , 33] . Thus , in a final step we build on all previously determined parameters to model human HIV-1 penile-vaginal transmission . This analysis yielded predictions of HIV-1 male-to–female per intercourse transmission probabilities that match empirical data , and provided estimates of mucosal nAb concentrations expected to provide protection from HIV-1 infection .
The topic of HIV-1 neutralization stoichiometry has previously been studied experimentally and computationally , with N = 1 being the most common conclusion [26 , 29 , 34] . However , ambiguity in these estimates prevailed , predominantly due to previously poorly defined parameters of HIV-1 entry stoichiometry , T , and mean virion trimer number , η¯ , which are essential for the analysis of N [27–29] . Here , we utilized recently determined values of T and η¯ [28] to conclusively estimate N across different HIV-1 strains and bnAbs identified in recent years , for which estimates of N are currently lacking . As precise information on N across nAbs is indispensable for an in-depth numerical understanding of HIV-1 nAb neutralization ( Fig 1A ) , we set out to estimate N for a range of nAbs and HIV-1 strains . To estimate N , we measured nAb neutralization of HIV-1 pseudoviruses carrying mixed trimers of nAb-sensitive and resistant Envs and analyzed the data with mathematical models ( Fig 2A ) . Relative virus infectivities ( RI ) under saturating nAb concentrations were set in relation to the fraction of resistant Env ( fR ) , shown for nAb 2F5 with Envs JR-FL wt ( 2F5 sensitive ) and mutant JR-FL D664N ( 2F5 resistant ) ( Fig 2B , S1 Fig , S2 Table ) . Taking the JR-FL entry stoichiometry ( T = 2 ) and mean trimer number per virion ( η¯=11 . 8 ) into account ( S3 Table ) , our model predicts different RI profiles for different N ( Fig 2C ) . Mathematical analysis of the experimental data indicated a neutralization stoichiometry of N = 1 for JR-FL by nAb 2F5 ( Fig 2C ) . We confirmed the robustness of the N = 1 estimate against variation in virus entry stoichiometry ( T ) and mean virion trimer number ( η¯ ) by sensitivity analyses ( Fig 2D ) . To investigate the observed deviations of the model fit from the experimental data ( Fig 2C ) , we performed a goodness-of-fit analysis ( Fig 2E ) . This analysis indicated that lower values of T and/or η¯ could result in better curve fits . While we expect T to be constant for each viral strain , fluctuations in η¯ from experiment to experiment are conceivable and provide a potential explanation for the deviation between experimental data and model predictions . In addition , mean virion trimer numbers of a given virus preparation may decrease over time as spontaneous Env shedding can occur resulting in non-functional trimers [36] . To further test variation in T and η¯ on predictions of N , we assessed N for nAb 2F5 against Env variants of HIV-1 strains JR-FL and NL4-3 that differ in T or η¯ ( S2A–S2F Fig , S3 Table ) [28 , 37] . We indeed observed divergent RI profiles upon variation in T or η¯ as predicted by our model . However , for all Env variants tested we retrieved estimates of N = 1 . To explore if nAb avidity influences estimation of N , we compared nAb 2F5 IgG and Fab fragment . Both yielded overlapping RI profiles for strains JR-FL and NL4-3 and identical estimates of N = 1 ( S2G–S2I Fig ) , confirming that estimation of N is independent of nAb valency . Having validated our approach to estimate N ( Fig 2 ) , we sought to obtain a comprehensive analysis of N for various HIV-1 nAbs including VRC01 , NIH45-46 , PGV04 , b12 , PGT121 , PGT128 , PGT135 , 2G12 , PG9 , PGT145 , and 2F5 ( S4 Table ) . We therefore generated a panel of Env mutants in five divergent HIV-1 strains with single or combined nAb resistance mutations ( S1 Fig ) . Several of these Env mutants showed a significant reduction in virus infectivity ( S2 Table ) . This is critical to note , as matched virus entry parameters , notably T and η¯ , are a prerequisite for the analysis of N ( S2 Fig ) . Indeed , we observed that strong infectivity differences between nAb-resistant and sensitive Envs in mixed trimer assays result in substantial deviations of the RI profiles that prohibit determination of N ( S3 Fig ) . We thus restricted our analysis to mixed trimer assays with nAb sensitive-resistant Env pairs of comparable infectivity ( ≤ 2-fold infectivity difference ) . In a direct comparison of eight different nAbs across five Env mixed trimer combinations , we thereby obtained a consistent estimate of N = 1 irrespective of nAb epitope specificity , potency or HIV-1 strain ( Fig 3A–3E , S4 Fig ) . The estimate of N = 1 was confirmed by bootstrap analyses ( S5 Fig ) and goodness-of-fit plots across all analyzed Env-nAb pairings ( S6 Fig ) . As shown before ( Fig 2E ) , the goodness-of-fit would in many cases be improved for lower values of T and/or η¯ , possibly indicating slight deviations in these parameters within the experimental setup . In the above analysis we employed a hard threshold model in which , according to our analyses , binding of one nAb to a trimer results in complete loss of function . Soft threshold models , which allow for partial loss of trimer functionality upon nAb binding , were introduced in earlier studies by us and others [26 , 29] . To investigate how a soft threshold model would fit to our current data set , we re-analyzed our data accordingly [26] . Here , we allow the probability of a virion to infect a cell to scale with the number of functional trimers ( a soft threshold for virus entry ) and the functionality of a trimer to decrease by successive nAb binding ( a soft threshold for neutralization ) . Our analysis revealed that the functionality loss of a trimer upon nAb binding to one subunit is dominant , that is , a soft threshold model is not supported ( S7 Fig ) . This finding thereby underscores the assumption of a hard threshold for the stoichiometry of HIV-1 nAb neutralization . To validate that N = 1 across all nAbs tested here , we tested a set of nAbs targeting different epitopes against Env variants with multiple nAb escape mutations , allowing parallel analysis of diverse nAbs on the same set of mixed trimer virus stocks ( Fig 3F and 3G , S8 Fig , S2 Table ) . Two Env combinations ( Fig 3F and 3G ) were infectivity matched and allowed mathematical analysis of N , yielding N = 1 for all nAbs tested . To derive an estimate of N in the non-infectivity matched setups we included nAb 2F5 throughout , as we previously established that it neutralizes with N = 1 ( Fig 2 ) . Furthermore , nAb PG9 , which is known to bind with only one nAb per trimer [38 , 39] , was included in two of these setups ( S8 Fig ) . As 2F5 and PG9 yielded highly similar curves compared to the other nAbs , we conclude that our analysis univocally inferred N = 1 for all probed nAbs . Regarding this consistent N = 1 estimate across all nAbs tested , we asked whether non-antibody Env inhibitors would show a similar neutralization behaviour . To test this , we generated Env mutants resistant against the peptide T-20 , a clinically used HIV-1 entry inhibitor targeting the Env gp41 subunit [40] . Using mixed trimer assays , we found that T-20 neutralizes with an N = 1 stoichiometry ( S9 Fig ) . This indicates that , regardless of inhibitor type , interference with one Env trimer subunit is sufficient for HIV-1 trimer neutralization . Antibody responses in the majority of HIV-1 infections are largely ineffective in neutralizing the virus within the patient , and bear only limited neutralization potency and breadth against other HIV-1 strains . In contrast , the bnAbs available to date were isolated from rare HIV-1+ patients with high HIV-1 neutralization potency and breadth [41–43] . In our stoichiometry analysis , these bnAbs uniformly showed N = 1 ( Fig 3A–3G ) . This raises the question if weakly neutralizing Abs as most commonly elicited during HIV-1 infection require a higher N , and whether N = 1 is a distinguishing feature of bnAbs . To test this , we determined N of the polyclonal antibody mix in HIV-1+ patient plasma . We first tested plasma from an individual with typical HIV-1 neutralization escape [44] and corresponding plasma neutralization-resistant and sensitive Env variants ( ZA110 wt and ZA110-V1V21 . 7 , respectively; S2 Table ) . Mixed trimer assays yielded N = 1 for the plasma Abs of this individual ( Fig 3H ) . We further tested plasma from three chronically HIV-1 infected individuals ( Pat117 , Pat118 , Pat122 ) showing only weak HIV-1 neutralization activity . We tested these plasma on mixed trimer stocks of Envs ZA110 wt ( resistant ) and ZA110 ΔV1V2 ( sensitive ) and further included weakly neutralizing Abs b6 , 17b and 48d in this analysis ( Fig 3I ) . Comparison of the RI profiles indicated identical N for all probed plasma and nAbs . Additionally , plasma Pat122 neutralized HIV-1 strain JR-FL in an infectivity-matched setup with N = 1 , as did several nAbs ( Fig 3J ) . Thus , irrespective of nAb potency , breadth or epitope specificity , neutralization of HIV-1 trimers requires only a single Env subunit to be bound by antibody ( Fig 3K ) . This estimation of N defines numerical requirements for HIV-1 antibody neutralization [27 , 29–31 , 45] , which we employed in subsequent modelling steps to assess HIV-1 virion population neutralization in vitro and in vivo . We previously established a mathematical framework that predicts the number of nAbs required to neutralize a given HIV-1 virion population based on the stoichiometry parameters N , T and η¯ [27] . The conclusive estimation of N ( Figs 2 and 3 ) , together with previously determined parameters T and η¯ [28] , now enabled us to use our framework for quantitative predictions of nAb neutralization . We extended the framework by including the affinity of nAb binding to Env trimers ( represented by the binding constant , KD ) [35 , 46] to predict the fraction of neutralized HIV-1 virions for given nAb concentrations . In essence , this allowed us to simulate HIV-1 nAb neutralization curves in silico ( Fig 4A ) . For this analysis we utilized nAb binding constants , KD , recently reported for the HIV-1 strain BG505 SOSIP trimer [49] . We employed these KD values together with T and η¯ of BG505 ( S3 Table ) and N = 1 to model nAb neutralization of an HIV-1 BG505 virion population ( Fig 4B–4D ) . As expected , we found that nAbs with high Env binding affinity ( low KD ) are predicted to require lower concentrations to achieve virion population neutralization ( Fig 4B ) . The required nAb concentrations increase slightly with fewer trimers needed for HIV-1 entry ( low T ) ( Fig 4C ) and higher virion trimer content ( high η¯ ) ( Fig 4D ) . The latter two trends can be rationalized as follows: in both cases ( a small T or a high η¯ ) , more nAbs will be needed to neutralize a given virion population , resulting in higher predicted nAb concentrations . The relation between nAb trimer binding and HIV-1 virion population neutralization is influenced by all parameters included in our model ( Fig 4A ) . Especially the influence of T and η¯ on nAb neutralization predictions should not be underestimated . This is highlighted in Fig 4E , depicting estimated values of nAb concentrations required for 50% virus population neutralization ( IC50 ) in dependence on T and η¯ . In addition to predicting nAb neutralization curves and IC50 values , we determined nAb concentrations required to achieve sterilizing neutralization of HIV-1 virion populations . Due to unproductive nAb binding to Env trimers ( Fig 4H ) [27] , the fraction of Env subunits required to be bound by nAb for sterilizing neutralization of a virion population increases with virion population size ( Fig 4F ) . Likewise , the predicted nAb concentrations required for sterilizing neutralization of virion populations increase with virion population size ( S10A–S10C Fig ) . As shown in Fig 4B–4D for nAb neutralization curves , these nAb concentrations are influenced by nAb KD , T and η¯ ( S10A–S10C Fig ) . Of note , only nAb KD has a direct linear relationship with sterilizing nAb concentrations , while the influence of T and η¯ follows non-linear relations ( S10D–S10F Fig ) . We noted that our predicted neutralization curves show steeper slopes than commonly observed in HIV-1 neutralization assays ( S11A Fig ) . We thus asked which parameters of our model could explain this deviation , and found that assuming broader virion trimer number distributions ( i . e . , higher variance in trimer numbers between virions ) results in less steep predicted neutralization curves ( S12 Fig ) . Of note , we also observed that our predicted curves are in closer agreement to in vitro neutralization data obtained with replication-competent virus and PBMC target cells ( S11B Fig ) . While a detailed analysis of this relation and the underlying parameters is beyond the scope of this manuscript , this observation may be taken as indication that our model predicts neutralization curves better for replication-competent virus than for pseudovirus preparations . To test the predictive power of our in silico approach , we compared experimentally derived IC50 values of HIV-1 strain BG505 and nAbs VRC01 , PGV04 , PGT121 , PGT123 , PGT145 and 2G12 , originating from in vitro experiments by three independent laboratories [47–49] , with nAb IC50 values estimated by our model ( Fig 4G , S11 Fig , S5 Table ) . Our predicted IC50 values for nAbs VRC01 and PGT121 were in close agreement with the measured values . For nAbs PGT123 and 2G12 , we observed wide variations in experimentally derived IC50 values; interestingly , the IC50 values predicted by our model lay in between these values . For nAbs PGV04 and PGT145 , our estimated IC50s are slightly lower and higher than experimentally determined IC50s , respectively , though not far off ( see below for a detailed discussion of the PGT145 IC50 estimation ) . This good agreement between experimental and predicted nAb IC50s highlights that the parameters included in our model ( N , T , η¯ , nAb KD and nAb concentration ) capture relevant steps of HIV-1 virion neutralization . Of note , our model performs significantly better than a simpler model based on nAb KD alone ( S13 Fig ) . We also derived predictions of nAb neutralization for scenarios of N = 1 but assuming that exclusively one nAb can bind one of the three epitopes displayed on each trimer ( a binding behavior described for nAbs PG9 , PG16 or PGT145 [38 , 39 , 49] ) . We compared these predictions to our standard model where we assume that all three trimer subunits can potentially be bound , although only one Env subunit needs to be bound to achieve neutralization . In this case , binding of the second and third nAb represents unproductive binding of nAbs , since the trimer is already neutralized by binding of the first nAb ( Fig 3K ) . Importantly , we assume that also for PG9-like nAbs , initially the same number of epitopes is present as for “typical” nAbs ( i . e . , three per trimer ) , with the difference that binding of PG9-like nAbs shows negative cooperativity: after binding of the first antibody the remaining two epitopes are inaccessible . In our analysis , we assume two nAbs with the same KD for their protomeric epitopes , but either a three-nAb-per-trimer or one-nAb-per-trimer binding behavior ( Fig 4H ) . According to our predictions , binding of only one nAb per trimer results in a 2-fold decreased IC50 , thereby demonstrating the effect of unproductive nAb binding ( Fig 4H ) . Intriguingly , the PGT145 IC50 predictions using this model are closer to experimental PGT145 IC50 values than the predictions obtained with the standard model ( Fig 4G ) . This indicates that nAbs with a one-nAb-per-trimer binding behavior should have a slight neutralization advantage , especially under conditions of low nAb concentrations . We next predicted nAb neutralization efficacy and host infection probability in vivo by re-assessing data from four studies of rhesus macaque vaginal virus challenge following passive immunization with nAbs b12 [8] , 2G12 [7] , PGT121 [10] and PGT126 [16] . Our approach consists of two parts: first , we utilized the nAb neutralization prediction model developed above ( Fig 4 ) to estimate how many virions of the challenge dose are neutralized in vivo in dependence on the nAb immunization regime . Secondly , we connected this number of non-neutralized virions to the number of animals protected or infected for a given immunization and challenge regime . Ultimately , this allows us to derive the probability for a single infectious virion to establish a host infection ( Fig 5A ) . This modelling of HIV-1 neutralization in vivo following vaginal challenge requires data on vaginal nAb concentrations and nAb KD as well as the virus-specific parameters of inoculum size , η¯ , T and N . Importantly , we also require the probability of virions to penetrate mucosal epithelial layers to come in contact with target cells ( ppen ) ; this parameter was recently estimated elegantly by Carias et al . [50] . The four selected macaque studies reported the majority of nAb and virus-specific parameters ( S6 and S7 Tables ) , allowing us to employ our modelling approach in a post-hoc analysis . Essential data of the four analyzed studies are listed below: To derive nAb KD data for the challenge virus Env P3 used in all four studies , we utilized the known IC50 for 2G12 , b12 , PGT121 and PGT126 against P3 and inferred the corresponding KD values by linear regression ( S6 and S7 Tables ) . We further assumed that virus strain P3 has T = 2 and η¯=20 as previously determined for the closely related strain P3N ( S3 Table ) . These data enabled us to predict neutralization curves for the four nAbs against SHIV-P3 ( S14 Fig ) . We then superimposed these nAb neutralization curves with the estimated nAb concentrations in the vaginal mucosa at the time of challenge . In this analysis , we allowed both nAb KD and mucosal nAb concentrations to vary 2-fold around the estimated values to account for potential inaccuracies in the extrapolation procedures . This analysis yielded windows for the extent of SHIV-P3 inoculum neutralization by the four nAbs in vivo ( S14A Fig , colored parts of the neutralization curves ) . We predicted almost complete SHIV-P3 inoculum neutralization for the highest PGT121 and PGT126 immunization regimes , while the b12 and 2G12 immunizations and the low doses of PGT121 and PGT126 immunizations yielded intermediate SHIV-P3 neutralization levels , mirroring the protective effects seen in the respective challenge studies [7 , 8 , 10 , 16] . Importantly , our model also allowed us to determine the probability that a single infectious virion starts a host infection . To this end , we first calculated the fraction of SHIV-P3 virions that remained potentially infectious in the four macaque challenge studies , i . e . virions with at least T non-neutralized trimers ( S14B Fig ) . Multiplying these fractions of non-neutralized virions with the total number of virions in the respective challenge inocula provided an estimate of the number of virions that remained potentially able to infect each animal ( Fig 5B ) . To investigate how this number of non-neutralized , infectious virions translates into systemic host infection we utilized a further parameter . During vaginal challenge not all virions of the inoculum will penetrate the vaginal epithelial layers to come in contact with mucosal CD4+ target cells; the probability of epithelial penetration is given by ppen . We assumed that only 0 . 235% of virions will penetrate the genital tract epithelium as experimentally estimated [50] . Based on this , we derived the number of infectious virions in the four challenge studies that can potentially contact mucosal target cells and set these number in relation to the observed infection outcomes ( Fig 5A ) . This delivered the probability for a single infectious virion to establish a systemic host infection , denoted ψ ( Fig 5C ) . Intriguingly , we obtained closely matching ψ-estimates for the four independent macaque studies , with an average value of ψ^=1 . 65×10−5 . Next , we asked whether a similar analysis would be possible using in vitro nAb neutralization data ( i . e . , nAb IC50 and Hill parameter ) . For this analysis we utilized in vitro neutralization data of SHIV-P3 with nAbs PGT121 , PGT126 and b12 ( S11B Fig ) and analyzed the respective macaque challenge studies . We obtained a closely matching estimate of ψ , i . e . ψ^=2 . 95×10−5 ( S15 Fig ) as with our mechanistic population neutralization model ( Fig 5C ) . Based on these two complementary analyses , we conclude that only one in ~30 . 000 to 60 . 000 infectious virions that have penetrated the vaginal epithelial layers will succeed in systemically infecting the host . This estimate thereby provides a quantitative evaluation of the bottlenecks encountered by HIV-1 during transmission in the genital mucosa . We next used the estimate of HIV-1 virion host infection probability ( ψ^=1 . 65×10−5 ) to predict male to female HIV-1 transmission risk per sex act and nAb protection efficacy . In this analysis , we define the per-act HIV-1 inoculum as the number of HIV-1 virions per ejaculate; this number is given by per-ejaculate semen volume and semen viral load . We then assumed that only a fraction of virions in the inoculum will penetrate the vaginal epithelial layers , defined by ppen [50] . We next determined the number of virions that are potentially infectious , i . e . have at least T trimers , and multiplied the fraction of penetrating infectious virions with the probability for each infectious virion to initiate host infection , ψ . Thus , we obtained the probability of host infection in dependence on HIV-1 inoculum size . In addition , we modelled protective effects of nAbs present in the vaginal mucosa; here , the extent of HIV-1 inoculum neutralization ( and hence the magnitude of protection ) depends on mucosal nAb concentration and nAb binding affinity ( KD ) for the Env trimer . In a first step , we calculated infection probabilities of women during a single penile-vaginal intercourse in dependence on HIV-1 inoculum size , in absence of nAbs . We performed this analysis for HIV-1 virions with the entry characteristics of the transmitted-founder strain BG505 ( η¯=9 . 5 trimers per virion , T = 2 entry stoichiometry; S3 Table ) . The resulting relation between HIV-1 inoculum size and per-act female host infection probabilities is shown in Fig 6A . This relation needs to be interpreted with regard to empirical data of human HIV-1 semen viral loads and per-act transmission risk . In chronic HIV-1 infection , semen viral load typically ranges between ≤ 200 and 20 . 000 HIV-1 RNA genome copies per mL seminal plasma ( i . e . , ≤ 100 to 10 . 000 virions / mL , assuming two viral genomic RNA copies per virion ) . During acute infection , semen viral load typically ranges between 20 . 000 and 200 . 000 RNA copies / mL ( 10 . 000 to 100 . 000 virions / mL ) [33] . In rare cases , semen viral loads of several million RNA copies / mL ( >1 . 000 . 000 virions / mL ) were reported [51] . Assuming an average per-ejaculate semen volume of 3 mL [51 , 52] , this results in typical HIV-1 inoculum sizes of ≤ 300 to 30 . 000 virions during chronic HIV-1 infection , 30 . 000 to 300 . 000 virions during acute infection , and > 3 . 000 . 000 virions in rare cases . The corresponding per-act female infection probabilities predicted by our model are ≤ 0 . 001 to 0 . 11% during chronic infection ( one in ≥ 87 . 000 to one in 876 sexual contacts ) , 0 . 11 to 1 . 14% during acute infection ( one in 876 to one in 88 sexual contacts ) , and maximum values exceeding 11% ( one in 9 sexual contacts ) ( Fig 6A ) . How do these estimates compare to empirical data on female per-act infection probabilities ? A frequently stated number for male-to-female penile-vaginal transmission risk is 1 in 1000 sexual contacts ( 0 . 1% ) . However , as previously discussed [32 , 53 , 54] , this likely represents a lower bound of per-act infection risk and may be dangerously misleading . Our model predicts that an HIV-1 inoculum of ~26 . 000 virions ( i . e . , a semen viral load of ~17 . 000 RNA copies per mL , assuming 3 mL semen volume ) would result in a female per-act infection risk of 0 . 1% . This semen viral load lies within the range typically observed during chronic HIV-1 infection . In contrast , several studies reported per-act penile-vaginal female infection risks between 0 . 5 and 10% , presumably linked to acute or late-stage HIV-1 infection of the male partner and correspondingly higher semen viral loads ( [32 , 53 , 54] and references therein ) . Indeed , our model predicts HIV-1 inoculums of 130 . 000 to 2 . 770 . 000 virions ( semen viral loads of 87 . 000 to 1 . 850 . 000 HIV-1 RNA copies / mL ) to result in per-act infection risks of 0 . 5 to 10% . These values represent the range of semen viral loads typically observed during acute HIV-1 infection and rare cases of exceedingly high semen viral load . Thus , our model appears to precisely capture and recapitulate the interplay between HIV-1 inoculum size and infection probability during HIV-1 transmission in the female genital tract suggested by empirical studies . Having thus obtained good indications that our model and the decisive parameter of ψ reliably mirror empirical data of human mucosal HIV-1 transmission , we next aimed to model the protective effects of nAbs present in the vaginal mucosa . First , we assumed an HIV-1 inoculum of 100 . 000 virions and modelled neutralization by four nAbs with a broad range of KD ( Fig 6B ) . Not surprisingly , this showed that with increasing nAb KD ( decreased Env trimer binding affinity ) , higher mucosal nAb concentrations are required to provide protection from infection . Specifically , we found that nAbs with a high Env binding affinity , such as PGT121 , would afford protection from infection well below mucosal nAb concentrations of 0 . 1 μg/mL . This suggests that nAbs with high Env binding affinity could exert protective effects in vivo at relatively low doses . We next looked more closely at the protective effects of a nAb with intermediate Env affinity and reduced potency , reasoning that such nAbs may be more readily inducible by vaccination . We chose nAb b12 as an example and modelled its protective effects in the vaginal mucosa against various HIV-1 inoculum sizes ( Fig 6C ) . This analysis showed that such medium-potent nAbs may provide considerable protective effects starting in the range of 1 μg/mL mucosal nAb concentration . In summary , the analyses shown in Fig 6 suggest that our model has the potential to retrieve information relevant to human mucosal HIV-1 transmission , offering opportunities for further model extension and application in HIV-1 vaccine research ( Fig 7 ) .
Broadly neutralizing antibodies are considered a crucial component of many vaccines or infectious disease therapeutics . For HIV-1 , defining the best lead antibodies has proven complex and affords intensive in vitro and in vivo efficacy testing in both animal models and humans . Thus , in silico modelling approaches that support such trials would be of enormous value; however , establishment of such models depends on detailed mechanistic knowledge of HIV-1 transmission and nAb neutralization processes . Here , we provide a step towards this by presenting an experimental and mathematical analysis of HIV-1 nAb neutralization spanning from the molecular to the organismal level , providing highly relevant quantitative insights into the initial steps of mucosal HIV-1 infection and its inhibition by nAbs . We first assessed the molecular interplay between nAbs and the HIV-1 Env trimer and confirmed that nAb occupancy of one Env subunit is sufficient for trimer neutralization ( N = 1 ) . It is important to note that N refers to the number of subunits within one trimer required to be nAb-bound for loss of functionality . This definition does not exclude the possibility that one nAb binds two adjacent trimers on the virion surface , which would result in lower nAb concentrations required for virion population neutralization . N = 1 proved true for all nAbs tested here , irrespective of epitope specificity or nAb breadth and potency , including weakly neutralizing polycloncal IgG in HIV-1+ patient plasma ( Figs 2 and 3 ) . This indicates that the intricate machinery of the HIV-1 Env trimer , required to mediate binding to and fusion with the target cell membrane , is dependent on full functionality of all three trimer subunits . By confirming that HIV-1 neutralization follows an N = 1 stoichiometry , we defined a decisive numerical requirement for HIV-1 virion neutralization by nAbs . Indeed , in combination with the mean number of trimers per virion ( η¯ ) and the number of trimers required for entry ( T ) , N defines the threshold number of antibodies required to neutralize a single HIV-1 virion or entire virion populations [27] . Alongside a range of additional parameters that were only recently determined ( Fig 1 ) , the estimation of N enabled us to use a mathematical model to analyze the interplay between nAbs , HIV-1 virion populations and the animal or human host during HIV-1 infection . We demonstrate the power of this model through several analyses . First , we predicted IC50 values of various nAbs and found that the predicted values closely matched empirical IC50 values ( Fig 4 ) . Next , we analyzed published data of macaque passive antibody immunization , vaginal virus challenge studies ( Fig 5 ) . We utilized the data presented in these studies to recapitulate virus inoculum neutralization by nAbs in vivo and to estimate the probability that a single infectious virion starts a productive host infection ( ψ^=1 . 65×10−5 ) . In a final set of analyses , we applied the estimate of ψ to model HIV-1 infection in the female genital tract and its neutralization by nAbs ( Fig 6 ) . We found that the per-act likelihood of female HIV-1 infection is clearly driven by the size of the virus inoculum , and we retrieved per-act virus transmission probabilities in good agreement with empirical estimates [32 , 33 , 51 , 53] . Furthermore , we provide estimates for mucosal nAb concentrations required to provide protection from infection , indicating that nAb concentrations in the low μg/ml range may provide protection from mucosal HIV-1 transmission . Similar concentrations of HIV-1-specific IgG are readily detectable in the vaginal mucosa of women with chronic HIV-1 infection , raising the possibility that such vaginal IgG concentrations may be achievable by vaccination [55] . The analyses of female infection risk shown in Fig 6 represent the synthesis of all previous analyses ( Figs 2 to 5 ) , incorporating all modelling parameters ( Figs 1C and 7 ) . It will remain difficult if not impossible to precisely determine per-act HIV-1 inoculum sizes and infection outcomes in a human study population as well as the effect of nAbs thereon . The data we obtained here suggests that our mathematical framework has the potential to retrieve some of this much needed information by in silico modelling of in vivo HIV-1 infection and nAb neutralization . Given this potential relevance , in the following we discuss these estimates and the underlying model assumptions in more detail . First , our model framework was built on data from macaque vaginal challenge studies . While host differences certainly exist , basic principles of mucosal HIV transmission in macaques and humans are similar [56 , 57] . Penetration of mucosal barriers has been shown in both cases to be a rapid but inefficient process resulting in focal infection of few mucosal CD4+ cells , with productive host infection frequently ensuing from a single transmitted-founder virus [50 , 58–60] . The probabilities we derive here for virion infectivity in macaque vaginal mucosal transmission should thus provide valuable insight into human infection processes . Secondly , we would like to point out the importance of patient-to-patient variation , especially for the relation between HIV-1 inoculum size and host infection probability ( Fig 6A ) . Of note , the per-patient and per-act HIV-1 inoculum size may differ widely based on variation in semen viral load and semen volume , which may both range over several orders of magnitude [33 , 51 , 52 , 61 , 62] . Our analyses support the hypothesis that HIV-1 transmission probability from an infected male partner with semen viral loads as typically observed in the chronic stage of HIV-1 infection is relatively low ( ≤ 0 . 1% ) , and that the HIV-1 pandemic may be primarily driven by transmissions occurring through high semen viral loads during acute or late-stage infections [33 , 63] . Third , we focused our analysis on male to female HIV-1 infection , as it represents the most frequent pathway of human HIV-1 transmission [64] . However , our model is not per se limited to the setting of penile-vaginal transmission and can be adapted to capture rectal , mother-to-child or intravenous HIV-1 transmission . For example , many recent non-human primate studies tested passive nAb immunization followed by rectal virus challenge [11–13 , 16 , 18 , 19] . The data presented in these studies could be leveraged by our model framework , as demonstrated here for vaginal challenge , to estimate ψ for rectal infection and subsequently test hypotheses for rectal HIV-1 transmission risk . Fourth , mucosal nAb levels are challenging to measure and were not available in all macaque challenge experiments analysed here . Thus , we extrapolated mucosal nAb concentrations based on blood plasma nAb levels , using the ratio between plasma and mucosal nAb concentrations from studies where both parameters were experimentally determined . While this provided valuable estimates , precise measurement of mucosal nAb levels would be ideal for future studies building on our model framework . Overall , we would like to note that our model should be viewed as a starting point to further investigate in vivo HIV-1 infection and nAb neutralization processes , as it focuses solely on virus-antibody interactions leaving additional factors , such as the mucosal milieu and nAb effector functions , not accounted for . The model can and should be fine-tuned by incorporation of additional parameters once they become known ( Fig 7 ) . Mucosal HIV-1 infection is incompletely understood and bottlenecks in transmission that may specifically select viral variants have not been specified . It remains debated whether HIV-1 transmitted-founder strains show distinct properties or whether transmission is purely stochastic [65–68]; our approach may help to shed light on this important question . Additionally , a range of factors are considered to influence mucosal HIV-1 transmission including epithelial micro-trauma , local inflammation , presence of other sexually transmitted infections , mucosal target cell availability , and innate immune defences [24 , 69–71] . Reliably parametrizing these conditions will remain challenging but could add valuable information in forthcoming studies utilizing our model framework . Furthermore , our model currently does not include selected nAb features that may impact on neutralization efficacy , including the effect of neutralization plateaus [13 , 49 , 72] , the contribution of Fc-mediated mechanisms [73–75] , the effect of non-neutralizing Abs [76–79] and the role of IgA in HIV-1 inhibition [80–82] . Most importantly with respect to neutralization efficacy , information on nAb half-life and tissue distribution is only starting to emerge [83 , 84] . In combination , these factors likely contribute substantially to inter-patient variation in susceptibility to HIV-1 infection , and it will be highly interesting to incorporate them in future model extensions . By estimating the probability that an infectious HIV-1 virion establishes an infection ( ψ ) , and by being able to predict the effect of mucosal nAbs on HIV-1 infection risk , our study occupies a sweet spot between HIV-1 mathematical epidemiology and virus dynamics studies . Modelling studies of the epidemiological spread of HIV-1 [85–87] , on the one hand , are typically not accounting for the transmitted virus dose , but rather assume a fixed probability of transmission between donor and recipient upon encounter . While the probability of transmission is often stratified by cofactors ( such as sex or age of donor or recipient , or set-point viral load in the donor ) it lacks the detailed mechanistic underpinning that our approach provides . Virus dynamics studies of HIV-1 , on the other hand , are mostly concerned with the virus dynamics once the infection has been established , often focusing on changes brought about by treatment [88 , 89] . In these studies , the anatomy of the host is usually not considered in detail . The necessity of integrating the within-host dynamics into the epidemiological modelling of any pathogen has been theoretically conceived [90 , 91] . In the context of HIV-1 , however , such so-called nested or embedded approaches have so far been used only in theoretical studies on the evolutionary dynamics of HIV-1 [92–94] . In a few studies , the probability of establishment of an infection along with its potential modulators , such as microbicides , T cells , exposure history , or latency , has been investigated for HIV-1 and SIV [95–101] as well as for other pathogens [102] . However , these studies did not provide the bottom-up empirical link between the establishment of an infection and mucosal antibody levels that is central to our approach . A notable exception is the study by McKinley et al . [103] who presented a model for the early virus dynamics and the effect of antibodies . In contrast to our model , however , they predict neutralization success purely on the basis of the binding kinetics ( KD ) of antibodies to the HIV-1 Env trimer . In summary , our study thus provides a comprehensive set of essential and empirically-derived parameters for modelling efforts that aim to combine the within and between host dynamics of HIV-1 infection . In conclusion , our combined experimental-mathematical approach delivers precise estimates of virion-antibody interaction stoichiometry , single-virion mucosal transmission probability , male to female per-act infection risk and in vivo nAb neutralization efficacy . These data represent novel quantitative insight into both the molecular details of HIV-1 antibody neutralization and the systemic level of mucosal HIV-1 infection . Our findings suggest that the model framework introduced here incorporates essential parameters that capture decisive steps of early HIV-1 infection and nAb neutralization , and thus provides means to predict and analyse the effects of nAbs on blocking mucosal virus transmission in vivo . Furthermore , our framework offers vast options for model extensions to investigate additional parameters or entirely different infection scenarios ( Fig 7 ) . Thus , our work represents a versatile , generalizable modelling tool to enhance our fundamental mechanistic knowledge of virus-antibody interactions and viral mucosal transmission , and can serve as stepping stone for planning and post-hoc evaluation of HIV-1 antibody-based treatment and vaccine trials .
Plasma samples from chronically infected individuals ( ZA110 , Pat117 , Pat118 , Pat122 ) were obtained from biobank samples previously collected during two approved clinical trials , the Swiss treatment interruption trial [104–108] and the Zurich Primary HIV-infection ( ZPHI ) study ( ClinicalTrials . gov identifier NCT00537966 ) [109] . Written informed consent was obtained from all individuals in the respective studies according to the guidelines of the ethics committee of the canton Zurich . 293-T cells ( obtained from the American Type Culture Collection ) and TZM-bl cells [110] ( obtained from the NIH AIDS Reagent Program ) were cultured as described [111] . The origin of HIV Env plasmids is listed in S3 Table . Env point mutations were generated by site-directed mutagenesis ( Agilent QuikChange II XL kit ) . All Env mutants were confirmed by sequencing . V1V2-deleted Envs were previously described [44] . The Luciferase reporter HIV-1 pseudotyping vector pNLLuc-AM was previously described [111] . MAbs ( see S4 Table ) were kindly provided by: PG9 , PGT121 , PGT128 , PGT135 , PGT145 , b12 and b6 by Dr . Dennis Burton , The Scripps Research Institute , La Jolla , USA . 2F5 and 2G12 by Dr . Dietmar Katinger , Polymun Scientific , Vienna , Austria . 17b and 48D by Dr . James Robinson , Tulane University , New Orleans , USA . 447-52D was purchased from Polymun Scientific . Expression plasmids for VRC01 and PGV04 were provided by Dr . John Mascola , National Institutes of Allergy and Infectious Diseases , Bethesda , USA . Expression plasmids for NIH45 . 46 and 1 . 79 were provided by Dr . Michel Nussenzweig , The Rockefeller University , New York , USA . T-20 was purchased from Roche Pharmaceuticals . To produce HIV-1 pseudovirus stocks expressing mixed trimers with varying ratios of nAb-sensitive to nAb-resistant Env , 293-T cells in 6-well plates ( 250 . 000 cells per well in 2 ml complete DMEM , seeded 24 h pre-transfection ) were transfected with 3 μg pNLLuc-AM and 1 μg Env expression plasmids , using polyethyleneimine ( PEI ) as transfection reagent . The ratio of sensitive to resistant Env was varied to yield combinations with 100 , 90 , 70 , 50 , 30 , 10 and 0% of resistant Env . After overnight incubation the transfection medium was replaced with 2 . 5 ml fresh complete DMEM and virus-containing supernatants were harvested 48 h post transfection . To determine virus infectivity , serial dilutions of virus stocks were added to TZM-bl cells in 96-well plates ( 10 . 000 cells per well ) in DMEM supplemented with 10 μg/ml DEAE-Dextran . TZM-bl infection was quantified 48 h post-infection by measuring activity of the firefly luciferase reporter ( in arbitrary relative light units , RLU ) . The neutralization activity of mAbs and patient plasma against the mixed trimer virus stocks was evaluated on TZM-bl cells as described [111] . Sufficiently high starting concentrations of inhibitors were chosen to yield clear neutralization plateaus , and data were fitted in GraphPad Prism version 7 . 0 using the sigmoidal dose-response ( variable slope ) function . In cases where no clear neutralization plateaus were obtained ( less than two consecutive data points giving the same level of neutralization ) , the position of the expected plateau was provided by the curve fit . Subsequently , the relative infectivity ( RI ) of each virus stock under saturating inhibitor concentrations was calculated . The resulting RI values were plotted over the fraction of resistant Env ( fR ) of each virus stock and the data were analyzed with mathematical models . 293-T cells were transfected with Env and Rev plasmids and processed for flow cytometry as described [37] . Env on the cell surface was detected with biotinylated mAb 2G12 ( 5 μg/ml ) and Streptavidin-APC ( BioLegend , San Diego , USA; 1:400 dilution ) or Abs 1 . 79 and PG9 ( 5 μg/ml ) and Cy5-conjugated F ( ab′ ) 2 goat anti–human IgG ( Jackson ImmunoResearch , West Grove , USA; 1:500 dilution ) followed by cell analysis on a CyAN ADP flow cytometer ( Beckman Coulter , Brea , USA ) . To tackle the question how high nAb concentrations must be to neutralize a given virion population , we first study the number of nAbs required to perform this task . We start with a virus population with nv virions . Each virion has a random trimer number Si that follows a discretized Beta distribution with mean η¯ and variance v=49/14×η¯ ( see above ) . We let nAbs bind to these virions until all virions are neutralized . Neutralization of virion i is achieved when at least ( Si − T + 1 ) trimers are bound to at least N nAbs . This simulation is repeated nr = 1000 times and the mean of the nAb numbers to reach neutralization is calculated . We introduced this procedure in Magnus and Regoes , 2011 [27] . To transition from nAb numbers required for virion neutralization to nAb concentrations , we model the binding of a nAb , Ab , to an envelope protein , E , with a chemical binding equation [35]: E+Ab⇌kdkaEAb where kd is the off-rate constant and ka the on-rate constant . Assuming a first order reaction , the quotient of the product of the reactant concentrations divided by the product concentration follows: KD≔kdka= c ( E ) c ( Ab ) c ( EAb ) The fraction of bound envelope proteins , fb , when the equilibrium is reached can then be calculated by fb= c ( EAb ) c ( E ) +c ( EAb ) = ( c ( E ) c ( EAb ) +1 ) −1= ( KDc ( Ab ) +1 ) −1 This equation can be transformed to c ( Ab ) = KDfb1−fb With this equation it is possible to determine the nAb concentrations needed for sterilizing neutralization by determining the fraction of bound envelope proteins with the simulation tool described above . We start with calculating the percentage of neutralized virions when nAb nAbs are bound to nv virions . Each virion has a discretized Beta distributed trimer number , Si , with mean η¯ and variance v=49/14×η¯ . We then distribute the nAb nAbs to the virions such that each envelope protein has the same probability of being bound . Thus Ai nAbs bind to virion i . The probability that a virion with s trimers is neutralized when a nAbs bind to the complete virion can be calculated as follows: P ( neut|S=s , A=a ) =0 if a< ( s−T+1 ) N; P ( neut|S=s , A=a ) =1 if a≥3 ( s−T ) + ( N−1 ) T+1 and P ( neut|S=s , A=a ) =∑ ( y1 , y2 , y3 ) ∈εa , sP ( Y1=y1 , Y2=y2 , Y3=y3 ) where P ( Y1 = y1 , Y2 = y2 , Y3 = y3 ) is the probability that yj trimers are bound to j nAbs and ℇa , s is the set of all combinations of a nAbs to s trimers such that at least ( s − T + 1 ) trimers are bound to at least N nAbs . For N = 1 this set is εa , s= ( ( y1 , y2 , y3 ) ∈ℕ03|y1=m , y2=3ζ−a−2m , y3=a+m−2ζ with 0≤m≤min ( s , a ) and s−T+1≤ζ≤s ) The fraction of neutralized virions , fnv , can thus be calculated by: fnv=1nv∑i=1nvP ( neut|Si=si , Ai=ai ) To calculate the mean fraction of neutralized virions , the above described procedure is performed nr = 1000 times . Several nAbs have been shown to bind with only one antibody per trimer , including nAbs PG9 , PG16 and PGT145 [38 , 47 , 49] . To account for this binding behavior , we extended the above described model . We still assume that each trimer has three epitopes . However , as soon as one nAb binds to a trimer , no additional nAbs can bind . The fraction of bound epitopes is then fb , epitopes = 1/3fb , trimers and a virion with s trimers is neutralized when ( s − T + 1 ) trimers are bound by one nAb . | Successful solicitation of the potential of neutralizing antibodies for HIV-1 prevention will require a deepened understanding of HIV-1 transmission and antibody neutralization . In this study , we experimentally determined molecular parameters of the HIV-1-antibody interaction , and subsequently used this knowledge to devise a mathematical model of HIV-1 infection and antibody neutralization in vivo . First , we experimentally confirmed that binding of one antibody per HIV-1 envelope trimer is sufficient for trimer neutralization . This finding , in combination with the number of trimers per HIV-1 virion , the number of trimers required for virus entry , and the affinity of antibody-trimer binding , enabled precise modelling of HIV-1 antibody neutralization . We employed our model for a post-hoc analysis of non-human primate infection studies , thereby obtaining estimates of HIV-1 neutralization in vivo and the probability for a single HIV-1 virion to initiate host infection . We further modelled HIV-1 infection and antibody neutralization during male-to-female transmission in the human host , which delivered estimates for the likelihood of HIV-1 transmission per sexual act and predictions of protective mucosal antibody concentrations . The quantitative insights into HIV-1 infection and antibody neutralization derived here , spanning from the molecular to the systemic level , contribute to a refined understanding of HIV-1 transmission and may prove useful for pre-study planning or post-hoc analyses of HIV-1 clinical trials and vaccine studies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"viral",
"transmission",
"and",
"infection",
"pathogens",
"immunology",
"microbiology",
"vertebrates",
"viral",
"structure",
"animals",
"mamm... | 2017 | Predicting HIV-1 transmission and antibody neutralization efficacy in vivo from stoichiometric parameters |
Interaction between the hepatitis C virus ( HCV ) envelope protein E2 and the host receptor CD81 is essential for HCV entry into target cells . The number of E2-CD81 complexes necessary for HCV entry has remained difficult to estimate experimentally . Using the recently developed cell culture systems that allow persistent HCV infection in vitro , the dependence of HCV entry and kinetics on CD81 expression has been measured . We reasoned that analysis of the latter experiments using a mathematical model of viral kinetics may yield estimates of the number of E2-CD81 complexes necessary for HCV entry . Here , we constructed a mathematical model of HCV viral kinetics in vitro , in which we accounted explicitly for the dependence of HCV entry on CD81 expression . Model predictions of viral kinetics are in quantitative agreement with experimental observations . Specifically , our model predicts triphasic viral kinetics in vitro , where the first phase is characterized by cell proliferation , the second by the infection of susceptible cells and the third by the growth of cells refractory to infection . By fitting model predictions to the above data , we were able to estimate the threshold number of E2-CD81 complexes necessary for HCV entry into human hepatoma-derived cells . We found that depending on the E2-CD81 binding affinity , between 1 and 13 E2-CD81 complexes are necessary for HCV entry . With this estimate , our model captured data from independent experiments that employed different HCV clones and cells with distinct CD81 expression levels , indicating that the estimate is robust . Our study thus quantifies the molecular requirements of HCV entry and suggests guidelines for intervention strategies that target the E2-CD81 interaction . Further , our model presents a framework for quantitative analyses of cell culture studies now extensively employed to investigate HCV infection .
HCV entry into target cells is a complex process involving the interactions of the viral envelope proteins E1 and E2 and several cell surface receptors , namely , scavenger receptor class B type I ( SR-BI ) [1] , the tetraspanin CD81 [2] , [3] , and the tight junction proteins claudin-1 ( CLDN1 ) [4] and occludin [5] . Several recent studies suggest a central role for CD81 in HCV entry: E2 has been shown to interact directly with SR-BI and CD81 following viral attachment to a target cell [1] , [2] . Patient derived neutralizing antibodies appear to target the CD81 binding domains on E2 [6] . Indeed , anti-CD81 antibodies were able to block infection in vitro [3] and in a mouse model [7] . Graft reinfection following liver transplantation was observed recently to select for HCV strains capable of more efficient entry , achieved partly through mutations in the CD81 binding domains on E2 [8] . Expression of human CD81 and occludin was essential for infection of genetically humanized mice [7] . Besides , CLDN1 appears to mediate HCV entry through its association with CD81 [9] , [10] . Consequently , the E2-CD81 interaction presents a potent target for intervention; drugs that block the E2-CD81 interaction are currently under development [11] , [12] . How many E2-CD81 complexes must be formed between a virion and a target cell to enable HCV entry ? Knowledge of this threshold would determine the number of E2-CD81 complexes that a drug or a vaccine must prevent from forming in order to block viral entry , thus presenting a quantitative guideline for intervention strategies targeting the E2-CD81 interaction . This threshold is currently unknown . Direct observation of the number of E2-CD81 complexes formed before HCV entry has not been possible . Recent cell culture studies have determined the dependence of viral entry and kinetics in vitro on the CD81 expression level on target cells [10] , [13]–[18] . In particular , cells expressing higher levels of CD81 were found to be more susceptible to infection [13] . Further , the frequency of cells with low CD81 expression typically increased with time following the exposure of cells to HCV [14] , [15] . We reasoned that analysis of these observations using a mathematical model of viral kinetics , akin to studies of HIV entry ( for example , see [19] , [20] ) , may allow estimation of the threshold number of E2-CD81 complexes necessary for HCV entry . While models of HCV viral kinetics in vivo have been employed successfully to analyse patient data and elucidate guidelines for treatment [21]–[31] , models of HCV viral kinetics in vitro have just begun to be formulated . Here , we constructed a mathematical model of HCV viral kinetics in vitro that mimics cell culture studies of the dependence of viral entry and kinetics on CD81 expression . Model predictions captured data from several independent experiments quantitatively and yielded estimates of the threshold number of E2-CD81 complexes necessary for HCV entry .
We considered in vitro experiments where a population of target cells , , with a known distribution of the CD81 expression level across cells is exposed to a population of HCVcc ( cell culture adapted ) virions , , and the progression of infection followed [13]–[15] . We modelled the ensuing viral kinetics as follows ( Fig . 1 ) . We first considered a single virus-cell pair with the virus attached to the cell by interactions that precede E2-CD81 binding [12] . E2 and CD81 then interact to form E2-CD81 complexes . We computed the mean number of these complexes formed at equilibrium , , as a function of the CD81 expression level , , on the cell . Assuming that the number of complexes formed , , followed a Poisson distribution with mean , we computed the probability , , that was larger than a threshold number . We assumed that viral entry ( and subsequently infection ) occurred if ( Fig . 1A ) . Thus , yielded the relative susceptibility to infection of a cell with CD81 expression level . We next considered the population of cells exposed to virions ( Fig . 1B ) . We divided the cells into different subpopulations with distinct CD81 expression levels and hence different susceptibilities , where . Cells in each subpopulation were assumed to proliferate , die , or be infected at a rate proportional to . The resulting infected cells , , were lost at enhanced rates compared to due to virus-induced cytopathicity in vitro [15] , [32] . Free virions were produced by infected cells and were cleared . With this description , we constructed dynamical equations to predict the time-evolution of each of the uninfected and infected cell subpopulations and the population of free virions and compared our predictions with experiments ( Methods ) .
Analysis of experimental data using mathematical models has provided crucial insights into disease pathogenesis and the effectiveness of drugs and established guidelines for rational optimization of therapy for HCV infection [21]–[31] , [33] , [34] . The recent development of cell culture systems that allow persistent HCV infection in vitro [35]–[38] has yielded a wealth of new data on HCV replication , evolution , and the impact of drugs . Analysis of this data is expected to provide further insights into HCV pathogenesis and outcomes of therapy , but has been precluded by the lack of mathematical models of HCV viral kinetics in vitro . Indeed , significant efforts are underway to construct models of the intracellular replication and evolution of HCV with the aim of elucidating the activity of direct acting antiviral drugs [39]–[41] . Here , we have constructed a mathematical model of HCV viral kinetics in vitro . Applying it to the analysis of data from several recent cell culture studies , we obtained quantitative insights into the molecular requirements of HCV entry . We estimated that depending on the binding affinity of E2 and CD81 , between 1 and 13 E2-CD81 complexes are necessary for HCVcc entry into human hepatoma-derived cells . Our estimate provides a quantitative guideline for the optimal usage of drugs and vaccines that target the E2-CD81 interaction: A potent drug or vaccine must ensure that not more than 1–13 E2-CD81 complexes are formed across a virus-cell pair in order to prevent viral entry . This guideline assumes significance as drugs that target the E2-CD81 interaction are under development [11] , [12] and may become part of future treatments involving direct acting antiviral agents that seek to overcome the limitations of the current interferon-ribavirin-based treatments [42] . Further , a recent analysis of the HCV quasispecies in six patients who underwent liver transplantation revealed that viral strains capable of more efficient entry , achieved through modulation of the CD81 dependence of viral entry , were selected following liver transplantation [8] . Blocking E2-CD81 interactions effectively , for which our estimate presents a quantitative criterion , may thus be a promising avenue to prevent graft reinfection following liver transplantation . In an earlier study , Koutsoudakis et al . [13] estimated that a cell must express >70000 CD81 molecules to allow HCV entry . This threshold was identified as follows . In independent experiments , cells with widely varying distributions of CD81 were exposed to HCV at low MOI in culture and the percentage of cells infected at day 5 was measured . This latter percentage was found to correlate well with the percentage of cells initially expressing >70000 CD81 molecules/cell . For instance , with Huh7 . 5 cells , the percentage of cells infected at day 5 was 91 . 5 and the percentage expressing >70000 CD81 molecules/cell at day 0 was 93 , and with Huh7-Lunet cells , which express fewer CD81 molecules than Huh7 . 5 cells , the percentages were 11 . 2 and 17 , respectively . Based on this correlation , Koutsoudakis et al . suggested 70000 CD81 molecules/cell as the threshold for entry . Our present analysis suggests that the underlying viral kinetics may render this estimate an upper bound . Virus-induced cytopathicity in culture would result in the loss of susceptible cells and therefore a continuous decrease in the frequency of susceptible cells with time . Consequently , the percentage of cells susceptible initially is expected to be higher than the percentage of cells susceptible–and hence even higher than the percentage infected–at day 5 post-infection . The higher percentage of cells susceptible initially would therefore imply a threshold smaller than 70000 molecules/cell . This is also evident in the experiments of Koutsoudakis et al . , where cells with lower CD81 expression ( MFI∼50 ) than 70000 molecules/cell ( MFI∼100 ) were infected ( see Fig . 3C ) . From our analysis , we identified the threshold number of E2-CD81 complexes and not CD81 molecules necessary for entry . Because stochastic events can result in the formation of the requisite number of complexes , cells expressing few CD81 molecules have small but nonzero susceptibilities to infection . Indeed , we found using parameters employed in Fig . 3D that a small percentage of cells expressing as few as ∼10000 CD81 molecules was infected by day 5 . More recently , Zhang et al . [18] have argued that a substantially smaller expression level than suggested by Koutsoudakis et al . may suffice for entry and that CD81 may also be necessary , perhaps at higher expression levels , for post-entry events . Our model does not distinguish between entry and post-entry requirements of CD81 . Our estimate of 1–13 E2-CD81 complexes defines successful infection of a cell and combines the requirements for entry and any post-entry steps [17] , [18] . Our model yielded good fits to data with in the range , which is higher than the value , , determined using recombinant E2 and soluble CD81 [43] . ( Fits with were poor ( not shown ) . ) This discrepancy may be because the binding affinity when the proteins are in solution may be different from that when the proteins are restricted to membranes [44] , recombinant E2 may not accurately mimic the true E2-CD81 interaction [45] , and/or only a fraction of the CD81 may lie outside tetraspanin-enriched microdomains and/or be associated with CLDN1 and therefore available for binding E2 [10] , [46] . The E2-CD81 binding affinity in situ remains to be determined . Our model yielded best-fit values of the threshold number of complexes , , that decreased as increased ( Fig . 4 ) . For a given CD81 expression level , the mean number of E2-CD81 complexes formed decreased as increased , lowering susceptibility ( Eqs . ( 4 ) and ( 5 ) ) . Decreasing restored this susceptibility ( Fig . 2B inset and Fig . S6 ) , thus ensuring that the resulting viral kinetics was conserved and in agreement with data . This does not imply , however , that viral strains with lower E2-CD81 affinity ( higher ) would require fewer E2-CD81 complexes for entry . On the contrary , given a value of , our model predicts that a cell would be less susceptible to entry of viral strains with higher . Our model was designed to mimic experiments that examined the influence of CD81 expression on viral entry and kinetics [13]–[15] . In these experiments , cells with high CD81 expression were preferentially infected and lost due to virus-induced cytopathicity , and cells with low CD81 expression , refractory to infection , eventually dominated the culture , suggesting that CD81 expression limited entry . Accordingly , our model assumed that other entry receptors were not limiting . Our model then predicted triphasic viral kinetics in vitro , in agreement with experiments . We note that the origins of the triphasic pattern here are distinct from the triphasic viral load decline in some patients undergoing combination therapy , the latter due to liver homeostatic mechanisms [25] . Further , the triphasic kinetics is a short-term phenomenon ( ∼2–3 weeks ) . Over longer periods , viral evolution may alter the kinetics substantially [15] , which our model does not consider . Nonetheless , our model can be readily adapted to the scenario where a receptor other than CD81 is limiting and may thus serve to quantify the requirements of that receptor for HCV entry . We recognize a few additional simplifications in our model . First , our model ignored cell-cell transmission of infection . CD81 appears to be necessary for direct cell-cell transmission [47] . If the susceptibility of a cell to cell-cell transmission depends on CD81 expression in a manner similar to its susceptibility to viral entry , which remains to be ascertained , then we can show that our model with the pseudo-steady state approximation , , implicitly accounts for cell-cell transmission: the net infection rate from both modes , , is in agreement with our model ( Eq . ( 1 ) ) with an effective infection rate constant that lumps the rate constants of infection by free virions , , and cell-cell transmission , . Second , our model assumes that reaction equilibrium is attained rapidly compared to viral entry and that the diffusion of CD81 on the target cells continually replenishes the free CD81 in the virus-cell contact region lost due to binding . Accordingly , our model predicts an upper bound on the mean number of E2-CD81 complexes formed in the contact region . Thus , if CD81 diffusion or its binding with E2 were rate limiting , a threshold smaller than 1–13 E2-CD81 complexes is expected to describe the data we considered . Finally , we ignored the splitting of cell culture at confluence . We employed the data at day 3 post-infection from Koutsoudakis et al . [13] , when cells are not expected to have reached confluence . Further , the best-fit parameter estimates obtained from the data of Koutsoudakis et al . [13] were close to those from Zhong et al . [15] ( Fig . 4 ) . Also , accounting for splitting did not significantly alter our comparisons with the data of Tscherne et al . [14] so long as the splitting was performed after day 5 post-infection ( Fig . S7 ) . Also , including a logistic term to limit the proliferation of cells as they approached confluence did not alter our parameter estimates significantly ( not shown ) . Nonetheless , that model predictions described several independent experimental observations quantitatively indicated that even with the above simplifications our model captured the essential features of HCV viral kinetics in vitro successfully . At the same time , the simplifications restricted model parameters to a number that allowed robust parameter estimation through fits to available data .
We considered in vitro experiments where a population of uninfected cells , , is exposed to a population of HCVcc virions , . We divided the cells into subpopulations , denoted , where , with cells in each subpopulation expressing CD81 in a range around molecules per unit area . At the start of infection ( ) , the variation of with was determined from a known distribution , , of CD81 expression levels across cells ( Fig . 1 ) . The following equations described the ensuing viral kinetics ( ) : ( 1 ) ( 2 ) ( 3 ) Here , and are the proliferation and death rates of . is the infection rate of cells expressing excess CD81 . is the death rate of . Following observations of HCV-induced cell cycle arrest in vitro [32] , [48] , we neglected the proliferation of . and are the per cell production rate and the clearance rate of free virions , respectively . Here , represents the combined rate of the natural degradation of virions , the loss of viral infectivity , and the loss of virions due to entry and attachment [49] , [50] . For simplicity , we assumed to be a constant . To determine , we considered a cell , with CD81 expression level , closely apposed to a virion with E2 molecules per unit area . We assumed , as with HIV [20] , that the E2-CD81 interactions across the virus-cell interface attain equilibrium well before viral entry . If is the surface density of E2-CD81 complexes , that of unbound CD81 and that of unbound E2 molecules in the contact area , then at equilibrium , where is the equilibrium dissociation constant of E2-CD81 complexes when the proteins are restricted to membranes . From Bell's analysis , , where is the equilibrium dissociation constant when the proteins are in solution and is the encounter distance between the proteins for bond formation [51] . The virus-cell contact area , , is small compared to the surface area of the cell . Further , free CD81 can diffuse on the cell membrane and therefore be recruited to the contact area . Consequently , is expected not to decrease substantially below , as suggested also by an independent reaction-diffusion model [52] . In contrast , the viral surface area is comparable to and assuming E2 to be less mobile than CD81 , it follows that the surface density of E2 in the contact area obeys the species balance equation: . Under the latter two constraints , the mean number of complexes formed across the virus-cell contact at equilibrium , , is given by ( 4 ) We recognized next that the E2 expression level on the virion and the virus-cell contact area are subject to stochastic variations . We assumed therefore that the number of complexes formed during a virus-cell contact , , follows a Poisson distribution with mean . Viral entry ( and subsequently infection ) occurred if , where is the threshold number of E2-CD81 complexes necessary for HCV entry . The probability that , ( 5 ) thus yielded the relative susceptibility to infection of a cell with CD81 expression level . ( The Poisson distribution does allow to exceed the limit of 180 set by the number of E2 molecules present on a virion , but in all our calculations remained fewer than 30 ( Fig . S8 ) so that the probability that was negligibly small . ) Equations ( 1 ) – ( 5 ) yielded a model of HCV kinetics in vitro that accounted explicitly for the dependence of viral entry on the CD81 expression level on cells . We solved model equations using a computer program written in MATLAB and computed quantities measured experimentally , namely , the time-evolution of the populations of uninfected and infected cells , and , the viral titre , , the fraction of cells infected , , the fraction of cells infected within each subpopulation , , and the populations of viable and dead cells , and , respectively . We employed the following parameter values and initial conditions: corresponding to 180 E2 molecules on a virion with average diameter 50 nm [53]; , corresponding to a virus-cell contact radius of [54]; the target cell diameter was [17]; was varied over the range ( see above ) . The initial CD81 expression was assumed to follow the log-normal distribution , , where and were the mean and standard deviation of . For comparisons with experimental data , the initial distributions were obtained from measurements ( Fig . S2; see below ) . Thus , the initial cell subpopulations , , where was the total initial target cell population . We divided the range of CD81 expression levels into intervals , which determined ; finer discretisation did not improve the accuracy of our solution ( Fig . S1 ) . The remaining parameters , estimated from fits to data , are listed in Tables S1 and S2 . We have summarized model parameters in Table 1 . We considered data from three recently published cell culture studies of HCV kinetics . First , we considered data of Zhong et al . [15] , where Huh-7 . 5 . 1 cells were exposed to JFH-1 virions and the kinetics of infection followed for 21 d . Specifically , we employed the data of the time-evolution of the supernatant infectivity , the population of attached ( viable ) cells , and the ratio of the populations of floating ( dead ) and attached cells , the latter two datasets with mock infection as well as with HCVcc infection ( Fig . S1 in [15] ) . Second , we considered data of Koutsoudakis et al . [13] , where a 1∶1 mixture of Huh7-Lunet and Lunet/CD81 cells was exposed to HCVcc Venus-Jc1 virus and the fractions of cells infected in subpopulations with distinct CD81 expression levels were measured after 72 h . The initial viral population was at an MOI of ∼5 and ∼1 TCID50/cell ( where TCID50 is the 50% tissue culture infective dose ) , respectively , in two independent experiments ( Fig . 5 in [13] ) . Third , we considered the data of Tscherne et al . [14] , where Huh7 . 5 cells were exposed to J6/JFH virus and the time-evolution of the fraction of cells infected as well as of the distribution of CD81 expression levels across cells was followed for 17 d ( Fig . 8 in [14] ) . The CD81 expression on cells is usually measured in terms of fluorescent intensity . To convert the measurements to CD81 surface densities , we adopted the following procedure . Measured distributions of the CD81 expression level on Huh-7 . 5 ( silRR ) cells were digitized from Zhang et al . [55] and on Huh-7 . 5 cells from Koutsoudakis et al . [13] . The log-normal distribution , , yielded good fits to the data ( Fig . S9 ) . The best-fit parameter values ( 95% CI ) were = 4 . 29 ( 4 . 27–4 . 31 ) and = 0 . 45 ( 0 . 43–0 . 47 ) for the data of Zhang et al . and = 6 . 1 ( 6 . 07–6 . 12 ) and = 0 . 7 ( 0 . 67–0 . 72 ) for the data of Koutsoudakis et al . Because the same cell lines were used , the underlying distributions of CD81 expression are expected to be similar in the two experiments . The differences in the reported fluorescence intensities may therefore be attributed to different conversions of CD81 surface density to fluorescence intensity in the two measurements . To unravel the underlying distribution , we assumed that log fluorescence intensity was linearly proportional to log CD81 surface density ( or that the fluorescence intensity had a power law dependence on the surface density ) . Thus , a cell with CD81 surface density would yield a fluorescence intensity and in the measurements of Zhang et al . and Koutsoudakis et al . , respectively , such that and , where a , b , c , and d are constants . To determine the latter constants , we employed the following observations: The mean surface density of CD81 for the Huh-7 cell line is 2×105 molecules/cell [3] , which corresponded to a fluorescence intensity of 48 units in the measurements of Zhang et al . , so that . Also , Koutsoudakis et al . measured the number of CD81 molecules/cell corresponding to a fluorescence intensity of 100 units and found it to be 7×104 molecules/cell , so that . Further the above relationship between fluorescence intensity and CD81 expression also implied that and . Solving the latter equations using the best-fit parameters above , we obtained a = 2 . 8 , b = 2 . 43 , c = 3 . 96 and d = 1 . 56 , which enabled conversion of measured fluorescence intensities to CD81 surface densities , ( Fig . S2 ) . The measured distributions in terms of counts versus fluorescence intensities were then converted to probability distributions , versus , by normalizing the counts such that the areas under the versus curves equalled unity . The resulting distributions ( Fig . S2 ) were employed as the initial CD81 distributions for our fits to data ( Fig . 3 ) . We digitized data using Engauge digitizer and fit model predictions to data using the nonlinear regression tool NLINFIT in MATLAB . | The interaction between the hepatitis C virus ( HCV ) envelope protein E2 and the host cell surface receptor CD81 is critical for HCV entry into hepatocytes and presents a promising drug and vaccine target . Yet , the number of E2-CD81 complexes that must be formed between a virus and a target cell to enable viral entry remains unknown . Direct observation of the E2-CD81 complexes preceding viral entry has not been possible . We constructed a mathematical model of HCV viral kinetics in vitro and using it to analyze data from recent cell culture studies obtained estimates of the threshold number of E2-CD81 complexes necessary for HCV entry . We found that depending on the E2-CD81 binding affinity , between 1 and 13 complexes are necessary for HCV entry into human hepatoma-derived cells . Our study thus presents new , quantitative insights into the molecular requirements of HCV entry , which may serve as a guideline for intervention strategies targeting the E2-CD81 interaction . Further , our study shows that HCV viral kinetics in vitro can be described using a mathematical model , thus facilitating quantitative analyses of the wealth of data now emanating from cell culture studies of HCV infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"medicine",
"infectious",
"diseases",
"theoretical",
"biology",
"gastroenterology",
"and",
"hepatology",
"biology",
"computational",
"biology"
] | 2011 | Mathematical Model of Viral Kinetics In Vitro Estimates the Number of E2-CD81 Complexes Necessary for Hepatitis C Virus Entry |
Asymmetric cell division and apoptosis ( programmed cell death ) are two fundamental processes that are important for the development and function of multicellular organisms . We have found that the processes of asymmetric cell division and apoptosis can be functionally linked . Specifically , we show that asymmetric cell division in the nematode Caenorhabditis elegans is mediated by a pathway involving three genes , dnj-11 MIDA1 , ces-2 HLF , and ces-1 Snail , that directly control the enzymatic machinery responsible for apoptosis . Interestingly , the MIDA1-like protein GlsA of the alga Volvox carteri , as well as the Snail-related proteins Snail , Escargot , and Worniu of Drosophila melanogaster , have previously been implicated in asymmetric cell division . Therefore , C . elegans dnj-11 MIDA1 , ces-2 HLF , and ces-1 Snail may be components of a pathway involved in asymmetric cell division that is conserved throughout the plant and animal kingdoms . Furthermore , based on our results , we propose that this pathway directly controls the apoptotic fate in C . elegans , and possibly other animals as well .
Asymmetric cell division is essential for the generation of cellular diversity during animal development [1] . In certain cases , one of the cells derived from an asymmetric division is specified to undergo apoptosis ( programmed cell death ) [2–11] . However , the genetic and cell biological mechanisms that permit the coupling of asymmetric cell division and the adoption of the apoptotic fate are not well understood . Previous studies have implicated members of the Snail family of transcriptional repressors in both asymmetric cell division and apoptosis , but there has been no demonstration that these processes are integrated in a given developmental context . In mammals , Snail-related proteins have been shown to have pro-survival as well as anti-apoptotic activities , and they have been causally linked to tumorigenesis and tumor progression in mammals [12 , 13] . In Drosophila melanogaster , the Snail-related genes snail , escargot , and worniu have been shown to function redundantly in asymmetric cell division [14 , 15] . Specifically , rather than dividing asymmetrically and producing two daughter cells of different sizes and fates , neuroblasts in the central nervous system of D . melanogaster mutants lacking snail , escargot , and worniu function divide symmetrically to produce two daughter cells of similar sizes and fates . The effect of snail , escargot , and worniu on asymmetric neuroblast division is mediated in part by their ability to promote the expression of the gene inscuteable , which encodes an adaptor protein that is required for the establishment and maintenance of neuroblast polarity [16] . The protein Inscuteable is localized in a polar fashion through its interaction with a complex composed of the PDZ domain–containing proteins Bazooka ( also referred to as Par-3 ) and Par-6 , which are found on the apical cell cortex of the neuroblasts [17] . Inscuteable in turn recruits the adaptor protein Pins and the alpha subunit of the heterotrimeric G protein Gi ( Gαi ) , which initiates the displacement of the mitotic spindle along the cell division axis of the neuroblasts , resulting in their asymmetric division . Inscuteable is also at least partially required for the enrichment of cell-fate determinants such as Prospero and Staufen on the basal cell cortex of the neuroblasts and their asymmetric segregation into the basal daughter cell [17] . In Caenorhabditis elegans , the Snail-related protein CES-1 ( cell-death specification ) has been implicated in the suppression of a specific apoptotic event , the death of the neurosecretory motoneuron ( NSM ) sister cell [18 , 19] . During embryonic development , the NSM neuroblast divides to give rise to two daughter cells: the NSM , which differentiates into a serotonergic motoneuron , and the NSM sister cell , which undergoes apoptosis [3] . A gain-of-function mutation of the ces-1 gene , which most likely results in the mis- or overexpression of ces-1 in the NSM lineage , prevents the death of the NSM sister cell [18] . The death of the NSM sister cell is dependent on the transcriptional upregulation of the pro-apoptotic BH3-only gene egl-1 ( egg-laying abnormal ) in the NSM sister cell , a process that is at least partially dependent on a heterodimer composed of the bHLH transcription factors HLH-2 and HLH-3 ( HLH-2/HLH-3 ) [20] ( Figure 1 ) . HLH-2/HLH-3 can bind to E-boxes/Snail-binding sites in a cis-regulatory region of the egl-1 locus referred to as Region B , which is required for the expression of egl-1 in the NSM sister cell in vivo . Therefore , it has been proposed that HLH-2/HLH-3 is a direct activator of egl-1 transcription in the NSM sister cell . The ces-1 ( gf ) mutation prevents the death of the NSM sister cell by blocking the HLH-2/HLH-3–dependent expression of egl-1 in the NSM sister cell [18 , 20] . Furthermore , CES-1 can also bind to the E-boxes/Snail-binding sites in Region B of the egl-1 locus , and the ability of elevated levels of CES-1 to prevent the death of the NSM sister cell in vivo is dependent on these E-boxes/Snail-binding sites . On the basis of these observations , it has been proposed that by competing with HLH-2/HLH-3 for binding to Region B of the egl-1 locus , elevated levels of CES-1 protein in ces-1 ( gf ) animals directly block egl-1 transcription in the NSM sister cell [20] . Finally , it has been hypothesized that the CES-1 protein might normally function in the NSM to block HLH-2/HLH-3–dependent egl-1 transcription , thereby allowing the survival of the NSM . The function of the ces-1 gene in the NSM sister cell is negatively regulated by the gene ces-2 , which is required for the death of the NSM sister cell and which encodes a bZIP transcription factor similar to the human proto-oncoprotein HLF ( hepatic leukemia factor ) [18 , 21] ( Figure 1 ) . Here we report that the previously uncharacterized protein DNJ-11 ( DnaJ domain–11 ) , a MIDA1 ( mouse Id associated 1 ) –like chaperone , cooperates with the CES-2 protein to reduce ces-1 transcription in the NSM lineage , thereby excluding CES-1 protein from the NSM sister cell and allowing the death of the NSM sister cell . Furthermore , we show that the NSM neuroblast , which gives rise to the NSM and NSM sister cell , divides asymmetrically and that the genes dnj-11 , ces-2 , and ces-1 also function to cause asymmetric NSM neuroblast division . Our results reveal new developmental roles of MIDA1-like chaperones and HLF-like transcription factors . Furthermore , our results delineate a pathway involved in asymmetric cell division that directly controls the apoptotic fate .
The NSM sister cells are generated about 410 min after the first cleavage of the C . elegans zygote and undergo apoptosis at about 430 min [3] . We screened for mutations that cause the NSM sister cells to survive , and we isolated the recessive mutation bc212 ( J . Hatzold , B . Conradt , unpublished data ) . bc212 causes 12% and 50% NSM sister cell survival when raised at 25 °C or 15 °C , respectively , and hence causes a cold-sensitive NSM sister cell survival phenotype ( Table 1 , dnj-11 ( bc212 ) ) . In addition , bc212 is maternally rescued: in homozygous bc212 progeny ( dnj-11 ( bc212 ) ) of heterozygous bc212 hermaphrodites ( dnj-11 ( bc212 ) /+ ) , only 1% of the NSM sister cells survive ( Table S1 ) . To determine whether bc212 prevents the death of cells other than the NSM sister cells , we analyzed additional cells that undergo apoptosis during C . elegans development . We found that bc212 had no effect on their deaths , which demonstrates that bc212 does not block apoptosis in general ( Table S2 ) . However , we found that bc212 causes a variety of other defects such as morphological defects , lethality , slow growth , and reduced brood size ( Figure S1 and Tables S3–S5 ) . Therefore , the gene defined by bc212 is required for the death of the NSM sister cells and for additional processes that are important for C . elegans development and fertility . We cloned the gene defined by bc212 and found that it is identical to the previously uncharacterized gene dnj-11 ( F38A5 . 13 ) ( DNJ-11 accession number NP_501006 ) on linkage group ( LG ) IV ( Figure 2A ) . dnj-11 encodes a 589–amino acid ( aa ) protein that is most similar to members of the family of MIDA1-like proteins , which are found in plants and animals ( see below ) . bc212 is a C-to-T transition at position 21 of the nucleotide sequence of the dnj-11 gene and transforms codon 7 into a stop codon , which is predicted to truncate the DNJ-11 protein after aa 6 ( Figure 2A and 2B ) . tm2859 , a 641–base pair ( bp ) deletion of the dnj-11 gene , removes base pairs 402-1042 of the dnj-11 coding region . This deletion is predicted to result in a frameshift leading to the generation of a truncated protein composed of the first 116 aa of the wild-type protein and a 67-aa , C-terminal extension in a different reading frame ( Figure 2A and 2B ) . Like bc212 , tm2859 causes a cold-sensitive NSM sister cell survival phenotype and 50% NSM sister cell survival at 15 °C ( Table 1 , dnj-11 ( tm2859 ) ) . In addition , bc212 and tm2859 cause 24% and 26% lethality , respectively , at 15 °C ( Table S3 ) . Based on these results , we conclude that bc212 represents a strong loss-of-function mutation and putative null allele of the dnj-11 gene . Using a functional transgene that drives the expression of a DNJ-11::green fluorescent protein ( GFP ) fusion protein under the control of the dnj-11 promoter ( Pdnj-11dnj-11::gfp ) , we determined the expression pattern of the dnj-11 gene and the sub-cellular localization of the DNJ-11 protein . Pdnj-11dnj-11::gfp expression was observed in most if not all somatic cells of embryos , larvae , and adult animals , including the cells of the NSM lineage ( Figure 2C and unpublished data ) . DNJ-11::GFP protein primarily localized to the cytoplasm in a punctate pattern and could not be detected in nuclei . In addition , the dnj-11 gene is most likely also expressed in the adult germ line , because dnj-11 ( bc212 ) is maternally rescued ( Table S1 ) . The death of the NSM sister cells is dependent on the transcriptional upregulation in the NSM sister cells of the BH3-only gene egl-1 ( EGL-1 accession number NP_506575 ) [20] . The transcriptional upregulation of egl-1 in the NSM sister cells can be blocked by the Snail-related transcription factor CES-1 ( accession number NP_492338 ) , which in turn is negatively regulated by the HLF-like transcription factor CES-2 ( accession number NP_493610 ) [18–21] . By using an egl-1 transgene ( Pegl-1his-24::gfp ) , we found that dnj-11 ( bc212 ) resulted in the loss of egl-1 expression in 69% of the NSM sister cells , indicating that dnj-11 acts upstream of egl-1 to promote its transcription in the NSM sister cells ( Table S6 ) . To determine where dnj-11 functions with respect to the genes ces-1 and ces-2 , we analyzed the interactions of dnj-11 ( bc212 ) with a putative null mutation of ces-1—n703n1434 [19]—and a partial , temperature-sensitive loss-of-function ( lf ) mutation of ces-2—n732 [18] . We found that 0% of the NSM sister cells survived in ces-1 ( n703n1434 ) ; dnj-11 ( bc212 ) double mutants , demonstrating that the ability of dnj-11 ( bc212 ) to prevent the death of the NSM sister cells requires a functional ces-1 gene ( Table 1 , ces-1 ( n703n1434 ) ; dnj-11 ( bc212 ) ) . Finally , we found that at 20 °C , dnj-11 ( bc212 ) greatly enhanced the NSM sister cell survival phenotype caused by ces-2 ( n732 ) ( Table 1 , ces-2 ( n732 ) ; dnj-11 ( bc212 ) ) . These results indicate that , at least in the NSM lineage , dnj-11 acts upstream of and as a negative regulator of ces-1 . Furthermore , these results suggest that dnj-11 and ces-2 cooperate to antagonize ces-1 function . MIDA1-like proteins contain three major regions: ( 1 ) an N-terminal J domain , which is a protein–protein interaction domain found in members of the J protein or Hsp40 family of chaperones [22] , ( 2 ) a central M domain , which is another protein–protein interaction domain found specifically in MIDA1-like proteins [23] , and ( 3 ) two C-terminal Myb domains , which are DNA-binding domains typically found in transcription factors [24 , 25] ( Figure 1A and 1B ) . MIDA1-like proteins have mainly been implicated in growth control [23 , 26 , 27] and are thought to function by regulating transcription and translation [28–32] . BLAST searches [33] revealed that the DNJ-11 protein is highly similar to mouse MIDA1 ( 40% identical , 61% similar ) ( accession number NP_033610 ) [23] , human MPP11 ( M phase phosphoprotein ) ( 38% identical , 54% similar ) ( accession number NP_055192 ) [27] , Volvox carteri GlsA ( gonidialess ) ( 34% identical , 52% similar ) ( accession number AF_106963 GenBank ) [26] , and Saccharomyces cerevisiae Zuotin ( 37% identical , 62% similar ) ( accession number NP_011801 ) [34] ( Figure 1B ) . To determine whether the J domain is important for DNJ-11 activity , we replaced the histidine residue of the conserved tripeptide of the J domain with a glutamine residue ( H129Q ) and tested a transgene that expresses a DNJ-11 ( H129Q ) ::GFP fusion protein under the control of the dnj-11 promoter ( Pdnj-11dnj-11 ( H129Q ) ::gfp ) for its ability to rescue the NSM sister cell survival phenotype caused by dnj-11 ( bc212 ) . In contrast to Pdnj-11dnj-11::gfp , Pdnj-11dnj-11 ( H129Q ) ::gfp failed to rescue the NSM sister cell survival phenotype of dnj-11 ( bc212 ) animals ( Table 2 ) . To determine whether the Myb domains are important for DNJ-11 activity , we replaced the tryptophan residue at position 456 in the first Myb domain ( W456 ) and/or the phenylalanine residue at position 578 in the second Myb domain ( F578 ) with glycine and tested the resulting transgenes ( Pdnj-11dnj-11 ( W456G ) ::gfp , Pdnj-11dnj-11 ( F578G ) ::gfp , and Pdnj-11dnj-11 ( W456G F578G ) ::gfp ) for their ability to rescue the NSM sister cell survival phenotype caused by dnj-11 ( bc212 ) . We found that the transgenes expressing a DNJ-11::GFP fusion protein with either one of the Myb domains mutated partially rescued the NSM sister cell survival phenotype of dnj-11 ( bc212 ) animals ( Table 2 , dnj-11 ( bc212 ) ; Pdnj-11dnj-11 ( W456G ) ::gfp , dnj-11 ( bc212 ) ; Pdnj-11dnj-11 ( F578G ) ::gfp ) . However , the transgene expressing a DNJ-11::GFP fusion protein with both Myb domains mutated was no longer able to rescue ( Table 2 , dnj-11 ( bc212 ) ; Pdnj-11dnj-11 ( W456G F578G ) ::gfp ) . These results demonstrate that the J domain and at least one functional Myb domain are required for the ability of DNJ-11 to cause the death of the NSM sister cells . The V . carteri ortholog of DNJ-11 , GlsA , has been implicated in asymmetric cell division [35] . Specifically , during V . carteri development , an asymmetric cell division occurs that results in the generation of two daughter cells of different sizes and fates , namely a large reproductive cell and a small somatic cell . In glsA mutants , this cell division occurs symmetrically , resulting in two cells of equal size , both of which differentiate into somatic cells [26] . Most of the 131 cells that undergo apoptosis during C . elegans development are thought to be the result of an asymmetric cell division that gives rise to a large cell , which is programmed to survive , and a small cell , which is programmed to die [2 , 3] . Therefore , we determined whether the NSM neuroblast , which gives rise to the NSM and the NSM sister cell , divides asymmetrically with respect to size as well . To that end , we identified the NSM neuroblast based on its position in the embryo , observed its division at about 410 min , and determined the sizes of its daughter cells immediately after its division had been completed ( Figure 3A ) . Using a transgene that expresses a plasma membrane–targeted GFP fusion protein ( Ppie-1gfp::ph ( PLC1δ1 ) ) [36] , we found that in wild-type animals , the size of the NSM sister cell on average is 0 . 46 times the size of the NSM ( Figure 3B , +/+ ) . Therefore , the NSM neuroblast divides asymmetrically to give rise to a large cell ( the NSM ) and a small cell ( the NSM sister cell ) . Furthermore , we found that in dnj-11 ( bc212 ) animals , the difference in size between the NSM sister cell and the NSM is highly variable , ranging from a 2-fold difference as observed in wild-type animals to no difference ( Figure 3B , dnj-11 ( bc212 ) ) . To determine whether the range observed reflects the incomplete penetrance of the NSM sister cell survival phenotype caused by dnj-11 ( bc212 ) ( see Table 1 ) , we followed the fate of NSM sister cells after the division of the NSM neuroblasts . We found that in five out of five embryos in which the division of the NSM neuroblast had occurred asymmetrically ( NSM sister cell size on average 0 . 51 times the size of the NSM ) , the NSM sister cells died ( Figure 3B , red diamonds ) . By contrast , four out of four NSM sister cells that were similar in size to the NSMs ( NSM sister cell size on average 0 . 83 times the size of the NSM ) survived ( green diamonds ) . To rule out the possibility that the increase in NSM sister cell size observed in dnj-11 ( bc212 ) animals is a result of inappropriate NSM sister cell survival rather than a defect in asymmetric cell division , we analyzed egl-1 ( lf ) mutants , in which many apoptotic events , including the death of the NSM sister cells , are blocked . We found that , like in wild-type animals , the NSM neuroblast divided asymmetrically in egl-1 ( lf ) mutants ( NSM sister cell size on average 0 . 52 times the size of the NSM ) ( Figure 3B , egl-1 ( n1084n3082 ) ) . Thus , the increase in NSM sister cell size observed in dnj-11 ( bc212 ) animals is a result of a defect in asymmetric cell division . Therefore , we conclude that dnj-11 is required for asymmetric NSM neuroblast division and , by inference , is required for the displacement of the mitotic spindle along the cell division axis . Furthermore , because in dnj-11 ( bc212 ) animals the defect in asymmetric NSM neuroblast division correlates with the defect in NSM sister cell death , asymmetric NSM neuroblast division might be critical for the specification of the apoptotic fate of the NSM sister cell . To determine whether dnj-11 is required for asymmetric cell division in lineages other than the NSM lineage , we analyzed a subset of the other asymmetric cell divisions ( including the first division of the zygote ) that take place during C . elegans development and that give rise to daughter cells of different sizes and fates . We found that none of these divisions was affected by dnj-11 ( bc212 ) , suggesting that dnj-11 is not required for asymmetric cell division in general ( Table S7 and unpublished data ) . Like dnj-11 ( bc212 ) , the ces-1 ( gf ) mutation n703 and a putative null mutation of ces-2 , bc213 , prevent the death of the NSM sister cells ( Table 1 , ces-1 ( n703gf ) , ces-2 ( bc213 ) ) . Unexpectedly , we found that ces-1 ( n703gf ) and ces-2 ( bc213 ) also disrupt the asymmetric division of the NSM neuroblast and result in the production of two cells of similar sizes ( NSM sister cell sizes on average 0 . 94 and 1 . 06 times the size of the NSM ) ( Figure 3B , ces-1 ( n703gf ) , ces-2 ( bc213 ) ) . These results demonstrate that increased levels of ces-1 function can prevent asymmetric NSM neuroblast division and that , like dnj-11 , ces-2 is required for asymmetric NSM neuroblast division . To determine whether ces-1 also acts downstream of dnj-11 and ces-2 during asymmetric NSM neuroblast division , we determined whether the loss of ces-1 function can suppress the defects in asymmetric NSM neuroblast division observed in dnj-11 ( bc212 ) and ces-2 ( bc213 ) animals . We found that ces-1 ( n703n1434 ) partially suppressed the defects in asymmetric NSM neuroblast division that are observed in dnj-11 ( bc212 ) and ces-2 ( bc213 ) animals ( Figure 3B ) . This finding suggests that ces-1 acts downstream of dnj-11 and ces-2 during asymmetric NSM neuroblast division as well . Shortly after the division of the NSM neuroblast has been completed , the NSM is located at a position that is medial and ventral to the NSM sister cell ( Figure 3A ) . The position of the NSM relative to the NSM sister cell implies that in wild-type animals , the cleavage plane of the NSM neuroblast is along the ventral/lateral to dorsal/medial axis . While determining the size of the NSM and NSM sister cell using the plasma membrane–targeted GFP fusion protein , we observed that in two out of nine ces-2 ( bc213 ) animals , the medial daughter cell was located dorsally rather than ventrally to the lateral daughter cell ( Figure 3A ) . This observation suggests that the loss of ces-2 function not only affects the displacement of the mitotic spindle along the cell division axis in the NSM neuroblast , but also the orientation of the cell division axis and , hence , the orientation of the cleavage plane . To observe more directly the division of the NSM neuroblast , we used a transgene that expresses a DNA-targeted GFP fusion protein ( Phis-24his-24::gfp ) ( M . Dunn , G . Seydoux , J . Waddle , personal communication ) . This fusion protein allowed us to determine the axis along which the separated chromatids move during anaphase of the NSM neuroblast division and therefore the cleavage plane of the dividing NSM neuroblast . In wild-type embryos , the chromatids of the future NSM move to the ventral/medial side , whereas the chromatids of the future NSM sister cell move to the dorsal/lateral side , confirming that the cleavage plane of the NSM neuroblast is along the ventral/lateral to dorsal/medial axis ( Figure 4 ; +/+ ) . As in wild-type embryos , we found that in ces-2 ( bc213 ) embryos , the cleavage plane in six out of 11 NSM neuroblasts was along the ventral/lateral to dorsal/medial axis . In three cases , however , the cleavage plane was reversed , and the cells divided along the ventral/medial to dorsal/lateral axis . Furthermore , in two cases , the cleavage plane was either along the lateral to medial or ventral to dorsal axis ( Figure 4B and 4D; ces-2 ( bc213 ) ) . We next analyzed dnj-11 ( bc212 ) and ces-1 ( n703gf ) embryos and found that the orientation of the cleavage plane of the NSM neuroblast was disrupted in two out of nine and three out of 10 NSM neuroblasts , respectively ( Figure 3D; dnj-11 ( bc212 ) , ces-1 ( n703gf ) ) . Based on these findings , we conclude that dnj-11 , ces-2 , and ces-1 not only play a role in the displacement of the mitotic spindle along the cell division axis in the NSM neuroblast but also in the orientation of the cell division axis . To investigate how dnj-11 and ces-2 antagonize ces-1 function in the NSM lineage , we constructed a functional transgene that drives the expression of a CES-1::yellow fluorescent protein ( YFP ) fusion protein under the control of the ces-1 promoter ( Pces-1ces-1::yfp ) . In wild-type animals , we failed to detect CES-1::YFP in NSM neuroblasts or NSM sister cells . However , we observed CES-1::YFP in two out of 17 NSMs analyzed ( Figure 5A and 5B , +/+ ) . ( CES-1::YFP was observed in lineages other than the NSM lineage in all animals examined . ) This observation suggests that the level of CES-1::YFP and hence , most probably of endogenous CES-1 protein , is relatively low in the NSM lineage and that it is higher in NSMs than in NSM neuroblasts or NSM sister cells . In dnj-11 ( bc212 ) and ces-2 ( bc213 ) animals , we observed CES-1::YFP not only in NSMs but also in NSM sister cells and NSM neuroblasts ( Figure 5A and 5B , dnj-11 ( bc212 ) , ces-2 ( bc213 ) ) . These results demonstrate that dnj-11 and ces-2 antagonize ces-1 function in the NSM lineage by reducing the level of CES-1 protein in the NSM neuroblast , thereby restricting the presence of CES-1 protein to the NSM . To determine whether dnj-11 and ces-2 act at the transcriptional or posttranscriptional level to affect the level of CES-1 protein , we constructed a transgene that drives the expression of the GFP protein under the control of the ces-1 promoter ( Pces-1gfp ) and analyzed GFP expression in the NSM lineage . In wild-type animals , GFP was not detected in NSM neuroblasts , NSMs , or NSM sister cells ( Figure 5B , +/+ ) . However , in dnj-11 ( bc212 ) and ces-2 ( bc213 ) animals , GFP was detected in NSM neuroblasts , NSMs , and NSM sister cells ( Figure 5B , dnj-11 ( bc212 ) , ces-2 ( bc213 ) ) . Based on these results , we conclude that dnj-11 and ces-2 restrict the presence of CES-1 protein to the NSM by reducing ces-1 transcription in the NSM lineage .
Snail family transcription factors have previously been implicated in the regulation of apoptosis in mammalian cells . In the case of C . elegans , it has been suggested that the Snail protein , CES-1 , might normally function within the NSM lineage to repress transcription of the pro-apoptotic gene , egl-1 , thereby allowing NSM survival . Our data support this idea , because we find that CES-1 protein is present at higher levels in the NSMs than in the NSM sister cells . Therefore , we propose that CES-1 acts as a cell-fate determinant in the NSMs to ensure the survival of the NSMs and their differentiation into serotonergic neurons . Conversely , we propose that the absence of CES-1 protein in the NSM sister cells determines the death of the NSM sister cells ( Figure 6A ) . Our results also indicate that the CES-1 protein acts in the NSM neuroblasts to affect the orientation of the cleavage plane and , hence , asymmetric cell division ( Figure 6 ) . Therefore , at least in the NSM lineage , the CES-1 protein represents a functional link between the cellular machinery that causes asymmetric cell division and the cellular machinery that causes the apoptotic death of specific cells during C . elegans development . At least to our knowledge , this is the first demonstration that apoptosis can directly be controlled by asymmetric cell division . Like the function of ces-1 in NSM fate determination , the function of ces-1 in asymmetric NSM neuroblast division is redundant to that of one or more unidentified genes . Similarly , the Snail-related proteins Snail , Escargot , and Worniu of D . melanogaster act redundantly to cause asymmetric cell division [14 , 15 , 17] . The HLF-like bZIP transcription factor , CES-2 , is thought to act as a direct repressor of CES-1 Snail transcription [19 , 21] . We have found that the dnj-11 MIDA1-like gene acts in concert with CES-2 as a negative regulator of CES-1 expression . Reducing either dnj-11 or ces-2 function results in an increased level of ces-1 transcription within the NSM lineage , disrupts asymmetric NSM neuroblast division , and prevents the death of the NSM sister cells . Based on our data , dnj-11 and ces-2 could either act in parallel or in a single linear pathway to antagonize ces-1 function ( Figure 6 ) . MIDA1-like proteins have been implicated in transcriptional regulation [28–30] and are components of a ribosome-associated chaperone referred to as RAC , which co-translationally interacts with nascent polypeptides thereby affecting translational accuracy and termination as well as protein folding [31 , 32] . The DNJ-11 protein predominantly localizes to the cytosol in a punctate pattern , which suggests that rather than regulating ces-1 transcription directly , DNJ-11 might affect the translation and/or folding of a regulator of ces-1 transcription . Based on our results , we propose that by repressing the transcription of the snail-related ces-1 gene , the HLF-like bZIP transcription factor CES-2 and the MIDA1-like chaperone DNJ-11 ensure that the CES-1 protein is present at an appropriate , low level in the early NSM neuroblast ( Figure 6A ) . A low level of CES-1 protein in the early NSM neuroblast would allow the expression at a certain level of a “polarity factor” that is required for the asymmetric division of the late NSM neuroblast . Furthermore , a complex that localizes to the ventral/medial side of the late NSM neuroblast would restrict the polarity factor to the ventral/medial side , thereby promoting the displacement of the mitotic spindle along the cell division axis and resulting in a shift of the cleavage plane and the asymmetric division of the cell . Finally , the asymmetrically localized polarity factor would also restrict CES-1 to the ventral/medial side and thereby cause its segregation predominantly into the NSM , thus resulting in the repression of egl-1/BH3-only transcription in the NSM and the survival of the NSM ( Figure 6A ) . In this model , the CES-1 protein not only is a component of the cellular machinery that causes asymmetric NSM neuroblast division , but also one of its targets . The identity of the polarity factor as well as the signals and mechanisms that cause its asymmetric localization or activation remain to be determined . The Snail-related genes snail , escargot , and worniu of D . melanogaster function in asymmetric neuroblast division by promoting the expression of the gene inscuteable , which encodes an adaptor protein required for asymmetric neuroblast division . It will be of interest to determine whether the C . elegans ortholog of inscuteable , the gene insc-1 ( F43E2 . 3 ) ( INSC-1 accession number AAC71125 ) [37] , plays a role in asymmetric cell division and in asymmetric NSM neuroblast division , in particular . dnj-11 ( bc212 ) does not affect asymmetric cell divisions that are known to be regulated by the C . elegans genes ham-1 [4 , 38] , pig-1 [5] , dsh-2 [39] , or hlh-14 [40] . Conversely , the loss of ham-1 , pig-1 , dsh-2 , or hlh-14 function does not affect the asymmetric division of the NSM neuroblast . Therefore , the dnj-11 MIDA1 , ces-2 HLF , and ces-1 Snail pathway might be independent of these genes . Furthermore , the first division of the zygote , which occurs asymmetrically and is known to be regulated by the C . elegans par genes [41] , is not affected by dnj-11 ( bc212 ) , ces-2 ( bc213 ) , or ces-1 ( n703gf ) ( J . Hatzold , B . Conradt , unpublished data ) . Whether the par genes function in the asymmetric division of the NSM neuroblast remains to be determined . At least to our knowledge , this is the first evidence that a HLF-like bZIP transcription factor plays a role in asymmetric cell division . Furthermore , our studies provide a functional link between the known roles in asymmetric cell division of MIDA1-like chaperones and Snail-related transcription factors , and hence suggest the existence of a pathway involved in asymmetric cell division that is conserved throughout the plant and animal kingdoms [14–16 , 26] . Therefore , it will be of interest to determine whether HLF-like and Snail-related proteins also contribute to asymmetric cell division in V . carteri and whether MIDA1-like and HLF-like proteins also participate in asymmetric neuroblast division in D . melanogaster . Furthermore , our results hint at the possibility that MIDA1-like , HLF-like , and Snail-related proteins might play a role in asymmetric cell division in vertebrates . Specifically , it has recently been reported that HLF and the Snail-related protein Slug of mammals have functions in stem cells [12 , 42] . In addition , there is increasing evidence that the process of asymmetric cell division plays a crucial role in the ability of stem cells to self renew [43] . Therefore , we hypothesize that a dnj-11 MIDA1 , ces-2 HLF , ces-1 Snail–like pathway might be important for stem cell renewal by allowing asymmetric stem cell division . At least to our knowledge , this is also the first evidence that a MIDA1-like protein plays a role in the regulation of apoptosis . Furthermore , our studies have identified a new component of a conserved cell-death specification pathway composed of C . elegans ces-2 HLF and ces-1 Snail . Like CES-2 and CES-1 , HLF and the Snail-related protein Slug of mammals have previously been implicated in the regulation of apoptosis [19 , 21 , 44–46] . Therefore , it will be of interest to determine whether MIDA1-like proteins also have an apoptotic role in mammals and act in a HLF- , and Slug-dependent pathway . Finally , based on our work , we consider it a possibility that the roles in vertebrates of a dnj-11 MIDA1 , ces-2 HLF , ces-1 Snail–like pathway in asymmetric cell division and apoptosis might be functionally linked as well . Specifically , we speculate that such a pathway could cause stem cell renewal through asymmetric cell division and control stem cell numbers through apoptosis [47] . The oncogenic form of human HLF , the E2A-HLF fusion protein , found in patients carrying the t ( 17;19 ) ( q22;p13 ) translocation , gives rise to pro–B cell acute lymphoblastic leukemia ( ALL ) in adolescents [48] . The E2A-HLF fusion protein is composed of the trans-activation domain of the bHLH protein E2A and the DNA-binding domain of HLF [49] . It has been proposed that E2A-HLF causes leukemic transformation of pro-B cells by blocking their apoptotic death . Specifically , it has been proposed that E2A-HLF inappropriately activates the transcription of the snail-related gene , slug , which encodes a direct repressor of the egl-1-like , pro-apoptotic BH3-only gene puma , thereby causing the survival of pro-B cells that are normally programmed to undergo apoptosis [44–46] . Based on our results , we speculate that like C . elegans CES-2 and CES-1 , the proteins HLF and Slug might not only function to control the expression of a pro-apoptotic BH3-only gene but to cause asymmetric cell division . Hence , in patients with the t ( 17;19 ) ( q22;p13 ) translocation , the presence of the E2A-HLF fusion protein and elevated levels of Slug protein might affect aspects of the pro-B cell fate other than their apoptotic fate and/or might alter the division of the lymphoid progenitors that produce pro-B cells . Finally , it will be of interest to determine whether the human MIDA1-like gene MPP11 [27] , which is expressed in hematopoietic lineages as well as other tissues , plays a role in E2A-HLF- and Slug-mediated tumorigenesis .
C . elegans strains were cultured as described [50] . Bristol N2 was used as the wild-type strain . The strain CB4856 ( Hawaii ) was used in conjunction with N2 for SNP mapping . Mutations and transgenes used in this study are listed below and are described [51] , except where noted otherwise: LGI: ces-1 ( n703n1434 ) , ces-1 ( n703gf ) , ces-2 ( n732ts ) , ces-2 ( bc213 ) ( this study ) . LGII: rrf-3 ( pk1426 ) [52] . LGIII: gmIs12 ( Psrb-6gfp ) [53] , ced-4 ( n1162 ) . LGIV: unc-5 ( e53 ) , dnj-11 ( bc212 , tm2859 ) ( this study ) , dpy-20 ( e1282 ) , nDf41 [54] , bcIs25 ( Ptph-1gfp ) [20] . LGV: nIs83 ( Pmec-4gfp ) , bcIs37 ( Pegl-1gfp ) [20] , egl-1 ( n1084n3083 ) [55]; akIs3 ( Pnmr-1gfp ) [56] . ( LGX ) nIs106 ( Plin-11gfp ) [57] , dtIs372 ( M . Dunn , G . Seydoux , J . Waddle . personal communication ) . Additional stable transgenes used: tIs38 ( Ppie-1gfp::ph ( PLC1δ1 ) ) [36] , kyIs39 ( Psra-6gfp ) [53] , bcIs58 ( Pces-1ces-1::yfp ) ( this study ) . RNA-mediated interference ( RNAi ) by feeding was performed as described using 6 mM IPTG [58] . Standard genetic techniques were used to map dnj-11 between unc-5 and dpy-20 on LGIV . SNP mapping was used to locate dnj-11 between the SNPs C43G2:22057 and C17H12:33927 . The cosmid F38A5 as well as a 3 , 565-bp subclone of F38A5 contained in plasmid pBC484 ( Pdnj-11dnj-11 ) rescued the NSM sister cell survival and brood size phenotype observed in bc212 animals ( Table 2 and Table S5 , dnj-11 ( bc212 ) ; Pdnj-11dnj-11 ) . In addition , partially reducing dnj-11 function by RNA-mediated interference ( RNAi ) causes 22% NSM sister cell survival ( Table 1 ) . dnj-11 plasmids: pBC484 ( Pdnj-11dnj-11 ) was generated by inserting a EcoRV-PvuI subclone of cosmid F38A5 into the EcoRV site of pBluescript II KS + ( Stratagene ) . The gfp sequence was amplified from pPD95 . 02 ( gift of A . Fire , Stanford Scool of Medicine , Stanford , California ) with appropriate primers ( sequence of these and all other primers are available on request ) and inserted into the BsmI site of pBC484 to create a C terminal in frame fusion of dnj-11 to gfp ( Pdnj-11dnj-11::gfp ) . The Pdnj-11dnj-11::gfp transgene rescued the NSM sister cell survival phenotype of dnj-11 ( bc212 ) mutants , demonstrating that it is functional ( Table 2 , dnj-11 ( bc212 ) ; Pdnj-11dnj-11::gfp ) . To generate Pdnj-11dnj-11::gfp mutant plasmids , site-directed PCR mutagenesis was performed to mutate CAC to CAA ( H129Q ) , TGG to GGG ( W456G ) , and TTC to GGC ( F578G ) . The expression levels of the transgenes Pdnj-11dnj-11 ( H129Q ) ::gfp , Pdnj-11dnj-11 ( W456G ) ::gfp , Pdnj-11dnj-11 ( F578G ) ::gfp , and Pdnj-11dnj-11 ( W456G F578G ) ::gfp were similar to that of the wild-type transgene ( Pdnj-11dnj-11::gfp ) , and the subcellular localization of the resulting fusion proteins was indistinguishable from that of DNJ-11::GFP ( unpublished data ) . ces-1 plasmids: The plasmid pBC510 ( Pces-1ces-1::yfp ) was generated by cloning an AflII-SpeI fragment from cosmid F43G9 containing the ces-1 rescuing fragment [19] into the EcoRV site of pBluescript II KS + ( pBC482A ) , and inserting yfp amplified from pvdB#3 [59] into the SwaI site to generate a C terminal in frame fusion . The Pces-1ces-1::yfp transgene was able to block the ability of ces-1 ( n703n1434 ) to suppress the NSM sister cell survival phenotype of dnj-11 ( bc212 ) and ces-2 ( n732 ) animals , demonstrating that it is functional ( unpublished data ) . To generate plasmid pBC664 ( Pces-1gfp ) , first ces-1 upstream and downstream regulatory regions were amplified by PCR , a PmeI site was introduced 3′ of upstream and 5′ of downstream regulatory regions , and PCR products were cloned into XmaI/NcoI digested pBC482A to obtain the ces-1 locus without coding region , 5′ and 3′ untranslated regions ( UTRs ) , and introns ( pBC656 ) . Because intron 4 of ces-1 is highly conserved between C . elegans and C . briggsae , intron 1 of the gfp sequence of plasmid pPD95 . 77 ( gift of A . Fire ) was replaced with intron 4 of ces-1 by PCR fusion . The SmaI/SpeI fragment of pPD95 . 77_ces-1 intron containing the gfp sequence and the unc-54 3′ UTR was inserted into the PmeI site of pBC656 to generate pBC664 . Germline transformation was performed as described [60] . Cosmids were injected at a concentration of 10 ng/μl with pPD93 . 97 ( Pmyo-3gfp ) at 50 ng/μl as coinjection marker . Plasmids were injected at a concentration of 10 ng/μl with pRF4 ( rol-6 ( su1006 ) ) at 50 ng/μl as coinjection marker . pBC510 ( Pces-1ces-1::yfp ) was injected into N2 to create an extrachromosomal array and integrated using EMS mutagenesis [50] to generate bcIs58 . The strain carrying bcIs58 was backcrossed three times to N2 . The NSM sister cell survival was scored as previously described [20] . Microscopy of living embryos was performed by mounting embryos on 2%–5% agar pads in M9 buffer , sealing them with petroleum jelly , and using a Zeiss Axioskop2 equipped with epifluorescence , a Micromax CCD camera ( Princeton Instruments ) , and Metamorph software . NSM neuroblasts , NSMs , and NSM sister cells were identified based on the position of their nuclei using Nomarski optics . Z-series were taken with a Z-distance of 0 . 5 μm ( analysis of Pces-1ces-1::yfp , Pces-1gfp , and Phis-24his-24::gfp expression ) and 0 . 25 μm ( determination of cell size ) . Epifluorescence Z-series were deconvolved using the AutoDeblur Gold WF AutoVisualize software ( Media Cybernetics ) . The cell size of NSMs and NSM sister cells was determined 10 to 15 min after the NSM neuroblast had started to divide , as indicated by the breakdown of the nuclear envelop , which was observed by Nomarski optics . To visualize the outline of a cell , a plasma membrane–targeted GFP fusion protein ( Ppie-1gfp::ph ( PLC1δ1 ) ) was used . The area of the cross section of a cell was measured in each section of a Z-series using Metamorph software . To estimate the difference in volume between NSM and NSM sister cell , the values of each section were added , and the sum obtained for the NSM sister cell was divided by the sum obtained for the NSM . To determine the orientation of the cleavage plane of the NSM neuroblast , chromatids were visualized using a His24-GFP fusion protein ( Phis-24his-24::gfp , dtIs372 ) ( gift of M . Dunn and G . Seydoux , Johns Hopkins University , Baltimore , Maryland; and J . Waddle , Southern Methodist University , Dallas , TX ) . The NSM neuroblast was identified by Nomarski optics before the start of the division and consecutive Z series were taken in 1 min time intervals until the completion of the NSM neuroblast division . Embryos were prepared in 10 μl on poly L-lysine coated slides and fixed and stained as described [61] . Slides were mounted in 1 μg/ml DAPI in PBS 1:1 diluted with VectaShield ( Vector Laboratories ) . GFP was detected using the anti-AFP antibody monoclonal antibody 3E6 ( Qbiogene ) .
The NCBI Entrez Protein ( http://www . ncbi . nlm . nih . gov/sites/entrez ? db=protein ) accession numbers for genes and gene products discussed in this paper are: NP_501006 ( DNJ-11 ) , NP_506575 ( EGL-1 ) , NP_492338 ( CES-1 ) , NP_493610 ( CES-2 ) , NP_033610 ( mouse MIDA1 ) , NP_055192 ( human MPP11 ) , NP_011801 ( S . cerevisiae Zuotin ) . The GenBank ( http://www . ncbi . nlm . nih . gov/ ) accession number for V . carteri GlsA is AF_106963 . | Asymmetric cell division and apoptosis ( programmed cell death ) are two fundamental processes that are important for the development and function of multicellular organisms . Asymmetric cell division creates daughter cells of different fates , and this is critical for the generation of cellular diversity . Apoptosis eliminates superfluous cells from the organism , which is critical for cellular homeostasis . We found that the processes of asymmetric cell division and apoptosis can be functionally linked . Specifically , we show that asymmetric cell division in the nematode Caenorhabditis elegans is mediated by a pathway involving three genes , dnj-11 MIDA1 , ces-2 HLF , and ces-1 Snail , that directly control the enzymatic machinery responsible for apoptosis . Interestingly , the role of this pathway in asymmetric cell division and the control of apoptosis might be evolutionarily conserved . Furthermore , it might have an unexpected role in stem cell biology: the process of asymmetric cell division plays an essential role in the ability of stem cells to self-renew , and the mammalian counterparts of two components of the dnj-11 MIDA1 , ces-2 HLF , ces-1 Snail pathway have recently been implicated in stem cell function . For this reason , we speculate that a dnj-11 MIDA1 , ces-2 HLF , ces-1 Snail–like pathway might function in stem cells to coordinate self-renewal and apoptosis and , hence , the number of stem cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"developmental",
"biology",
"cell",
"biology",
"genetics",
"and",
"genomics"
] | 2008 | Control of Apoptosis by Asymmetric Cell Division |
Cytokines such as TNF and FASL can trigger death or survival depending on cell lines and cellular conditions . The mechanistic details of how a cell chooses among these cell fates are still unclear . The understanding of these processes is important since they are altered in many diseases , including cancer and AIDS . Using a discrete modelling formalism , we present a mathematical model of cell fate decision recapitulating and integrating the most consistent facts extracted from the literature . This model provides a generic high-level view of the interplays between NFκB pro-survival pathway , RIP1-dependent necrosis , and the apoptosis pathway in response to death receptor-mediated signals . Wild type simulations demonstrate robust segregation of cellular responses to receptor engagement . Model simulations recapitulate documented phenotypes of protein knockdowns and enable the prediction of the effects of novel knockdowns . In silico experiments simulate the outcomes following ligand removal at different stages , and suggest experimental approaches to further validate and specialise the model for particular cell types . We also propose a reduced conceptual model implementing the logic of the decision process . This analysis gives specific predictions regarding cross-talks between the three pathways , as well as the transient role of RIP1 protein in necrosis , and confirms the phenotypes of novel perturbations . Our wild type and mutant simulations provide novel insights to restore apoptosis in defective cells . The model analysis expands our understanding of how cell fate decision is made . Moreover , our current model can be used to assess contradictory or controversial data from the literature . Ultimately , it constitutes a valuable reasoning tool to delineate novel experiments .
Only the apoptotic caspase-dependent pathway downstream of FAS and TNF receptors is considered here . Upon engagement by their ligands and in the presence of FADD ( FAS-Associated protein with Death Domain ) , a specific Death Inducible Signalling Complex ( DISC-FAS or DISC-TNF in Figure 1 ) forms and recruits pro-caspase-8 . This leads to the cleavage and activation of caspase-8 ( CASP8 ) . In the so-called type II cells , CASP8 triggers the intrinsic or mitochondria-dependent apoptotic pathway , which also responds to DNA damage directly through the p53-mediated chain of events ( not detailed here ) . CASP8 cleaves the BH3-only protein BID ( not explicitly included in the diagram ) , which can then translocate to the mitochondria outer membrane . There , BID competes with anti-apoptotic BH3 family members such as BCL2 for interaction with the proteins BAX or BAK ( BAX will stand here for both BAX and BAK ) . Consequently , oligomerisation of BAX results in mitochondrial outer membrane permeabilisation ( MOMP ) and the release of pro-apoptotic factors . Once released to the cytosol , cytochrome c ( Cyt_c ) interacts with APAF1 , recruiting pro-caspase-9 . In presence of dATP , this enables the assembly of the apoptosome complex ( referred to as ‘Apoptosome’ in Figure 1A , lumping APAF1 and pro-caspase-9 ) , responsible for caspase-9 activation , followed by the proteolytic activation of pro-caspase-3 ( CASP3 ) [9] . By cleavage of specific targets , the executioner caspases ( CASP3 in the model ) are responsible for major biochemical and morphological changes characteristic of apoptosis . SMAC/DIABLO ( SMAC ) is released during MOMP to the cytosol , where it is able to inactivate the caspase inhibitor XIAP [10] . CASP3 also participates in a positive circuit by inducing the activation of CASP8 [11] , [12] . In type I cells , CASP8 directly cleaves and activates executioner caspases such as CASP3 ( not described ) . Here , we consider mainly a mode of cell death with morphological features of necrosis , which occurs when apoptosis is impeded in cells treated with cytokines [13] or in some specific cell lines such as L929 cells when exposed to TNF [14] . In primary T cells , if caspases are inhibited , activation of TNFR or FAS causes necrosis via a pathway that requires the protein RIP1 and its kinase activity ( RIP1K ) [13] . This RIP1-dependent cytokine-induced necrotic death defines necroptosis [15] , [16] . A genetic screen recently identified other genes necessary for this type of cell death [17] . However , a precise description of this pathway is still lacking . Reactive oxygen species ( ROS ) were proposed to be involved downstream of RIP1 [18] . ROS are also thought to play a key role in the control of mitochondria permeability transition ( MPT ) , since they are produced by damaged mitochondria and can oxidize mitochondrial components , thus favouring MPT [19] , [20] , [21] . Furthermore , the role of mitochondria in necrosis is highlighted through the involvement of MPT , which causes a fatal drop in ATP level and leads to necrotic death . Indeed , MPT results from the inhibition of ATP/ADP exchange at the level of mitochondrial membranes , or from the inhibition of oxidative phosphorylation decreasing cellular ATP level and causing energy failure [21] , [22] . Although there is evidence that necrosis is also triggered by TNF- and FAS-independent pathways , these are not yet considered in this study . These pathways include , for example , calpain-mediated cleavage of AIF followed by its nuclear translocation [23] , [24] , or PARP-1-mediated NAD+ depletion [24] , [25] . NFκB represents a family of transcription factors that play a central role in inflammation , immune response to infections and cancer development [26] . The ubiquitination of RIP1 at lysine 63 by cIAP leads to the activation of IKK and ultimately that of NFκB [27] . In different cell types , especially in tumour cell lines , activation of NFκB inhibits TNF-induced cell death [28] . This effect is mediated by NFκB target genes: cFLIP inhibits recruitment of CASP8 by FADD [29]; anti-apoptotic BCL2 family members inhibit MOMP and MPT [30] , [31] , [32]; XIAP acts as a caspase inhibitor [33]; and ferritin heavy chain [34] or mitochondrial SOD2 [35] decrease ROS levels ( these mechanisms are represented in Figure 1A by a direct inhibitory arrow from NFκB to ROS ) . For the sake of simplicity , other NFκB target genes that are known to inhibit TNF-induced apoptosis are provisionally omitted in the model ( e . g . , A20; cf . [36] , [37] ) . Our goal here is to provide a simplified but yet rigorous model of the mechanisms underlying cell fate selection in response to the engagement of FAS and TNF receptors . We have proceeded in several steps . First , we have assembled a regulatory network covering the main experimental data . Species and interactions were selected on the basis of an extensive literature search and integrated in the form of a diagram or “regulatory graph” . This diagram is then translated into a dynamical model . Our analysis initially focuses on the determination of the asymptotic properties of the system for different conditions , which correspond to the possible phenotypes that the model can account for . Next , we analyse the different trajectories leading to each phenotype in the wild type and mutant situations . As quantitative data are still largely lacking for this system , we use a qualitative logical formalism and its implementation in the GINsim software [38] . As we shall see , proper model analysis can assess where and when cell fate decisions are made , provide novel insight concerning the general structure of the network , in particular concerning the occurrence of cross-talks between pathways , and predict novel mutant phenotypes and component activity patterns .
As mentioned before , the pathways are highly intertwined ( Figure 1A ) . For instance , the survival pathway interacts with the apoptotic pathway at different points: cFLIP inhibits CASP8; BCL2 blocks mitochondria pore opening through inhibition of BAX ( and BAK , implicitly represented in our model ) ; and XIAP blocks the activity of both CASP9 in the apoptosome and CASP3 . Conversely , the apoptotic pathway negatively regulates NFκB activity through the CASP8-mediated cleavage and inactivation of RIP1 upstream of NFκB . Because RIP1 operates upstream of the necrotic pathway , this regulation also impacts necrosis . Moreover , for the apoptosome to form , dATP ( or/and ATP ) is ( /are ) needed . Consequently , in our model , when necrosis occurs , ATP production drops , terminating apoptosis . Regarding the influence of the survival pathway on the necrotic one , NFκB tentatively stimulates the production of anti-oxidants that shuts off ROS level . Both the necrotic and the apoptotic pathways are able to interact with the survival pathway through the action of cIAP1/2 , referred to as cIAP in our model . More precisely , cIAP1 and 2 are E3-ubiquitin ligases that target RIP1 for K63-linked polyubiquitination . They are essential intermediates in the activation of NFκB downstream of TNF receptor [27] . Some synthetic molecules that mimic the N-terminal of SMAC IAP-interacting motif have been shown to induce cIAP1/2 auto-ubiquitination and subsequent proteasomal degradation , thus blocking TNF-dependent NFκB activation [39] , [40] . Tentatively , mitochondrial permeabilization in the apoptosis or necrosis pathways could block TNF-induced NFκB activation through the release of SMAC into the cytosol thereby causing the inhibition of c-IAP1/2 . Initially , cIAP was not included in the model , which led to discrepancies between model simulations and published data . Indeed , in FADD or CASP8 deletion mutants , our preliminary model predicted only survival ( not shown ) , whereas both necrotic and survival phenotypes were observed in experiments in the presence of TNF or FAS [41] , [42] , [43] . The consideration of the path MOMP⇒SMAC = |cIAP⇒NFκB enabled us to eliminate the discrepancies , both necrotic and survival phenotypes were then obtained in the simulations , although it does not preclude other mechanisms . To transform the static map shown in Figure 1A into a dynamical model accounting for the different scenarios or set of events leading to one of the three phenotypes , we have to define proper dynamical rules . Since there is little reliable quantitative information on reaction kinetics and cellular conditions leading to one or another phenotype , these rules must be sufficiently flexible to cover all possible scenarios following death receptor activation . The nodes encompassed in the map represent different things: simple biochemical components ( receptors , ligands , proteins or metabolites ) : TNF , FASL , TNFR , FADD , FLIP , CASP8 , RIP1 , IKK , NFκB , cIAP , BCL2 , BAX , Cyt_c , SMAC , ROS , XIAP , CASP3 , ATP ) ; specific modified forms of proteins: RIP1K ( active RIP1 kinase ) , RIP1ub ( K63-ubiquitinated RIP1 ) ; complexes of proteins: DISC-TNF ( corresponding to TRADD , TRAF2 , FADD , proCASP8 ) , DISC-FAS ( corresponding to FAS , FADD , proCASP8 ) , apoptosome; cellular processes: MPT ( Mitochondrial Permeability Transition ) and MOMP ( Mitochondria Outer Membrane Permeabilisation ) . A Boolean variable is associated with each of these nodes , which can take only two logical values: “0” ( false ) , denoting the absence or inactivity of the corresponding component , and “1” ( true ) , denoting its active state . Furthermore , a logical rule ( or function ) is assigned to each node , defining how the different inputs ( incoming arrows ) combine to control its level of activation . For example , CASP8 can be activated ( its value is set to “1” ) by DISC-TNF or DISC-FAS , but only in the absence of cFLIP protein . This can be encoded into a logical rule as follows: ( DISC-TNF OR DISC-FAS ) AND NOT cFLIP . Several nodes correspond to simple inputs ( TNF , FASL and FADD ) . Their initial values are kept fixed during most simulations . On the basis of the regulatory graph and the associated logical rules , we then proceeded with the exploration of the dynamical properties of our model . We first focused on the identification of all stable states and on their biological interpretation . Then , we investigated the reachability of these stable states for different initial conditions , for both wild type and mutant cases . Details on the computational methods used are provided in the Methods section and in the supplementary Text S1 . The logical model has been filed in the BioModels database with the reference MODEL0912180000 . Analysis of the cell fate decision model ( Figure 1A ) led to the identification of the 27 stable states showed in Figure 2 . These stable states are the sole attractors of the system under the asynchronous assumption ( see Methods ) . They thus represent all possible cellular asymptotical states . In other words , whatever the initial conditions , a wild type cell will end up in one of these states if we wait long enough . A closer look reveals that several stable states correspond to each cellular fate , with few differing ( minor ) component values . This consideration led us to address the following questions: ( i ) Does a cluster structure exist in the distribution of internal stable states of the network ? ( ii ) If so , in these clusters , could the corresponding states be interpreted as slightly different realisations of the same cellular phenotype ? ( iii ) What would be the characteristic signature of each cluster ( conserved values of variables inside each cluster ) ? ( iv ) What is the number of independent variables defining the internal stable states of the network ? Standard statistical methods and clustering algorithms are applied to group stable states . Figure 3 displays a projection of the internal ( without inputs and outputs ) stable values into the 2D space defined by the first two principal components of the corresponding distribution . The first two principal components explain 52% and 20% of the total variation , respectively ( Table S1 ) . The first principal component can be associated with the activity of NFκB pathway , while the second is determined mainly by ATP and MPT status . These factors do appear to determine the principal ( independent ) degrees of freedom for the internal state of the network . A typical trajectory starting from any set of initial conditions will thus quickly converge to the region under the influence of these three components . The 2-D graph ( Figure 3 ) reveals a striking separation of the stable states into 4 clusters: one cluster ( blue circles ) devoid of significant activity , which we call the ‘naïve’ state; one cluster ( green rhombs ) corresponding to survival , with NFκB pathway activated; and two clusters corresponding to the two different modalities of cell death , apoptosis ( orange squares ) and necrosis ( purple triangles ) . K-means clustering using Euclidean and L1 distance perfectly reproduces these groupings , demonstrating that the compact groups easily distinguishable on the PCA plot indeed represent well-separated clusters in the original multidimensional space . Some interesting conclusions and predictions can be drawn just by looking at the values of each component in each phenotypical group . For instance , in the necrotic ( purple ) stable states , when FADD is present ( i . e . normal wild type conditions ) , RIP1 is always OFF and CASP8 ON , even though RIP1 is required and CASP8 is dispensable for necrosis to occur . This observation suggests a transient activation of RIP1 protein when switching on the necrotic pathway in response to death receptors . However , inactivation or cleavage of RIP1 is not per se a prerequisite for necrosis , nor is CASP8 activation . Indeed , for the mutant models in which CASP8 activation is impaired , such as CASP8 or FADD deletion , there exist necrotic attractors with RIP1 = 1 ( not shown ) . Our model thus predicts that TNF-induced necrosis could occur despite CASP8-mediated cleavage of RIP1 . An attractive experimental model in which such a transient activation of RIP1 could be tested is the mouse fibrosarcoma cell line L929 . Upon TNF exposure , these cells die by necrosis [44] and they have a functional CASP8 [45] , which is cleaved during TNF-induced cell death [46] . Since RIP1 controls both the activation of NFκB and the level of ROS , the same transient behaviour could be expected for the survival phenotype . However , this is not observed with our model , as RIP1 = 1 in all survival ( green ) stable states . This can be explained by the regulatory circuit involving RIP1 and NFκB , which is not functional in necrosis . Indeed , when NFκB is active , it can mediate the synthesis of cFLIP , an inhibitor of CASP8 , itself an inhibitor of RIP1 . Moreover , RIP1 is part of the positive circuit that keeps NFκB ON . The model thus suggests that a sustained RIP1 activity is needed for survival . How could this hypothesis be experimentally assessed ? If an experiment would reveal that RIP1 is only transiently activated upon death receptor activation , while NFκB remains activated , the model would be contradicted . In that case , one would need to look for other components capable to maintain NFκB active . The stable state analysis described above provides a first validation of the master model presented in Figure 1A . On this basis , we performed a more detailed analysis of the dynamics of the system . We investigated which cell fates ( stable states ) can be reached from specific initial conditions . Given a set of reachable stable states , can we say something about their relative “attractivity” ? To avoid the combinatorial explosion of the number of states to consider , we have reduced the number of components while preserving the relevant dynamical properties of the master model ( Figure 1B ) . Details of how this reduction is performed are provided in the Methods section . The resulting network encompasses 11 components . The corresponding Boolean rules are listed in Table 1 . The size of the transition graph ( 211 = 2048 ) is now amenable to a detailed dynamical analysis . First , the set of attractors of this reduced model is identified: 13 attractors are obtained , which are all stable states matching those found for the master model when the input variable FADD = 1 . Recall that , in the master model , both values of FADD were considered leading to 13 stable states with FADD = 1 and 14 stable states with FADD = 0 . Using the theoretical results presented in [47] ( mainly Theorem 1 ) , we can conclude that the 13 stable states of Figure 2 are the only attractors of the master model when FADD = 1 . Based on the reduced model defined in Figure 1B and Table 1 , we derived 15 model variants representing biologically plausible perturbations . We will abusively use the term “mutant” to refer to these variants , even though they do not all technically correspond to mutations . For instance , the “z-VAD mutant” simulates the effect of caspase inhibitor z-VAD-fmk . Each mutant simulation consists in a local alteration of our reduced model , which can be qualitatively compared with results reported in the literature . In the Boolean framework , such alterations amount to force the level of certain variables to zero in the case of a gene deletion , or to one in the case of a component over-expression . As we are using the reduced version of the master model , some perturbed components may be hidden by the reduction process . In such cases , we change the logical rules of their ( possibly indirect ) targets to take into account their effects . Table 2 lists the 15 variants of the model considered , along with the modified logical rules , the expected effects on the phenotypes according to the literature , and short descriptions of simulation results . The references provided in Table 2 cover experiments performed on different cell types and with different experimental conditions . In contrast , our cell fate model represents mechanisms of cell fate decision in a generic cell , qualitatively recapitulating a wide variety of cellular contexts . Given a cellular system , its response to the activation of death receptors is determined by the logical rules . However , the generic model presented here considers equally all possible contexts and regulatory combinations . To evaluate the relative likelihood of having a particular response in a randomly chosen cellular system , we count the relative number of possible trajectories from the stimulated ‘naïve’ state to a given phenotype . This analysis gives an idea on what is possible or forbidden in a ‘generic’ cell . Using dedicated methods and software [48] , the set of reachable stable states is calculated , starting from selected physiological initial conditions , for the wild type and mutant models . The physiological state is defined by fixing the variables ATP and cIAP to “1” and all the other ones to “0” . Different combinations of TNF and FASL are considered . The probability to reach each phenotype is computed as a fraction of the paths in the graph that link physiological initial conditions to each cell fate ( Figure 4 for reduced model and Figure S4 for master model ) . As expected , the absence of TNF and FASL can only lead to the ‘naïve’ state ( except of course when caspase-8 or NFκB are over-expressed , for obvious reasons ) . This means that the inputs ( TNF and FASL ) are needed for the system to effectively trigger the decision process . This was expected since intracellular death signals are not yet taken into account in the model . When TNF = 1 ( Figure 4 , right panel ) , for the wild type system , we observe that three outcomes or phenotypes are reachable from the initial condition , with different probabilities: ∼10% for necrosis , ∼30% for active survival and ∼60% for apoptosis . Although these probabilities cannot be directly compared with experimental results , they become useful when comparing different variants of the model . For instance , an increase ( or decrease ) of a phenotype probability between the wild type and a particular mutant can be interpreted as a gain ( or a loss ) of effectiveness of the corresponding pathway in that mutant . Such qualitative observations can then be confronted with published experimental results , which are summarized in the last column of Table 2 . In most cases , activation of FASL and TNF lead to similar effects ( not shown ) , except in the case of the FADD deletion mutant ( Figure 5 ) . As expected , this mutant cannot lead to cell death when FASL is ON . In contrast , necrosis is still possible in the presence of TNF . Interestingly , TNF-induced apoptosis is expected to be blocked [49] whereas the qualitative analysis shows that apoptosis is actually reachable in the model . Nevertheless , the probability of this phenotype is very low ( around 0 . 61% ) , which means that very few trajectories may lead to apoptosis and it would thus be difficult to obtain the corresponding cellular context . In the reachability analysis presented above , the value of TNF and FASL are kept constant and therefore always ON ( or always OFF ) along all trajectories . These qualitative simulations are useful to characterize the asymptotic behaviour of the system when the death receptor is engaged for a sufficiently long time . The principle of ‘ligand removal’ experiments consists in characterizing the decision process when it is subject to a temporary pulse of TNF . Here , time is intrinsically discrete , meaning that the duration of TNF pulse denoted td is represented by an integer number . In order to simulate each experiment , N trajectories were generated , starting from the “physiological” condition with TNF = 1 . At time td , the value of TNF is forced to zero . The probabilities to reach the different phenotypes are then calculated as explained in the Methods section . The average probabilities , over the N computed trajectories , are represented in Figure 6 , for the wild type and the 15 mutants . The purpose of this study is to investigate the dynamics of all the mutants and how they reach the various possible phenotypes for different lengths of TNF pulses . It provides a measurable way to assess the appearance or disappearance of certain phenotypes upon TNF induction . The curves of Figure 6 allow to link explicitly the graphs of Figure 4 when TNF is ON ( right panels ) and OFF ( left panels ) with the subjacent dynamics . Let us compare the wild type case and the deletion of cFLIP as an example of how to read these graphs . For early events , the two cases behave similarly as expected ( up to event 3 ) . As TNF pulse is prolonged , the apoptotic phenotype becomes more and more pronounced and strongly favoured over the survival one in the cFLIP mutant as opposed to the wild type conditions . This leads to the complete disappearance of survival in the mutant . This observation reinforces the role of cFLIP in the control of the apoptotic pathway . With the ‘ligand removal’ experiment , we can evaluate the number of steps , in the reduced model , that are needed for the cell to decide on its fate after TNF exposure . For almost all mutants and wild type case , the choice is made around step 4 . This means that , after this point , even if TNF is removed , the cell has already committed to a specific fate . One surprise arises from the non-monotonic behaviour of mutants for which apoptosis is suppressed ( APAF1 , BAX , caspase-8 and FADD deletions and z-VAD-fmk treatment ) , tentatively indicating a competition between components of the survival and necrotic pathways . Indeed several inhibitory cross-talks could explain this behaviour . These mutants also indicate the existence of an optimal TNF induction for which the maximum rate of necrosis is achieved ( around step 2 in the corresponding mutants of Figure 6 ) . To complete our study of cell fate decision , we reasoned on the simplest model of cell fate that can be deduced from the master model described above . The purpose here was to further simplify the network to obtain a formal representation of the logical core of the network . We have selected three components to represent the three cellular fates: NFκB for survival , MPT for necrosis and CASP3 for apoptosis . Based on reduction techniques and on the identification of all possible directed paths between these three components [50] , a three-node diagram was deduced from the master model . In this compact model , each original path ( including regulatory circuit ) is represented by an arc whose sign denotes the influence of the source node on its target . All original paths and the corresponding arcs are recorded in Table 3 . In some ambiguous cases ( e . g . influence of MPT on CASP3 or of NFκB on MPT ) , the decision on the sign of the influence is based on the Boolean rules and not on the paths only . Indeed , two negative and one positive paths link NFκB to MPT . Therefore , the sign of the arc depends not only on the states of BCL2 and of ROS , both feeding onto MPT , but also on the rule controlling MPT value . Since the absence of BCL2 and the presence of ROS ( Boolean ‘AND’ gate ) participate in the activation of MPT , if BCL2 is active , then MPT is set to 0 , even when ROS is ON . By extension , if NFκB is ON , then MPT is 0 , justifying the choice for a negative influence . In the case of mutations eliminating all the negative influences , however , a positive arrow must be considered . The resulting molecular network is symmetrical: each node is self-activating and is negatively regulated by the other nodes ( Figure 7 , upper left panel ) . This is a conceptual picture representing the general architecture of the master model that can help address specific questions . Even for this relatively simple regulatory graph , there is a finite but quite high number of possible logical rules . For now , we use a simple generic rule involving the AND and NOT operators . For example , the logical rule for CASP3 is: NOT MPT AND NOT NFkB AND CASP3 . This compact model has four stable states , each corresponding to one cell fate , along with the ‘naïve’ state ( Figure 7 , upper right panel ) . This is coherent with what was observed from the analysis of the complete model . To validate our compact model , we verified that the simulations of known mutations correspond to the published observations . Here , when a hidden component is deleted , all the paths traversing this component in the original graph are broken . If all the paths corresponding to an arc of the compact model happen to be broken , then it is removed . In the case of auto-regulation , not only the link is broken but the node is also set to zero to avoid the node to become active in the absence of death receptor activation . Let us consider the CASP8 deletion mutant to illustrate this approach ( Figure 7 , middle panels ) . For this mutant , several arrows in the compact model have to be deleted . For example , the arcs CASP3⇒CASP3 ( paths 4+5 in Table 3 ) and CASP3⊣MPT ( path 17 ) clearly depend on the activation of CASP8 . Note , however , that CASP8 intervenes in other paths , which do not fully rely on its sole activity . In the case of the arc NFκB⇒NFκB , CASP8 depletion interrupts path 3 , while path 2 can still enable the NFκB auto-regulation . Consistent with the results from the previous section , CASP8 depletion leads to the loss of the apoptotic fate while the ‘naïve’ stable state cannot be attained . At this point , one could wonder how apoptosis could be re-established in a CASP8 mutant . The analysis of the broken paths suggests some experiments to bypass CASP8 and undergo CASP3 activation . On the basis of path 4 , BAX , MOMP , SMAC and XIAP are identified as potential targets , while path 5 points to cytochrome c and apoptosome . One way to experimentally assess this possibility would be to inject exogenous cytochrome c as it was done in ‘wild type’ conditions [51] , or yet provoke its release from the mitochondria by forcing the opening of the pores . This is possible only in the absence ( or with low activity ) of NFκB and in the presence of ATP . Again , since no quantitative information can be deduced from the path analysis proposed in this study , no prediction can be made on the concentrations of proteins needed to achieve a specific answer . In a previous section , we postulated that an inhibition of the survival pathway by the necrotic pathway is necessary to reproduce some mutant phenotype . We suggested that cIAP could play this role . Let us now test this hypothesis with our conceptual model . We build the corresponding 3-node model without cIAP . In the current version , cIAP plays two important roles , first as a mandatory intermediate in the inhibitory effect of MPT ( associated with necrosis ) onto NFκB ( survival ) ( path 13 ) , next as an obligate intermediate in the self-activation of NFκB ( path 2 ) . The simulation ( Figure 7 , lower right panel ) shows that in the absence of cIAP , it is impossible to obtain the necrosis cell fate in the CASP8 ( and FADD ) mutant ( s ) , in agreement with our previous conclusions and in support of our suggestion . A complete list of all possible gene knockouts is provided in the Table S2 . This conceptual model analysis underlines the importance to simplify in order to better understand the general structure of the network and reason on it . Indeed , the simple 3-node network enables us to grasp global functional aspects and propose specific qualitative predictions .
Mathematical models provide a way to test biological hypotheses in silico . They recapitulate consistent heterogeneous published results and assemble disseminated information into a coherent picture using a coherent mathematical formalism ( discrete , continuous , stochastic , hybrid , etc . ) , depending on the questions and the available data . Then , modelling consists of constantly challenging the obtained model with available published data or experimental results ( mutants or drug treatments ) . After several refinement rounds , a model becomes particularly useful when it can provide counter-intuitive insights or suggest novel promising experiments . Here , we have conceived a mathematical model of cell fate decision , based on a logical formalisation of well-characterised molecular interactions . Former mathematical models only considered two cellular fates , apoptosis and cell survival . In contrast , we include a non-apoptotic modality of cell death , mainly necrosis , involving RIP1 , ROS and mitochondria functions . Both the master and the reduced models were constructed on the basis of an extensive analysis of the literature . The master model ( Figure 1A ) summarises our current understanding of the mechanisms regulating cell fate decision and identifies the major switches in this decision . However , some important interactions , components ( caspase-2 , calpains , AIF , etc . ) or pathways ( JNK , Akt , etc . ) have not yet been considered . This model was built to be as generic as possible . Most of the mutants considered were analysed in Jurkat cells , T-cells , or L929 murine fibrosarcoma cells , thus in very different cellular contexts ( e . g . in response to TNF , Jurkat cells are resistant to cell death , whereas L929 cell lines undergo necrosis ) . We are trying to account for all those phenotypes in a unique model . The next step will be to provide a model variant for each cell type in order to better match cell-specific behaviours . The reduced models can be used to simulate observed experiments and to reflect on the general mechanisms involved in apoptosis , survival or necrosis . This led us to identify the principal actors involved in the decision process . The presence of RIP1 or FADD , for example , proved to be decisive in our simulation . However , the role of cFLIP appears less obvious than previously suggested [7] . We can easily perturb the structure of the system in silico and assess the dynamical effects of such perturbations ( e . g . novel knockouts ) . Our model can also be used to decide between antagonist results found in different publications . For instance , the inhibitory role of cIAP1/2 on the apoptotic pathway was initially attributed to a direct inhibition of caspases . However , detailed biochemical studies challenged this view [52] , [53] . We have tested this hypothesis by adding an inhibitory arc from cIAP onto CASP8 , but simulations do not support a functional inhibitory role of cIAP1/2 , since survival is favoured over apoptosis in many mutants , thus making apoptosis a very improbable phenotype ( Figure S1 ) . Similarly , we tested the role of the feedback circuit involving CASP8 and CASP3 . We found that the activation of CASP8 by CASP3 is not functional when TNF and FASL are constantly ON . However , when TNF or FAS signal is not sustained , CASP3⇒CASP8 activation becomes necessary to insure the persistence of the apoptotic phenotype . When TNF is sustained , this feedback is no longer needed ( see Figures S2 and S3 for details ) . The in-depth analysis of model properties led us to propose several predictions or novel insights . Some concern the structure of the network , as several interactions appear to be necessary to achieve specific phenotypes . For example , our simulations of FADD and CASP8 deletion mutants underline the need for a mechanism from the necrotic pathway that would inhibit the survival one . Here , we consider a mechanism involving MPT , SMAC and cIAP . Other simulations point to different roles of proteins: RIP1 activity is transient in necrosis whereas it is sustained in survival . Similarly , our model analysis shows the role played by the duration of the TNF pulses in the cell fate decision and enlights when this decision is made . Finally , some hints about possible scenarios for forcing or restoring a phenotype in mutants are provided . Deregulations of the signalling pathways studied here can lead to drastic and serious consequences . Hanahan and Weinberg proposed that escape of apoptosis , together with other alterations of cellular physiology , represents a necessary event in cancer promotion and progression [54] . As a result , somatic mutations leading to impaired apoptosis are expected to be associated with cancer . In the cell fate model presented here , most nodes can be classified as pro-apoptotic or anti-apoptotic according to the results of “mutant” model simulations , which are correlated with experimental results found in the literature . Genes classified as pro-apoptotic in our model include caspases-8 and -3 , APAF1 as part of the apoptosome complex , cytochrome c ( Cyt_c ) , BAX , and SMAC . Anti-apoptotic genes encompass BCL2 , cIAP1/2 , XIAP , cFLIP , and different genes involved in the NFκB pathway , including NFKB1 , RELA , IKBKG and IKBKB ( not explicit in the model ) . Genetic alterations leading to loss of activity of pro-apoptotic genes or to increased activity of anti-apoptotic genes have been associated with various cancers . Thus , we can cross-list the alterations of these genes deduced from the model with what is reported in the literature and verify their role and implications in cancer . For instance , concerning pro-apoptotic genes , frameshift mutations in the ORF of the BAX gene are reported in >50% of colorectal tumours of the micro-satellite mutator phenotype [55] . Expression of CASP8 is reduced in ∼24% of tumours from patients with Ewing's sarcoma [56] . Caspase-8 was suggested in several studies to function as a tumour suppressor in neuroblastomas [57] and in lung cancer [58] . On the other hand , constitutive activation of anti-apoptotic genes is often observed in cancer cells . The most striking example is the over-expression of the BCL2 oncogene in almost all follicular lymphomas , which can result from a t ( 14;18 ) translocation that positions BCL2 in close proximity to enhancer elements of the immunoglobulin heavy-chain locus [59] . As for the survival pathway , elevated NFκB activity , resulting from different genetic alterations or expression of the v-rel viral NFκB isoform , is detected in multiple cancers , including lymphomas and breast cancers [60] . An amplification of the genomic region 11q22 that spans over the cIAP1 and cIAP2 genes is associated with lung cancers [61] , cervical cancer resistance to radiotherapy [62] , and oesophageal squamous cell carcinomas [63] . A better understanding of the pro- or anti-apoptotic roles of these genes involved in various cancers and their interactions with other pathways would set a ground for re-establishing a lost death phenotype and identifying druggable targets . The cell fate model proposed here is a first step in this direction . In the future , we will consider additional signalling cascades and their cross-talks , following the path open by other groups [64] . In parallel , we are contemplating the inclusion of other modalities of cell death such as autophagy [65] , which inhibits apoptosis through BCL2 and is itself inhibited by apoptosis through Beclin1 . The functioning of the intrinsic apoptotic pathway and the internal cellular mechanisms capable of triggering it could be investigated in more details , taking advantage of recent molecular analyses [66] , [67] . Finally , when systematic quantitative data regarding the decision between multiple cell fates will become available , our qualitative model could be used to design more quantitative models adapted to specific cellular systems in order to predict the probability for a given cell to enter into a particular cell fate depending on stimuli .
The computation of trajectories in the state space consists in the calculation of sequences of states where each member of the sequence is a logical successor of the previous one . As we choose to use Boolean variables to encode the 25-dimensional master model , the state space is the set S = {0 , 1}25 . Although finite , the size of this set is huge ( more than 33 millions states ) . Furthermore , in the discrete framework , the mathematical definition of the trajectories assumes an updating rule for the variables . Two main strategies are usually considered to analyse discrete models of biological networks . The first one consists in updating all variables simultaneously , at each time step . This synchronous strategy [68] has the advantage to generate simple determinist dynamics , each state having one and only one successor . Drawing a directed arrow from each of the 225 states to its successor , one constructs the synchronous transition graph , comprising all synchronous trajectories of the system . The determinism of the synchronous transition graph is a very strong property that poorly portrays the complexity of the biochemical processes that are modelled ( some processes are likely to occur faster than others ) . The second strategy , which is used in this paper , consists in considering that only one component is updated at each time , implying that a state may have several successors [69] . More precisely , to compute the set of asynchronous successors of a state x = ( x1 , … , xn ) ∈{0 , 1}n , one has to follow the three steps: ( 1 ) compute the state F ( x ) = ( f1 ( x ) , … , fn ( x ) ) , where fi is the Boolean rule of the ith variable ( F ( x ) is thus the synchronous successor of x ) ; ( 2 ) select the indices i such that xi≠fi ( x ) ( those are the indices of the variables that are liable to change when the system is in state x ) ; and ( 3 ) for all such indices i , the state ( x1 , … , fi ( x ) , … , xn ) is an asynchronous successor of x . According to this definition , in the asynchronous approach , no a priori hypothesis is made on the order of the events: all possible orders are considered , which is much more satisfying from a modelling point of view , as it is very difficult to know the relative speeds of the different processes involved in the master model . Note that the stable states of the model are independent on the choice of the strategy ( synchronous or asynchronous ) . Therefore , the first analysis ( based on the clustering of stable states ) is valid regardless the updating strategy . Drawing an arrow from each state to its asynchronous successors leads to the construction of an asynchronous transition graph , which comprises all possible asynchronous trajectories of the system . To each arrow starting from the same state is associated an equal probability ( see [70] for details ) . This is a strong assumption , which is the main reason why the exact values of computed probabilities ( of the different phenotypes ) should not be compared to experimental data in a quantitative manner . Nevertheless , the same assumption has been made for all model variants ( mutants and drug treatments ) , thereby allowing comparative studies . A systematic method to assess the impact of the probability distribution is a key point towards a finer quantitative analysis ( work in progress ) . As pointed earlier , the size of the transition graph is exponential with respect to the number of variables , which constitutes a first obstacle to the dynamical analysis . A second difficulty resides in the fact that the asynchronous graph is not deterministic , as each vertex may have more than one successor , which , given the size of the graph , makes the application of classical graph algorithms computationally heavy . We have used a model reduction technique specifically adapted to discrete systems , which mainly consists in iteratively “hiding” some variables , while keeping track of underlying regulatory processes [47] . The main dynamical properties of the master model , including stable states and other attractors are conserved in the reduced model . Thanks to the computation of the reduced asynchronous transition graph , relevant qualitative dynamical properties of the model can be compared to experimental results for wild type and in different mutant cases . To reduce the number of species in the master model , each logical rule is considered . For each removed component , the information contained in its rule is included in the rules of its targets such that no effective regulation is lost . Many intermediate components could easily be replaced by a proper rewriting of the logical rules associated with their target nodes . For example , IKK has only one input ( RIP1ub ) and one output ( NFκB ) . Since its role in our model merely consists in transmitting the signal from RIP1ub to NFκB , it can be easily replaced by a straightforward change in the logical rule associated with NFκB ( implementing a direct activation from RIP1ub instead of IKK ) . We also relied on the results of the clustering of stable states and their associations with biologically plausible phenotypes to select the key components to keep in the reduced model: NFκB is the principal survival actor , while caspases-3 and -8 , together with the mitochondrial membrane permeability variables ( MOMP and MPT ) , determine apoptotic and non-apoptotic cell deaths . Let us consider the example of the removal of BAX and BCL2 ( Figure 1 A and B ) . The regulators ( or inputs ) of these variables are NFκB for BCL2 and CASP8 for BAX while their regulating targets ( or outputs ) are MPT for BCL2 and MOMP for BAX . BCL2 is directly activated by NFκB , and has two targets: MPT and BAX . Therefore , BCL2 removal is performed by replacing BCL2 by NFκB into the rules of the two targets , leading to the two new logical rules: MPT′ = ROS AND NOT NFkB and BAX′ = C8 AND NOT NFkB . Applying the same process to remove BAX , one obtains the following new rule for MOMP: MOMP′ = MPT OR ( C8 AND NOT NFkB ) . The variables MOMP and MPT have now as inputs the variables NFκB and CASP8 . One can see that , in spite of the disappearance of variables BAX and BCL2 , their regulating roles are still indirectly coded in the reduced system , ensuring that no “logical interaction” of the master model ( i . e . activation or inhibition ) is actually lost during the reduction process . Table S3 lists the variables of the master model that are removed to obtain the reduced model . Some hypotheses were made when reducing the model . First , FADD is considered to be constantly ON in wild type simulations . Second , since the two complexes TNFR and DISC-TNF have been removed together with the input FADD , the two deaths ligands TNF and L have the exact same action in the reduced model . Indeed , we consider that , in response to FAS death receptor engagement as well as that of TNF; the activations of both the survival and necrotic pathways RIP1-dependent . In this case , one could then merge these variables and consider only one input that could be called “external death receptor” . However , we choose to keep the two variables TNF and FASL , in the FADD deletion mutant , the phenotype differs for TNF and FAS signal: actually , only for that mutant is the symmetry of TNF and FAS broken . | Activation of death receptors ( TNFR and Fas ) can trigger either survival or cell death according to the cell type and the cellular conditions . In other words , the same signal can have antagonist responses . On one hand , the cell can survive by activating the NFκB signalling pathway . On the other hand , it can die by apoptosis or necrosis . Apoptosis is a suicide mechanism , i . e . , an orchestrated way to disrupt cellular components and pack them into specialized vesicles that can be easily removed from the environment , whereas necrosis is a type of death that involves release of intracellular components in the surrounding tissues , possibly causing inflammatory response and severe injury . We , biologists and theoreticians , have recapitulated and integrated known biological data from the literature into an influence diagram describing the molecular events leading to each possible outcome . The diagram has been translated into a dynamical Boolean model . Simulations of wild type , mutant cells and drug treatments qualitatively match current data , and predict several novel mutant phenotypes , along with general characteristics of the cell fate decision mechanism: transient activation of some key proteins in necrosis , mutual inhibitory cross-talks between the three pathways . Our model can further be used to assess contradictory data and address specific biological questions through in silico experiments . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"cell",
"biology/cellular",
"death",
"and",
"stress",
"responses",
"computational",
"biology/signaling",
"networks",
"computational",
"biology/systems",
"biology",
"cell",
"biology/cell",
"signaling"
] | 2010 | Mathematical Modelling of Cell-Fate Decision in Response to Death Receptor Engagement |
Cellular toxicity introduced by protein misfolding threatens cell fitness and viability . Failure to eliminate these polypeptides is associated with various aggregation diseases . In eukaryotes , the ubiquitin proteasome system ( UPS ) plays a vital role in protein quality control ( PQC ) , by selectively targeting misfolded proteins for degradation . While the assembly of the proteasome can be naturally impaired by many factors , the regulatory pathways that mediate the sorting and elimination of misassembled proteasomal subunits are poorly understood . Here , we reveal how the dysfunctional proteasome is controlled by the PQC machinery . We found that among the multilayered quality control mechanisms , UPS mediated degradation of its own misassembled subunits is the favored pathway . We also demonstrated that the Hsp42 chaperone mediates an alternative pathway , the accumulation of these subunits in cytoprotective compartments . Thus , we show that proteasome homeostasis is controlled through probing the level of proteasome assembly , and the interplay between UPS mediated degradation or their sorting into distinct cellular compartments .
Protein homeostasis encompasses the systems required by the cell for generating and maintaining the correct levels , conformational states , distribution , and degradation of its proteome . Maintaining protein homeostasis is crucial to cells given the toxic potential of misfolded proteins and aggregates . Cells therefore rely on a number of protein quality control ( PQC ) pathways that survey proteins both during and after synthesis to prevent protein aggregation , promote correct protein folding , and target terminally misfolded proteins to degradation . In eukaryotes , the ubiquitin proteasome system ( UPS ) plays a vital role in PQC by selectively targeting proteins for degradation [1–4] . The eukaryotic 26S proteasome is a highly conserved 2 . 5-MD multisubunit protease responsible for degrading a large fraction of intracellular proteins . The 26S proteasome comprises a 20S core particle ( CP ) and two 19S regulatory particles ( RP ) that are further divided into lid and base complexes [5] . The degradation of most proteins is mediated by polyubiquitin chains labeling , which leads to their recognition by the 26S proteasome [6] . A diverse array of fundamental biological processes are controlled by the UPS , including cell cycle progression , DNA repair , signal transduction , and PQC in which the cell removes abnormal and toxic proteins generated as a result of environmental damage [7 , 8] . Under such conditions , chaperones are tasked with accompanying terminally misfolded and aggregated proteins to disposal , or to limit inclusion of these proteins , thereby preventing protein aggregates from causing cell toxicity and from being transferred to the next generation [2] . This chaperone mechanism , alongside the UPS , is termed spatial quality control , and consists of the juxtanuclear quality control compartment ( JUNQ ) and the insoluble protein deposit ( IPOD ) , which were identified in yeast [9 , 10] . The JUNQ provides a specialized environment for chaperone-mediated refolding or proteasomal protein degradation . Proteins that are not refolded or degraded in the JUNQ are mobilized to the IPOD . Alongside its PQC role , the UPS plays an essential role in regulating the degradation and function of nuclear proteins [11–13] . Accordingly , immune-electron microscopy experiments [14] and fluorescent microscopy of GFP-tagged proteasome subunits [15] have established that the 26S proteasome is highly enriched in the nucleus . In exponentially growing yeast cells , 80% of the 26S proteasomes are localized inside the nucleus throughout the cell cycle [16] . Given the importance of the UPS , proteasomal nuclear mislocalization may have severe consequences , for example , a deleterious effect on DNA double strand break repair [13] . Although ubiquitin-mediated proteasomal degradation of many proteins plays a key role in the PQC system , the proteasome itself can become dysfunctional as a result of transcriptional and translational failures , genomic mutations , or diverse stress conditions , leading to misfolded proteins existing in every compartment of the cell . The regulatory pathways , and the identity of the cellular machinery that mediates the sorting sequestration and elimination of dysfunctional proteasomal subunits remain poorly understood . By using a mutated proteasome lid subunit ( termed rpn5ΔC ) , we show that the nuclear mislocalization , and the cytosolic aggregates formed by this mutant represent a misassembled proteasome lid . With this experimental tool in hand , we were able to demonstrate how the dysfunctional proteasome is controlled by the PQC machinery . We found that among the multilayered quality control mechanisms , the UPS-mediated degradation of its own dysfunctional subunits is the favored pathway . However , in the absence of a functional proteasome , peripheral aggregates that represent misassembled proteasome accumulate in the IPOD , a process that is mediated by the Hsp42 chaperone . We further demonstrate that while the proteasome structure can tolerate the structural defects of the rpn5ΔC mutant and assemble into a functional proteasome , the PQC machinery takes-over , and the mutated protein is spatially removed by the PQC machinery , leading to proteasome dysfunction . Overall , our results demonstrate that proteasome homeostasis is controlled through cellular probing of the quality of proteasome aggregates , and the interplay between UPS-mediated degradation of dysfunctional subunits and alternatively , their accumulation in cytoprotective compartments .
Recently , we screened a collection of temperature sensitive ( Ts ) mutants in the yeast Saccharomyces cerevisiae for those that show a chromosomal instability ( CIN ) phenotype [17] . This screen identified proteasome subunit genes , such as the Ts allele of the regulatory particle ( RP ) subunit RPN5 , and the core particle ( CP ) subunit PUP2 . The Ts allele of RPN5 was generated by random mutagenesis [17] . Sequence analysis of this mutant revealed that a single base pair insertion resulted in a premature stop codon , leading to a 34 amino acid ( aa ) truncation at the C-terminal domain ( CTD ) of Rpn5 ( termed rpn5ΔC ) . To examine the subcellular localization of Rpn5ΔC , GFP was fused at its N-terminus . An identical N-terminal GFP fusion was constructed for the wt RPN5 gene , which was used as a control ( both fusion proteins were expressed from a galactose-inducible promoter ( GAL1 ) ) . RPN5 is an essential gene , and therefore , growth on a glucose-containing medium ( which shuts-off the expression of both GAL1-GFP-RPN5 ( GFP-RPN5 ) , and GAL1-GFP-rpn5ΔC ( GFP-rpn5ΔC ) ) resulted in cell death ( Fig 1Ai-top ) . In contrast , when the expression of the wt and the truncated gene were induced by growing the cells on 2% galactose , the growth was fully restored at the semi-permissive temperature ( 30°C ) ( Fig 1Ai-bottom ) . Furthermore , as shown in Fig 1Aii-top , similarly to the wt proteasome , the control GFP-Rpn5 protein localized predominantly to the nucleus in logarithmically growing cells . These results indicate that GAL1-GFP-RPN5 fully supports the proteasome function . In contrast , while GFP-Rpn5ΔC cells could still grow at the semi-permissive temperature ( 30°C ) , the GFP-Rpn5ΔC protein were detected in cytosolic inclusions in 81% of the cells ( n>200 ) ( Fig 1Aii , bottom , other localization patterns are shown in S2A Fig ) . Next , we wanted to determine whether the mislocalization caused by rpn5ΔC could be attributed to a failure in proteasome assembly . Previous studies have suggested a model whereby the two parts of the 26S proteasome , namely the CP and RP , are formed and imported to the nucleus independently of each other [18 , 19] . Our results are in agreement of this model and show that rpn5ΔC leads to the specific mislocalization of another proteasomal lid subunit ( Rpn11 ) ( S1Ai and S1Aii Fig ) , while the core subunits are retained in the nucleus ( S1Aiii and S1Aiv Fig ) . Similar results were obtained in a reciprocal experiment in which the Ts mutant of the CP pup2 affects the nuclear enrichment of Pre6-GFP ( another CP ) , while the RP Rpn11-GFP is unaffected ( S1B Fig ) . To further demonstrate the importance of an intact CTD for proteasome integrity , we generated diploid cells in which the original truncation of 34 aa from the CTD was extended to 45 aa ( rpn5Δ45 ) . In this case , no viable haploid rpn5Δ45 spores could be obtained following tetrad dissection ( Fig 1B ) . Moreover , a cross-species complementation experiment revealed that the expression in yeast of a full-length human homolog of RPN5 , PSMD12 , but not psmd12ΔC , was able to rescue the temperature sensitivity of the rpn5ΔC strain ( Fig 1C ) . Next , we wished to test the interaction of truncated RPN5 with other proteasomal lid subunits . To this end , we used the protein fragment complementation assay ( PCA ) [20] to examine the interaction between Rpn5ΔC and several other proteasomal lid subunits ( Rpn3 , Rpn6 , Rpn7 , Rpn8 , Rpn11 , and Rpn12 ) that were previously shown to interact with Rpn5 [7 , 21 , 22] . In this assay ( for details see S1C Fig ) the interaction between two proteins of interest can be detected by cell growth on media in the presence of the dihydrofolate reductase ( DHFR ) enzyme inhibitor , methotrexate ( MTX ) . The results show that Rpn5ΔC fails to exhibit the expected interactions when compared to the wt Rpn5 control at the semi-pemissive temperature ( Fig 1D ) . This result was supported biochemically by an immunoprecipitation experiment showing that when Rpn8 fused to a FLAG-Tag ( Rpn8-Flag ) is pulled-down , it shows a decreased interaction with Rpn5 containing a 34 aa deletion in its CTD ( Rpn5ΔC-F[3] ) ( S1D Fig ) . By generating a series of strains with defined deletions at the C-terminus of Rpn5 ( Fig 1D and 1E ) , we next showed that the interaction with other proteasomal lid subunits is impaired only when the deletion at the CTD domain is greater than 20aa . The physiological importance of these interactions is highlighted by the clear correlation between the extent of truncation and cell viability at the restrictive temperature ( 34°C ) ( Fig 1F ) . Moreover , we analyzed total protein by immunoblotting with anti Ub Abs . The results ( Fig 1G ) show a clear increase in protein ubiquitination in GFP-Rpn5ΔC cells , when compared to the GFP-Rpn5 control , at the restrictive temperature of 34°C . We therefore concluded that the presence of the truncated form of RPN5 is associated with proteasome dysfunction . Our results are in agreement with a previous study showing that in rpn5-1 cells , containing a different CTD-truncated Ts mutant of RPN5 , lid subcomplexes are not assembled , even to a partial extent , at the restrictive temperature [18] . This study also specifically examined the effect of the rpn5-1 mutation on the UPS , by evaluating the stability of three model substrates of the ubiquitin–proteasome pathway . The results demonstrated that compared with the wild-type cells , rpn5-1 cells maintained the normally short-lived substrates at a higher level at the semi-permissive temperature , indicating that the rpn5-1 mutation caused a defect in the UPS . [18] . Additional studies mapped the interaction between the subunits of the RP by electron microscopy , tandem mass-spectrometry , affinity purification analysis , and other methods [7 , 21 , 23] . These approaches demonstrate that Rpn5 , Rpn8 , Rpn9 , and Rpn11 form a stable soluble subcomplex , and the authors have proposed a subunit interaction map , supporting the notion that Rpn5 is a core component in the lid formation . Furthermore , it was also shown that in yeast , Rpn5 is independently incorporated through its CTD into the proteasomal lid [24] . These results , together with our observation that the RP subunit Rpn11-RFP co-localizes with 98% of the cells containing a large GFP-Rpn5ΔC cytosolic aggregate ( n>100 ) ( Fig 2A ) , suggest that these aggregates represent misassembled proteasome lid intermediates . Next , we tested the degree of lid assembly into 26S proteasome holocomplexes in GFP-Rpn5 , or GFP-Rpn5ΔC at the semi-permissive and restrictive temperatures using the in-gel peptidase activity assay . In this assay proteasomes are resolved by non-denaturing PAGE according to their charge/mass ratio directly from whole cell extract , and visualized based on inherent peptidase activity as described in [25] . The results ( Fig 2B ) clearly show that similarly to cells expressing the endogenous levels of Rpn5 in the wt , the over-expression of GFP-Rpn5 by a galactose inducible promoter , had no effect on proteasome integrity , as in both cases similar amounts of proteasomes were found as a mixture of RP2CP , and RP1CP . In contrast , the over production of GFP-Rpn5ΔC resulted in structural defects , as evident by higher levels of RP1CP at the semi permissive temperature ( 30°C ) , and free BaseCP and CP mostly at the restrictive temperature ( 34°C ) . Similar defects were previously reported by Yu et al , when using rpn5ΔC mutant , expressed through its endogenous promoter [24] . Taken together , these results clearly show that proteasome assembly is inhibited in rpn5ΔC cells , and that the misassembled lid in rpn5ΔC is not associated with active proteasome . Aggregation prone proteins are partitioned between the JUNQ and the IPOD [9 , 10 , 26] . Proteins that are ubiquitinated by the PQC machinery are delivered to the JUNQ where they are processed for degradation by the proteasome [9] . We therefore hypothesized that the impairment of the PQC degradation pathway by the rpn5ΔC mutant , should lead to the accumulation of misassembled proteasomal lid subunits in the IPOD . To test this idea , we followed the localization of Hsp104 , a commonly used IPOD marker [9 , 10 , 27] , fused to TFP ( Hsp104-TFP ) , in a GFP-Rpn5ΔC strain grown in rich galactose containing medium . Our results show that in all cases where GFP-Rpn5ΔC cytosolic inclusions could be detected ( 81% of the cells , n>200 , S2A Fig ) , the largest inclusion always co-localized with Hsp104-TFP ( n>200 ) ( Fig 2C ) . Similar results were obtained with the glutamine-rich prion protein Rnq1 fused to mCherry ( Rnq1-mCherry ) , another well-established IPOD marker ( Fig 2D ) [9 , 10 , 27] . Taken together , these findings suggest that peripheral foci containing misassembled Rpn5ΔC are directed to the IPOD when a functional proteasome is scarce , such as in rpn5ΔC cells . It should be noted that since the RPN5 proteasome subunit is essential for cell viability [17 , 28] , our experiments were performed mainly at the semi-permissive temperature ( 30°C ) , to enable its partial ( hypomorphic ) function . At this temperature , some proteasomes apparently function sufficiently to support cell growth . Indeed , in 19% of the cells ( n>200 ) , the GFP-Rpn5ΔC signal was limited to the nucleus , the expected localization in proliferating cells ( S2A-top Fig ) . Since the sequestration of aggregates to the IPOD was shown to preclude their delivery by the parental cells to subsequent generations [10] , the nuclear GFP-Rpn5ΔC signal , may also represent the functional proteasomes in recently separated daughter cells , as demonstrated in S1 Movie . This idea is also supported by calcofluor staining showing that the GFP-Rpn5ΔC signal was not detected as cytosolic inclusions in 98% of the daughter cells issued from IPOD containing mother cells ( Fig 2E ) . In addition to the nuclear localization , 81% of cells showed cytosolic inclusions of the GFP-Rpn5ΔC signal with the following patterns ( representative images are shown in S2A Fig , n>200 ) : 25% exhibited a single large cytosolic aggregate representing the IPOD , and 39% contained a second smaller juxtanuclear protesomal signal , likely representing the JUNQ [9] . Finally , it was recently demonstrated that under acute stress ( such as in the case of partially assembled proteasome ) , misfolded proteins are initially collected and processed in the form of multiple puncta named Q-bodies , which are rapidly reversible structures , that can be dynamically directed to folding and refolding by the chaperone machinery , or to degradation by the proteasome or autophagy system [10 , 26] . Indeed in 17% of the cells , we detected many cytosolic aggregated bodies , which probably represent these structures . One of the challenges faced by the cell in maintaining protein homeostasis is the presence of misfolded proteins . The UPS , in particular , plays a critical role in PQC by selectively targeting proteins for degradation . To test whether the UPS can also regulate the degradation of its own misassembled subunits , we introduced a functional allele of RPN5 by mating the haploid GFP-rpn5ΔC strain to another haploid containing a wt copy of RPN5 . The complementation of GFP-rpn5ΔC by RPN5 was indicated by the restoration of growth at the restrictive temperature , and the nuclear localization of the Rpn5 wt protein ( Rpn5-RFP ) ( Fig 3A and 3B ) . Similarly Rpn11-GFP also localized to the nucleus in rpn5ΔC/RPN5 cells ( S2B Fig when compared to rpn5ΔC haploids ( Fig 2A ) . Interestingly , in the presence of a functional copy of RPN5 , the GFP-Rpn5ΔC puncta was no longer detected in 97% of the cells ( n>100 ) ( Fig 3C ) . Furthermore , the addition of the proteasome inhibitor , MG132 , stabilized the GFP-Rpn5ΔC signal , and led to its reproducible accumulation in the IPOD in 68% of the cells ( n>100 ) , compared to the DMSO control ( Fig 3D ) . The accumulation of GFP-Rpn5ΔC in cells treated with MG132 was further confirmed by GAL promoter shutoff experiments followed by Western Blot analysis ( S2C Fig ) . Previous studies suggested that the ubiquitination level of a protein determines whether it is sequestered into the IPOD or JUNQ compartments [9 , 29] . Impairing misfolded protein ubiquitination blocked their accumulation in the JUNQ , and instead resulted in excessive accumulation in the IPOD [9] . The UPS-mediated degradation of Rpn5ΔC in heterozygote diploids ( containing a wt copy of RPN5 ) , suggests that Rpn5ΔC is in a ubiquitinated form , when localized to the IPOD . In order to explain this discrepancy , we propose that in rpn5ΔC haploids , the cells are under constant acute stress , due to proteasome impairment . Under such conditions , the degradation mechanism is blocked , and therefore there is no alternative , aside from targeting the misassembled lid to the IPOD . This idea is supported by previous reports that proteasome inhibition leads to the accumulation of substrates that are normally degraded by the UPS ( such as Ubc9ts ) into the JUNQ and IPOD [9 , 26] . In conjunction with this , other studies have demonstrated that when the degradation capacity of the JUNQ declines , with JUNQ-localized proteasomes becoming inactive , a protective alternative is furnished by the IPOD , and toxic aggregating species are rerouted from the JUNQ to the IPOD [30] . We therefore suggest that in our case , when cells fail to degrade misassembled proteasome lid subunits , they can be targeted to the IPOD when still conjugated to ubiquitin . Taken together , our results suggest that the preferred PQC pathway of misassembled proteasome subunits is degradation by the UPS . As functional proteasomes are scarce in rpn5ΔC cells grown at the restrictive temperature , the UPS pathway is hindered . Thus , misassembled lid subunits are ultimately diverted to the IPOD . Molecular chaperones prevent aggregation and misfolding of proteins , and are thus central to maintaining protein homeostasis . Two chaperones that were previously shown to mediate spatial sequestration of misfolded proteins are the small heat shock proteins Hsp26 and Hsp42 . These chaperones efficiently co-aggregate with misfolded proteins , thereby altering the properties of protein aggregates and facilitating disaggregation by other chaperones [31] . This process is a key molecular event that determines whether such a protein is sorted to the JUNQ or to a peripheral site [32 , 33] . As shown in Fig 4A and 4B , both Hsp26 and Hsp42 , fused to TFP , colocalized with GFP-Rpn5ΔC in 100% of the cells containing a large cytosolic aggregate , representing the IPOD ( n>200 ) . Similar results were obtained when we investigated the co-localization of Hsp42-TFP with Rpn11-GFP on a rpn5ΔC background expressed from its endogenous promoter ( Fig 4C ) . Since we have shown that Rpn5ΔC co-localizes with Rpn11 ( Fig 2A ) , and that Hsp42 co-localizes with the IPOD marker Hsp104 on a rpn5Δc background in 95% of the cells ( n>200 ) ( Fig 4D ) , we conclude that Hsp42 colocalizes with the misassembled proteasome lid in the IPOD . This co-localization is probably associated with physical interactions between Hsp42 and misassembled proteasome subunits , as indicated by the co-immunoprecipitation between the lid subunit Rpn8 , and Hsp42 in rpn5ΔC cells ( S2D Fig ) . Given the role of HSP42 , HSP26 and HSP104 in controlling the sequestration of protein aggregates into deposition sites [32–34] , we hypothesized that the sequestration of GFP-Rpn5ΔC to the IPOD may also depend on these chaperones . To investigate this possibility , we examined the localization of GFP-Rpn5ΔC in Δhsp42 , Δhsp26 and Δhsp104 cells . In contrast to Δhsp26 , and Δhsp104 which only had a minor effect on GFP-Rpn5ΔC cytosolic peripheral focus ( S2E and S2F Fig ) , in cells lacking HSP42 , the GFP-Rpn5ΔC signal was no longer observed in the cytosolic periphery . Instead , it showed the nuclear enrichment expected of the wt proteasome ( Fig 4E-bottom ) . Similar results were obtained with Rpn11-GFP in a strain mutated in both HSP42 and RPN5 ( Δhsp42 , rpn5Δc ) ( Fig 4F-bottom ) . We therefore concluded that association of misassembled proteasome lid subunits with Hsp42 is required for their accumulation in the IPOD . Strikingly , while the deletion of HSP26 , and HSP104 had no effect , the nuclear relocalization of GFP-Rpn5ΔC in Δhsp42 cells was clearly associated with increased survival at 34°C ( Fig 4G ) , and as revealed by the in gel peptidase activity assay , with the reassembly of functional proteasomes ( Fig 4H ) . Taken together , these results show that HSP42 plays an essential role in mediating the sequestration of misassembled proteasome lid subunits to the IPOD . Our data is in agreement with previous studies that mapped the interactions between subunits of the proteasome regulatory particles , and led to the notion that Rpn5 is a core component in lid formation [7 , 21 , 23] , and that in yeast , Rpn5 is independently incorporated through its CTD into the proteasomal lid [24] . Our model ( Fig 5 ) suggests that there is competition between assembly , degradation and aggregation of proteasome subunits . The employment of rpn5ΔC shifts the balance , since this mutation partially impairs proteasome lid assembly , which in contrast to wt RPN5 , triggers the activation of the PQC machinery . At the restrictive temperature , the lid is mostly misassembled , and the degradation pathway is blocked , which results in the independent recruitment of Rpn5ΔC , and other lid subunits to the IPOD in an HSP42 dependent manner . This slower assembly has a dual effect: It increases the amount of the unassembled subunits , and at the same time decreases its degradation , because there is less proteasomes available . Hence , the misassembled subunits aggregate in the IPOD . The fact that the deletion of HSP42 restores cell growth at 34°C , suggests that at this temperature , cells can still tolerate the CTD truncation in Rpn5 , and assemble partially functional proteasomes ( Fig 4G and 4H ) . A previous study supports this idea by proteasome fractionation using a glycerol gradient , showing that although the truncation influenced integration of additional subunits , Rpn5ΔC could still integrate into the proteasome at the semi-permissive temperature [8] . However , the cell stress imposed by partial misassembly of the lid in the rpn5 mutant activates the PQC machinery . This activation causes further damage , since lid subunits are independently removed to the IPOD by Hsp42 , which in turn leads to complete lid misassembly , proteasome dysfunction , and cell death at the restrictive temperature ( 34°C ) . In agreement with this model , the omission of HSP42 probably prevents the rapid sequestration of Rpn5ΔC into aggregates , allowing more time for 26S proteasomes to assemble , and to degrade the unassembled lid subunits . Thus , while it was believed that the Rpn5ΔC mutation causes purely structural defects [18] , our study provides a plausible alternative mechanism . We suggest that spatial separation of misassembled proteasome lid subunits mediated by the PQC machinery is the key pathway leading to proteasome dysfunction , rather than the structural defects within the RPN5 mutant . Taken together , our results reveal that proteins harboring mutations that activate the PQC can be eliminated from the cells , even when the protein is still functional , and the damage ensuing from diverting essential protein products . In light of this , cells have adopted numerous PQC pathways to aid folding , mediate degradation , or to accumulate such proteins in stress foci [3] . This idea is nicely demonstrated by the ΔF508 mutation within the fibrosis transmembrane conductance regulator ( CFTR ) , highly associated with cystic fibrosis ( CF ) [35] . Although the mutated protein retains significant chloride-channel function , the protein is rapidly recognized by the PQC machinery and is degraded shortly after synthesis , before it can reach its site of activity at the cell surface [36] . Although the critical role-played by the UPS in PQC , and the severe consequences of impairing this pathway are well established , little was known about the mechanisms that control dysfunctional proteasome subunits . Our results demonstrate for the first time that proteasome homeostasis is controlled through the interplay between UPS mediated degradation of its misassembled subunits , and sorting into the IPOD , a process that is mediated by the Hsp42 chaperone , which determines how proteasome homeostasis is controlled in yeast cells . The assembly of the proteasome is an intricate process due to the number of subunits that must associate to form an active complex . We used a synthetic mutant that induces proteasome dysfunction . However , the UPS function can be naturally impaired by many factors , including mutations , errors during transcription , RNA processing and translation , trapping of a folding intermediate , incorrect incorporation into multimeric complexes or oxidative damage , all of which are processes that are accelerated during aging [37] . Dysregulation of this pathway results in intracellular deposits of ubiquitin protein conjugates which can be seen in age-related pathologies and in all the major chronic neurodegenerative disorders such as Alzheimer’s , Parkinson’s and Huntington’s diseases as well as amyotrophic lateral sclerosis ( ALS ) and others [37] [38] . The mechanism of proteasome regulation by the PQC in yeast may serve as a paradigm to understand how homeostasis of this essential complex is controlled in higher eukaryotes . Identifying additional chaperones that work in conjunction with Hsp42 , and elucidating the identity of the structurally abnormal features that the PQC machinery recognizes in the misassembled proteasome , will provide further insight into the recognition and targeting mechanisms of dysfunctional proteasomes in cells .
Unless otherwise stated , all the strains used in this study are isogenic to BY4741 , BY4742 , or BY4743 [39] . The relevant genotypes are presented in S1 Table . Deletions , GFP , TFP , and mCherry fusions were generated using one step PCR mediated homologous recombination as previously described [40 , 41] . For all deletions , the selection markers replaced the coding region of the targeted genes . GFP , TFP and mCherry were fused at the 3’ end of the coding region of the targeted genes , by replacement of their stop codons [40 , 41] . A GAL1 promoter was placed at the N-terminal of RPN5 and RPN5ΔC by replacement of their start codon . We have shown that the expression of the fusion protein GAL1-GFP-RPN5 has no effect on proteasome normal phenotype ( Fig 1A ) . Strains containing different mDHFR-F[1 , 2] , and mDHFR-F[3] C-terminal fusion proteins were obtained from the PCA collection ( commercially available from Open Biosystems ) , or in cases in which the strains were absent from the collection , by one step PCR mediated homologous recombination , as described by Tarassov , K , et al . [20] . For C-terminal truncations by mDHFR-F3 ( Fig 2D and 2E ) , this fragment was targeted to replace the truncated amino acids at the 3’ end . Microscopy was performed as previously described [42] . Briefly , cells were observed in a fully automated inverted microscope ( Zeiss observer . Z1 Carl Zeiss , Inc . ) equipped with an MS-2000 stage ( Applied Scientific Instrumentation ) , a Lambda DG-4 LS 300 W xenon light source ( Sutter Instrument ) , a 63x Oil 1 . 4 NA Plan-Apochromat objective lens , and a six-position filter cube turret with a GFP filter ( excitation , BP470/40; emission , BP525/50; Beamsplitter , FT495 ) , a HcRed filter ( excitation , BP592/24; emission , BP675/100; Beamsplitter , FT615 ) , and a DAPI filter ( excitation , G365; emission , BP445/50; Beamsplitter , FT395 ) from Chroma Technology Corp . Images were acquired using a CoolSnap HQ2 camera ( Roper Scientific ) . The microscope , camera and shutters ( Uniblitz ) were controlled by AxioVision Rel . 4 . 8 . 2 . Images are a single plane of z-stacks performed using a 0 . 5 μm step . The PCA was performed as described previously [20] . Strains were mated on YPD , and diploids were selected on YPD supplemented with clonNAT and hygB . SD supplemented with noble agar ( Difco ) , and methotrexate ( MTX; Bioshop Canada ) was used for the final selection steps . Drugs were added to the following final concentrations: clonNAT ( 100 μg/ml , Werner Bioagents ) ; MTX ( 200 μg/ml ( prepared from a 10 mg/ml methotrexate in DMSO stock solution , Bioshop Canada ) ; and HygromycinB ( 100 μg/ml , Calbiochem ) . Co-immunoprecipitations and Western blot analysis were carried out as described previously [43] . The antibodies used for the Western blot analysis were anti-DHFR-[F3] ( Sigma ) , anti-FLAG ( Sigma ) , anti-Ubiquitin ( Dako Dk-z045801-2 ) , anti-GFP ( Roche 11–814460001 ) . Cultures were grown overnight and washed twice with DDW and once with chilled buffer A ( 25 mM Tris [pH 7 . 4] , 10 mM MgCl2 , 10% glycerol , 1 mM ATP , and 1 mM dithiothreitol [DTT] ) . Pellet was resuspended in two volumes of buffer A and lysed using glass beads at 4°C . Native lysates were clarified by centrifugation at 16 , 000×g for 15 min . Proteasome peptidase activity was studied in native PAGE using the substrate succinyl-LLVY-7-amido-4-methylcoumarinfluorescent peptide ( Bachem , Bubendorf , Switzerland ) as previously published [24 , 25] . | The accumulation of misfolded proteins threatens cell fitness and viability and their aggregation is commonly associated with numerous neurodegenerative disorders . Cells therefore rely on a number of protein quality control ( PQC ) pathways to prevent protein aggregation . In eukaryotes , the ubiquitin proteasome system ( UPS ) , a supramolecular machinery that mediates the proteolysis of damaged or misfolded proteins , plays a vital role in PQC by selectively targeting proteins for degradation . Although the critical role-played by the UPS in PQC , and the severe consequences of impairing this pathway are well established , little was known about the mechanisms that control dysfunctional proteasome subunits . Here , we reveal that the interplay between UPS mediated degradation of its own misassembled subunits , and sorting them into cytoprotective compartments , a process that is mediated by the Hsp42 chaperone , determines how proteasome homeostasis is controlled in yeast cells . We believe that the mechanism of proteasome regulation by the PCQ in yeast may serve as a paradigm for understanding how homeostasis of this essential complex is controlled in major chronic neurodegenerative disorders in higher eukaryotes . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | The Protein Quality Control Machinery Regulates Its Misassembled Proteasome Subunits |
Iron sequestration by host iron-binding proteins is an important mechanism of resistance to microbial infections . Inside oral epithelial cells , iron is stored within ferritin , and is therefore not usually accessible to pathogenic microbes . We observed that the ferritin concentration within oral epithelial cells was directly related to their susceptibility to damage by the human pathogenic fungus , Candida albicans . Thus , we hypothesized that host ferritin is used as an iron source by this organism . We found that C . albicans was able to grow on agar at physiological pH with ferritin as the sole source of iron , while the baker's yeast Saccharomyces cerevisiae could not . A screen of C . albicans mutants lacking components of each of the three known iron acquisition systems revealed that only the reductive pathway is involved in iron utilization from ferritin by this fungus . Additionally , C . albicans hyphae , but not yeast cells , bound ferritin , and this binding was crucial for iron acquisition from ferritin . Transcriptional profiling of wild-type and hyphal-defective C . albicans strains suggested that the C . albicans invasin-like protein Als3 is required for ferritin binding . Hyphae of an Δals3 null mutant had a strongly reduced ability to bind ferritin and these mutant cells grew poorly on agar plates with ferritin as the sole source of iron . Heterologous expression of Als3 , but not Als1 or Als5 , two closely related members of the Als protein family , allowed S . cerevisiae to bind ferritin . Immunocytochemical localization of ferritin in epithelial cells infected with C . albicans showed ferritin surrounding invading hyphae of the wild-type , but not the Δals3 mutant strain . This mutant was also unable to damage epithelial cells in vitro . Therefore , C . albicans can exploit iron from ferritin via morphology dependent binding through Als3 , suggesting that this single protein has multiple virulence attributes .
Iron is an essential element for virtually all organisms , ranging from microbes to multicellular animals . Higher organisms can sequester iron using high-affinity iron-binding molecules , so that it is unavailable to microorganisms . Iron sequestration provides a natural resistance to infections which has been described as “nutritional immunity” [1] . Successful microbial pathogens have developed multiple iron acquisition and uptake systems ( reviewed in [2] , [3] ) . These systems include enzymes for reduction and oxidization of iron ions ( Fe2+ or Fe3+ ) , high-affinity permeases for iron transport , chelators ( siderophores ) and uptake systems for siderophores . In the human body , the majority of iron is bound to iron-containing proteins with physiological functions ( for example heme proteins such as hemoglobin ) , iron-binding transport proteins ( transferrin ) , antimicrobial proteins ( lactoferrin ) , or cellular iron storage proteins ( ferritin ) . With the exception of ferritin , each of these proteins has been reported to serve as an iron source for some pathogenic microbes . These iron sources are exploited via direct binding , degradation , and/or uptake [4]–[13] . In mammalian cells , extracellular ferric iron is bound by apotransferrin ( transferrin without iron ) . The binding of apotransferrin to two ferric iron molecules ( holotransferrin , hTF ) increases by two-fold its affinity for the transferrin receptor ( TFR ) present on the surface of virtually all mammalian cells . Following endocytosis of the hTF-TFR complex into the early endosome , acidification to low pH ( pH 5 . 6 ) results in the release of iron from holotransferrin . The released ferrous iron is then transported to the cytoplasm by the divalent metal transporter ( DMT1 ) and either used for cellular metabolism or stored within ferritin . The resulting apotransferrin is recycled to the cell surface and released at physiological pH ( 7 . 4 ) [14]–[16] . Ferritin is the main intracellular storage protein for iron ( reviewed in [17] ) , containing approximately 30% of the total human body iron ( 66% is bound to hemoglobin ) . Ferritin consist of a 24-subunit protein shell of approximately 500 kDa . One ferritin molecule can contain up to 4 , 500 Fe3+ ions . The quaternary structure of ferritin is dissociated at acidic pH [18] . Its intracellular concentration can be increased by addition of exogenous iron and decreased by addition of an iron chelator [19] . Under iron-limiting conditions , cytosolic ferritin is autophagocytosed and subsequently degraded within acidic lysosomes [19] , [20] and the iron becomes available to the cell . Outside of lysosomes , ferritin is an extremely robust and stable protein which seems to be resistant to all known microbial activities . In fact , the only microorganism that has so far been shown experimentally to exploit holoferritin as an iron source during interaction with host cells is Neisseria meningitidis [21] . N . meningitidis bacteria can trigger degradation of cytosolic ferritin within infected epithelial cells by manipulating the cellular machinery and lysosomal activity [22] . To secure sufficient iron availability whilst avoiding toxicity by iron mediated processes , a tight regulation of iron metabolism is essential . Some pathogenic microbes seem to have linked the availability of iron with expression of virulence attributes . For example , the expression of virulence genes in Listeria monocytogenes was found to be positively controlled by iron limitation [23] and infections with Mycobacterium tuberculosis were reported to be more fatal when iron was accessible [24] . In pathogenic Escherichia coli strains , more than 90 genes involved in iron acquisition and several other cellular functions such as chemotaxis , respiration , DNA synthesis , glycolysis and the tricarboxylic acid cycle are co-regulated by the global iron-dependent regulator FUR ( Ferric-Uptake Regulator protein ) [2] . The fungus , Cryptococcus neoformans has recently been shown to co-regulate iron uptake mechanisms with two key virulence properties , capsule formation and melanin production [25] , [26] . Such coordinated regulation indicates that sensing the low iron content of the host environment is a key signal for pathogenic microbes to initiate adaptation to the host and express factors such as toxins and siderophores to facilitate access to host iron sources [27] , [28] . Candida albicans is a polymorphic yeast which is part of the normal microbial flora of humans . The fungus lives as a harmless commensal on mucosal surfaces in healthy individuals , but can cause several types of infections in predisposed patients , ranging from superficial to life threatening disease [29] . During infection , C . albicans can grow in almost all body sites and organs , indicating an astonishing metabolic flexibility , a high level of stress resistance and effective immune evasion strategies . One of the key features of C . albicans is its ability to grow in different morphological forms – either as ovoid yeast , a filamentous hyphal form or as pseudohyphae [30] . Although the yeast form appears to be important for dissemination [31] , the hyphal form is of crucial importance for cell and tissue invasion [32]–[34] . Furthermore , genes known or proposed to be associated with adhesion , invasion , extracellular hydrolytic activity , detoxification or as yet unknown functions ( HWP1 , ALS3 , SAP4-6 , SOD5 , HYR1 , ECE1 ) are co-regulated with the yeast-to-hyphal transition [35]–[40] . Both cellular morphology and expression of hyphal associated genes are tightly regulated by a network of signal transduction pathways ( including MAP kinase , cAMP and Rim101 pathways [30] , [41] ) and transcriptional activators and repressors such as Efg1 , Tec1 , Bcr1 , Tup1 and Nrg1 [42]–[46] . C . albicans adaptation to the host environment is also reliant on a large number of genes associated with iron acquisition [47] . These genes contribute to the three known iron acquisition systems of C . albicans: ( 1 ) Uptake and utilization of iron from hemoglobin is mediated by Rbt5 and Hmx1 [13] , [48] , [49] . In vitro data have shown that Rbt5 is a hemoglobin receptor that binds hemoglobin on the surface . This binding seems to induce expression of HMX1 , which encodes a heme oxygenase . This activity is essential for iron utilization from heme [49] . ( 2 ) Iron in siderophores is taken up via the siderophore transporter , Sit1 [50] , [51] . C . albicans siderophore production had been demonstrated by biochemical assays in earlier studies [52] , [53] . However , in contrast to Aspergillus fumigatus [54] , genes encoding factors of a possible siderophore production pathway have not been discovered in the C . albicans genome [47] . Nevertheless , Sit1 can mediate uptake of a range of heterologous siderophores from other organisms and other iron complexes [50] , [55]–[57] . ( 3 ) To use free iron from the environment , iron from transferrin , and possibly iron from other so far unknown sources , C . albicans uses the reductive uptake system . This system is located in the plasma membrane and has three components . The first component consists of ferric reductases . At least two surface ferric reductases , which are able to reduce insoluble extracellular ferric ( Fe3+ ) ions into soluble Fe2+ ions , have been described [7] , [58] , [59] . In addition thirteen homologous genes , putatively encoding other ferric reductases have been identified in the C . albicans genome ( http://www . Candidagenome . org ) . The second component consists of multicopper oxidase . Reduced ferrous iron generated by surface reductase activity can be toxic due to spontaneous generation of free radicals . However , Fe2+ can also be oxidized to Fe3+ by multicopper oxidase activity and thus preventing the production of toxic free radicals [60] , [61] . The C . albicans genome contains five putative multicopper oxidase genes [62] . Due to the copper requirement of the oxidase activity , the intracellular copper transporter Ccc2 is essential for this reductive pathway [63] . The third component consists of iron permeases . These form a protein complex with multicopper oxidases and transport Fe3+ into the cell . C . albicans has two iron permeases that are encoded by two highly homologous genes . The high-affinity iron permease gene , FTR1 is induced by iron deprivation and the low-affinity iron permease gene , FTR2 is induced when higher levels of iron are available [64] . All three iron acquisition systems appear to be independent from each other and so far only Ftr1 has been shown to be crucial for C . albicans virulence in an experimental animal model of infection [64] . However , it is unclear which iron sources are used during the different types of C . albicans infection and within different anatomical sites . Recent in vitro and in vivo transcriptional profiling experiments have shown that C . albicans gene expression is tissue specific [33] , [34] , [65] . Since the relative proportion of iron-containing proteins differs among the different anatomical sites , we propose that usage of iron by C . albicans is niche specific . Within the oral cavity , extracellular iron is bound mostly to lactoferrin in saliva and intracellular iron is stored in ferritin . However , oral infections by C . albicans are frequent , suggesting that C . albicans must be able to exploit the host iron resources of the oral cavity . We observed that genes encoding the high-affinity reductive iron uptake system of C . albicans are up-regulated during oral infections in patients [33] . Also , C . albicans causes greater damage to oral epithelial cells that contain a high concentration of ferritin ( this study ) . Therefore , we hypothesized that host ferritin may be used as an iron source by this organism . Here we show that C . albicans can utilize iron from ferritin via morphology dependent binding through the adhesin and invasin Als3 , suggesting that this single protein has multiple virulence attributes .
C . albicans can attach to and proliferate on oral epithelial tissue and can invade and damage epithelial cells [66] . To elucidate which iron sources are exploited during growth on and invasion of oral epithelial cells and to determine how the availability of iron influences fungal-host cell interactions , we incubated oral epithelial cell monolayers in the presence of additional free iron or the iron chelator bathophenanthrolindisulphonic acid ( BPS ) . This chelator sequesters extracellular , but not intracellular iron [67] . Through immunocytochemical localization of ferritin within epithelial cells , we found that addition of BPS caused a dramatic decrease in cellular ferritin within 24 hours of incubation ( Figure 1A ) , in comparison to non-treated cells ( Figure 1B ) . In contrast , addition of free iron to the medium increased the concentration of ferritin within the same time period ( Figure 1C ) . The treatment with additional iron or the iron chelator itself did not cause cell damage , as monitored by measuring the release of epithelial lactate dehydrogenase ( LDH ) into the supernatant ( not shown ) . Next , ferritin enriched or depleted epithelial monolayers were incubated for 8 h with C . albicans in normal cell culture medium ( serum-free RPMI1640 ) and cell damage caused by C . albicans was monitored by LDH release . The epithelial monolayers depleted of ferritin were significantly protected from damage in comparison to untreated monolayers ( control ) ( Figure 1D ) . In contrast , ferritin enriched epithelial cells were significantly and dose dependently more susceptible to damage caused by C . albicans ( Figure 1D ) . These observations suggested that the ferritin content of epithelial cells directly correlates with cell damage and opened up the possibility that C . albicans can use ferritin as an iron source . To clarify whether the observed increased or decreased cytotoxicity was due to either reduced or increased invasion of epithelial cells by C . albicans , we quantified invasion ( after 3 h of co-incubation ) in iron depleted versus iron saturated epithelial cells . Invasion of C . albicans into iron depleted epithelial cells was drastically reduced ( Figure 2 ) . It is known that C . albicans must invade oral epithelial cells to cause epithelial cell damage [68] . Therefore , the decreased epithelial cell invasion likely contributed to the reduced epithelial cell damage caused by iron depletion . In contrast , iron saturated epithelial cells were invaded at the same proportion as compared to untreated cells ( Figure 2 ) , even though C . albicans caused much more damage to these cells . These results suggest that the iron content of epithelial cell influences their susceptibility to damage by C . albicans , a mechanism that is at least partially independent of the extent of epithelial cell invasion . One explanation for the increased susceptibility of iron saturated epithelial cells to damage by C . albicans is that the organism uses epithelial cell ferritin as an iron source and is thereby able to produce more cytotoxic factors . To test whether C . albicans can use ferritin as an iron source in vitro , we incubated fungal cells on agar with ferritin as the sole iron source . By addition of BPS to the medium , we were able to remove any residual iron from the agar , medium or plastic surfaces . Only the addition of an external iron source allowed fungal growth under these conditions . Moreover , to minimize possible iron contamination of the ferritin solutions ( not shown ) , we passed the ferritin through a column ( Microcon YM-100 , see Material and Methods ) and washed it once with 5 mM HEPES buffer ( pH 7 . 4 ) prior to use . Addition of free ferric iron , hemoglobin or ferritin to the agar promoted the growth of C . albicans at pH 7 . 4 ( Figure 3A ) . In contrast , the baker's yeast Saccharomyces cerevisiae , known to be unable to grow with hemoglobin as the sole source of iron [13] , only grew after addition of free iron to the medium ( Figure 3A ) . However , S . cerevisiae was able to grow with ferritin when the initial pH of the medium was calibrated to pH 5 . 0 ( not shown ) . This result suggested that the external pH of the medium influenced the bio-availability of iron from ferritin . The ferritin protein shell is unstable at acidic pH [18] . Therefore , our finding that S . cerevisiae can utilize iron from ferritin at acidic , but not alkaline pH , suggested the possibility that C . albicans is able to release iron from this protein by active acidification of the medium . In fact , C . albicans was able to acidify a medium buffered with 25 mM HEPES ( initial pH 7 . 4 ) during incubation with ferritin as sole source of iron as monitored by the pH indicator bromocresol green ( Figure S1 ) . To determine whether ferritin utilization is dependent on fungal-driven acidification , we substituted the glucose in the medium for casamino acids . This mixture of amino acids can be used as a carbon source by yeasts and avoids the acidification associated with glucose use [69] , [70] . Furthermore , we stabilized the buffering capacity of the medium by the addition of HEPES buffer ( pH 7 . 4 ) with increasing concentrations . As shown in Figure 3B , decreasing the capacity to acidify the medium , reduced the ability of C . albicans to grow with ferritin as a sole source of iron . Next , we sought to determine which of the three known iron uptake systems of C . albicans are involved in iron acquisition from ferritin . Mutants lacking key genes of each iron acquisition system were screened for growth on ferritin agar plates . A mutant lacking the high-affinity permease Ftr1 was able to grow with free iron , hemoglobin , but not with ferritin as the sole iron source ( Figure 3A; Table 1 ) . Similarly , a mutant , lacking the copper transporter Ccc2 , which is also essential for the reductive pathway , did not grow on ferritin plates ( Figure S2; Table 1 ) . In contrast to S . cerevisiae , Δftr1 and Δccc2 mutants did not grow on ferritin plates even when the initial pH was reduced to 5 . 0 ( not shown ) . As expected , the Δftr1+FTR1 and Δccc2+CCC2 re-integrant strains grew similarly to the wild-type strain in the presence of ferritin ( Figure S2 ) . These observations suggest that the reductive pathway is essential for C . albicans to acquire iron from ferritin . The Δsit1 and Δrbt5 mutants grew normally when ferritin was the sole iron source , indicating that C . albicans utilization of iron from ferritin is independent of the siderophore and hemoglobin uptake systems ( Figure S2; Table 1 ) . We also investigated the possibility that aspartic proteases secreted by C . albicans could break down ferritin and release iron . The Δsap1-3 and Δsap4-6 triple-mutants grew similarly to wild-type cells on ferritin plates , suggesting that secreted aspartic proteases of C . albicans are not involved in liberating iron from ferritin ( Figure S2; Table 1 ) . Pathogenic microbes frequently utilize iron from host proteins by binding these molecules via specific receptors [4]–[8] , [12] , [13] , [71] . Since our data showed that C . albicans can use ferritin as a sole source of iron , we investigated whether C . albicans can bind ferritin on its surface . C . albicans cells precultured in iron limited medium ( LIM0 ) were co-incubated with ferritin and then rinsed extensively . The ferritin that remained bound to the organisms was subsequently visualized with fluorescent labeled anti-ferritin antibodies . Hyphae of wild-type C . albicans bound ferritin whereas yeast-phase organisms did not ( Figures 4A and 4B ) . The binding of ferritin to hyphae was also visualized by electron microscopy . Due to their high-electron density , ferritin molecules appeared as black particles in the electron micrograph adjacent to the fungal cell wall , indicating that ferritin bound to the cell surface , and not within the fungal cell wall ( Figure 4C ) . C . albicans cells incubated under the same condition , but without ferritin , had no such electron dense particles on their surfaces ( not shown ) . The finding that ferritin was not bound by either yeast cells or the mother cells of hyphae suggested that ferritin binding was specific to C . albicans hyphae . To test this hypothesis further , we investigated the ferritin binding of C . albicans Δras1 and Δcph1/efg1 mutants that were unable to form hyphae , and did not express hyphal-specific genes [72] , [73] . Both mutants were unable to bind ferritin ( Figure 4A and 4B ) . Next , we tested the ferritin binding capacity of a Δhgc1 mutant , which forms pseudohyphae rather than true hyphae , but still expresses hyphal-specific genes [74] . When grown under hyphal-inducing conditions ( RPMI 1640 , 37°C with 5% CO2 ) , the Δhgc1 mutant bound ferritin even though it did not form true hyphae ( Figure 4A and 4B ) . These results suggest that the product of one or more hyphal specific genes is essential for C . albicans to bind ferritin . To uncover which hyphal-associated activities are involved in ferritin binding , wild-type hyphae were killed with thimerosal or UV light and tested for ferritin binding . Cells killed with thimerosal still bound ferritin; whereas cells killed with UV light did not ( Figure S3A ) . When untreated wild-type cells were mixed with 50% UV-killed cells , we observed 49 . 06%±4 . 27% ferritin binding . These data demonstrate that cell viability is not necessary for ferritin binding and that the ferritin receptor on the cell surface is inactivated by UV treatment . We also investigated whether iron availability influenced the extent of ferritin binding of wild-type hyphae . Cells grown under iron limiting conditions or in the presence of excess iron bound ferritin similarly ( Figure S3B ) . Also , C . albicans hyphae were able to bind ferritin and apoferritin ( a ferritin shell without iron ) with similar efficiency ( not shown ) indicating that iron molecules within the ferritin shell were dispensable for binding of ferritin . Thus , these data indicated that the binding of ferritin by C . albicans is morphology associated , but not iron-regulated . Transcriptional profiling was used to identify genes encoding putative ferritin receptors . We incubated a wild-type strain ( true hyphae and ferritin binding ) , the Δhgc1 mutant ( yeast or pseudohyphae and ferritin binding ) and Δras1 ( no hyphae and no ferritin binding ) under hyphal-inducing conditions ( RPMI medium , 37°C with 5% CO2 ) and in the presence of ferritin . After 1 . 5 h , the RNA from all three strains was isolated , labeled and hybridized to C . albicans microarrays . Microarray data from four independent experiments were analyzed . We reasoned that candidate genes encoding putative ferritin receptors should be up-regulated in wild-type and Δhgc1 cells , but should be unaltered or down-regulated in the Δras1 mutant ( Figure 5 ) . A total of 22 genes were identified with such an expression profile ( Figure 5 ) . Expression data shown in Table 2 indicate the genes that were up-regulated in wild-type cells but not in Δras1 mutant cells . Three of these genes were known to encode hyphal-specific proteins that are cell surface localized as would be expected for a receptor protein . Consequently , these three genes were further investigated . The three genes encoding cell surface localized and hyphal-specific proteins were ECE1 , HYR1 and ALS3 . ECE1 ( Extent of Cell Elongation ) is a hyphal-specific gene with yet unknown functions . ECE1 expression increases during elongation of the hyphal cell . This gene encodes a predicted cell membrane protein and the corresponding null mutant displays no obvious altered phenotype [40] . HYR1 ( HYphally Regulated ) encodes a GPI-anchored protein that is predicted to be cell wall localized and is of unknown function [39] . Finally , ALS3 ( Agglutinin-Like Sequence ) encodes a hyphal-specific cell wall protein which belongs to a family of adhesins ( Als family ) [75] and plays a crucial role in epithelial and endothelial adhesion and invasion [32] . The corresponding homozygous null mutants were tested for ferritin binding . Both , the Δece1 and the Δhyr1 mutants bound ferritin similarly to the wild-type strain ( Figure 6A and B ) . In contrast , ferritin binding of the Δals3 mutant was dramatically reduced ( Figures 6 and 7 ) . This defect in ferritin binding was restored when a wild-type copy of ALS3 was reintegrated into the Δals3 mutant ( Figures 6 and 7 ) . These results suggested that Als3 plays a crucial role in ferritin binding and may in fact be the hyphal-specific ferritin receptor . If Als3 is a ferritin receptor , one would expect that mutants lacking factors that govern ALS3 expression would also have an altered capacity to bind ferritin . Therefore , we tested two mutants that lacked ALS3 transcriptional regulators . BCR1 encodes a transcription factor which regulates the expression of certain hyphal-specific genes , including ALS3 [76] . Furthermore , expression of BCR1 itself depends on Tec1 [44] . Figure 6C shows that the presence of both transcriptional factors , Tec1 and Bcr1 , is necessary for C . albicans cells to bind ferritin . These data reinforce the view that Als3 plays a key role in the capacity of C . albicans to bind ferritin . To determine whether ferritin binding is necessary for the utilization of iron from this protein , we tested the growth of the Δals3 mutant with ferritin as the sole iron source . The Δals3 mutant grew very poorly on agar plates ( pH 7 . 4 ) with ferritin as the sole source of iron ( Figure 8 ) . The reconstitution of one copy of the gene ( Δals3+ALS3 re-integrant strain ) , improved growth , although not to wild-type levels ( Figure 8 ) . Growth of the Δals3 mutant in media with low iron content was not reduced , indicating that uptake of free iron is normal in the Δals3 mutant ( not shown ) . Therefore , Als3 is required for C . albicans hyphae to both bind and utilize ferritin as a source of iron . Moreover , a mutant unable to form hyphae ( Δras1 ) and thus unable to bind ferritin was also tested for growth on ferritin plates . As expected , Δras1 displayed a reduced ability to grow with ferritin as the sole source of iron ( Figure S4 ) . This result reforces the key role of hyphal development and the hyphal associated expression of ALS3 in the ability of C . albicans to obtain iron from ferritin . To elucidate whether Als3 itself can bind ferritin without an additional C . albicans surface factor , we tested the ferritin binding capacity of a strain of S . cerevisiae that expressed C . albicans ALS3 [77] . Because ALS3 is a member of a large gene family encoding similar proteins , we also analyzed two additional S . cerevisiae strains that expressed ALS1 or ALS5 , two closely related ALS genes . The strain that expressed ALS3 strongly bound ferritin , whereas the strains that expressed ALS1 or ALS5 did not ( Figure 9 ) . Next we investigated whether ferritin binding via Als3 occurs when C . albicans interacts with host cells . Oral epithelial cells were loaded with iron and then incubated with wild-type C . albicans , the Δals3 mutant , or the Δals3+ALS3 re-integrant strain for 6 h . To visualize ferritin molecules on cellular surfaces and to investigate whether the location of fungal cells had an influence on ferritin binding , we used an immunofluorescence approach with differential staining , which enabled us to discriminate between hyphae located on the epithelial cell surface and hyphae that had invaded into the epithelial cells ( Figure 10 columns 1 , 2 and 4 ) . In addition , we used an anti-ferritin antibody to localize ferritin ( Figure 10 column 3 ) . As shown in Figure 10 , hyphae of wild-type and Δals3+ALS3 re-integrant strains invaded the epithelial cells and were surrounded by ferritin ( white arrows in Figure 10 ) . Very little ferritin accumulated around wild-type hyphae that had not invaded the epithelial cells ( data not shown ) . In contrast , the few hyphae of the Δals3 mutant that invaded the epithelial cells displayed no accumulation of ferritin ( Figure 10G and K ) . These results indicate that C . albicans hyphae bind to ferritin in an Als3-dependent manner while invading epithelial cells . If binding to ferritin and utilizing host iron are important for C . albicans to cause an oral infection , one would expect that mutants lacking ALS3 or FTR1 would have a reduced potential to cause tissue damage as compared to wild-type cells . To test this prediction , we measured the extent of epithelial cell damage caused by wild-type , Δals3 mutant and Δftr1 mutant strains of C . albicans . We found that the Δals3 and Δftr1 mutants lost their capacity to damage epithelial cells ( Figure 11 ) . In contrast to Δals3 mutant cells , which displayed strongly reduced invasion abilities when co-incubated with epithelial cells for 3 hours ( not shown ) , Δftr1 mutant cells showed the same invasion rate than the wild-type strain ( Figure S5 ) . Although hyphae of this mutant seemed shorter than the wild-type hyphae , there was no morphological differences between Δftr1 mutant cells on epithelial cells and in RPMI medium alone ( control ) ( not shown ) .
Iron availability is a critical factor for all pathogenic microbes and iron excess can accelerate pathogenicity [1] , [78]–[80] . We observed that oral epithelial cells enriched in intracellular ferritin were more susceptible to tissue damage by wild-type C . albicans and that epithelial cells depleted of ferritin were significantly protected from damage . The reduced damage of iron depleted epithelial cells correlated with reduced invasion of C . albicans . It is possible that the treatment with the iron chelator affected both the host cells and the pathogen . Iron depleted epithelial cells may have a reduced ability to internalize fungal cells and limited accessibility of iron may reduce the capacity of C . albicans to both invade and damage epithelial cells . This model is supported by previous data . For example , endothelial cells incubated with an iron chelator before C . albicans infection were protected from injury by C . albicans [81] and the anti Candida activity of ciclopiroxolamine , a potent antifungal agent , is proposed to be mediated by iron chelation [82]–[84] . In contrast , when epithelial cells were loaded with exogenous iron , epithelial cell uptake of C . albicans was not affected . However , the increased iron reservoir was likely exploited by C . albicans , leading to increased epithelial cell damage . This increased damage was probably due to an enhanced production of virulence determinants ( e . g . hydrolases ) and hyphal extention . Therefore , it can be concluded that access to iron has a direct influence on the pathogenicity of C . albicans , probably by acting on both the host and the fungus . Furthermore , our data suggest that C . albicans is able to directly use ferritin as a source of iron . It is known that ferritin is an extremely robust and resistant protein . Prior to this study , the only microorganism that has been known to exploit holoferritin as an iron source during interaction with host cells is N . meningitidis [21] , [22] . However , this bacterium is not able to directly utilize iron from ferritin . Instead , it induces degradation of cytosolic ferritin by manipulating the host cellular machinery and thereby utilizes the resultant free cytosolic iron . To our knowledge , no published studies have so far demonstrated direct use of iron from host ferritin . Nevertheless , a number of studies have suggested that certain microbial pathogens can use ferritin as an iron source during in vitro growth . For example , Yersinia pestis can grow on agar containing hemin , myoglobin , hemoglobin or ferritin [85] . A siderophore produced by M . tuberculosis ( exochelin ) can sequester iron from transferrin , lactoferrin and to a lesser extent from ferritin [86] . L . monocytogenes and Burkholderia cenocepacia can grow in liquid medium with ferritin as the sole source of iron [87] , [88] . However , the microbial mechanisms of iron acquisition from ferritin are unknown and it is not clear whether ferritin from host cells can be used by any of these species . Furthermore , although ferritin seems to be almost indestructible under physiological conditions , iron may be released from ferritin in vitro , especially under condition of low pH . In our hands , even S . cerevisiae was able to utilize iron from ferritin under such conditions . Therefore , it is possible that previous observations of the microbial usage of ferritin in vitro were the results of non-physiological conditions . In contrast to S . cerevisiae , C . albicans can use ferritin as the sole source of iron in vitro even when the growth medium was buffered at a physiological pH . Which mechanisms and activities are involved in iron acquisition from ferritin ? One possibility is that ferritin is degraded by extracellular proteolytic activity since it is known that C . albicans can secrete a family of aspartic proteases ( Saps ) with very broad substrate specificity [89] . However , it appears that extracellular degradation due to fungal proteases is not necessary for growth with ferritin , since mutants lacking the protease genes SAP1-3 or SAP4-6 were still able to grow on such medium . Indeed , an earlier study by Rüchel demonstrated that ferritin was the only tested protein which was resistant to proteolysis by Sap2 , one of the major secreted proteases of C . albicans with an extremely broad substrate specificity [90] , supporting the view that proteases are not involved in the ability of C . albicans to utilize iron from ferritin . Since even S . cerevisiae was able to grow with ferritin when the pH of the medium was low ( pH 5 . 0 ) , we reasoned that the pH plays a crucial role in the release of iron from ferritin . It is known that ferritin is unstable at acidic pH [18] and that the natural recycling of iron from ferritin occurs in the acidic environment of lysosomes [19] , [20] . Thus , it may be possible that C . albicans actively lowers the pH in its proximate vicinity . In fact , C . albicans was able to lower the pH of the medium during growth even on buffered ferritin plates ( Figure S2 ) . Additionally , the fungus was only able to use ferritin as an iron source under conditions which allowed acid production ( glucose , but not casamino acids as a carbon source ) and acidification of the surrounding environment ( low concentrations of buffer at pH 7 . 4 ) . Similarly , it has been observed that the bacterial pathogen Staphylococcus aureus , under iron starvation , decreases the local pH resulting in the release of iron from transferrin [91] . It is also possible that C . albicans can produce and secrete reductants , which are able to sequester iron from ferritin . Such a process would indeed be favoured by acidification of the surrounding media . In agreement with this model , reductants or chelators such as thioglycolic acid , ascorbate , and aceto- and benzohydroxamic acids are capable of releasing iron from the ferritin core [92]–[94] . Underscoring the importance of pH in the release of iron from ferritin , this process is increased at pH 5 . 2 in comparison to pH 7 . 4 [94] . However , since we demonstrated that binding is necessary for ferritin iron exploitation by C . albicans , it can be hypothesized that a surface factor rather than a secreted factor is necessary for ferric iron reduction from the ferritin core . Another possible speculation is that reductases on the C . albicans cell surface can reduce ferric iron from the ferritin core and that this process may be facilitate under acidic pH . Although we do not have experimental evidence that local acidification occurs in vivo during infection , transcriptional profiling of C . albicans during experimental infections suggests that the local environment of at least some cells in fact changes from neutral to acidic pH during invasion and tissue damage . For example , we have found that the acid induced gene , PHR2 is up-regulated during tissue invasion [34] . In addition to the ability to acidify the environment , C . albicans requires the reductive high-affinity iron uptake pathway to exploit iron from ferritin . Mutants lacking either the high-affinity permease Ftr1 or the copper transporter Ccc2 ( which is essential for the reductive pathway ) [63] , [64] did not grow on ferritin plates even when the initial pH was low . Therefore , we conclude that a combination of active acidification and uptake via the high-affinity permease are key mechanisms in this process . As a third prerequisite , we hypothesized , that a close association between C . albicans cells and ferritin is required for the release of iron from ferritin and subsequent uptake into the fungal cell . This close contact is facilitated by binding of ferritin on the fungal surface . In principle it may also be postulated that a yet unknown molecule is secreted by C . albicans , which binds ferritin and subsequently delivers the iron protein to a surface receptor , similar to some bacteria which can secrete haemophores that bind extracellular haemoglobin and mediate its delivery to surface receptors [95] . However , such a mechanism is unlikely to be involved in ferritin-binding by C . albicans since fungal cells that were killed with thimerosal and then washed , removing any secreted factors , were still able to bind ferritin . Interestingly , fungal cells killed via exposure to UV-light lost their ability to bind ferritin . This result suggests that ferritin-binding at the cell surface is mediated by a receptor which is inactivated by UV treatment . In support of this possibility , it is known that certain proteins can be inactivated by exposure to UV light [96] . Several lines of evidence suggest that the cell surface protein , Als3 is a receptor that binds ferritin and facilitates iron acquisition from this protein . ( 1 ) Only hyphae , but not yeast cells bound ferritin and Als3 is known to be a hyphal-specific protein . However , the binding of ferritin did not need the hyphal morphology , since a mutant lacking Hgc1 [74] did not produce true hyphae , but still bound ferritin ( Figures 4A and 4B ) and expressed ALS3 ( not shown ) . ( 2 ) Mutants lacking transcription factors known to regulate ALS3 expression ( Tec1 , Bcr1 ) [44] , [45] had a reduced ability to bind ferritin . In agreement with this , a mutant that was unable to form hyphae ( Δras1 ) and that did not express ALS3 , also displayed reduced binding of ferritin and reduced growth on ferritin plates . ( 3 ) A mutant lacking ALS3 was dramatically reduced in its ability to bind ferritin and displayed poor growth on ferritin plates . The Δals3+ALS3 re-integrant strain had a restored ability to bind ferritin and a partially restored ability to grow on ferritin plates , although not to wild-type levels , possibly due to a gene dosage effect . Finally , ( 4 ) a S . cerevisiae strain expressing Als3 was able to bind ferritin . Binding of ferritin to hyphal surfaces was observed with both exogenously added purified ferritin and during the interaction of C . albicans with intact epithelial cells . Only hyphae , but not yeast cells showed bound ferritin during interaction with epithelial cells . Furthermore , ferritin accumulation was predominantly observed on those hyphae that had invaded the epithelial cells . Finally , the hyphae of the Δals3 mutant did not show ferritin accumulation . Taken together , these data suggest that ferritin can be used as an iron source by C . albicans via direct binding by Als3 on the surface of hyphae , iron release is then mediated by acidification and uptake is facilitated by the reductive pathway ( Figure 12 ) . Although we do not have direct evidence that ferritin is in fact used as an iron source during interaction with epithelial cells , these data at least suggest that ferritin is in close contact to invading C . albicans hyphae and thus may be exploited by the above proposed mechanism . This view is supported by the fact that both the Δals3 mutant and the Δftr1 mutant completely lost their capacities to damage epithelial cells in vitro . Furthermore , the Δals3 mutant has significantly reduced capacity to damage epithelial cells in the reconstituted human epithelium model [97] . However , it should be noted that this reduced damage is likely due to a combination of reduced adherence [97] , reduced invasion [32] , and reduced ability to use ferritin as an iron source . Interestingly , hyphae of the Δftr1 mutant displayed the same invasion rate than the wild-type strain , suggesting that this mutant can initially invade the epithelial cells , but is not able to damage host cells possibly because it can not use ferritin as an intracellular available iron source . Several studies have shown that pathogenic microbes link the availability of iron with virulence attributes . In this study , we show that a similar link between the regulation of an iron acquisition system and virulence attributes exists in C . albicans . In fact , the regulation of the ferritin receptor Als3 is independent from external iron sources and seems to be strictly linked to hyphal formation , one of the most extensively investigated virulence attributes of C . albicans [30] , [98] . Therefore , iron acquisition of the intracellular iron storage protein ferritin is hyphal regulated . Hyphal formation is also associated with adhesion , proteolytic activity , cellular invasion and damage [32]–[34] , [89] , [99] , and the hyphal form of the organism is the predominant morphology that reaches the intracellular compartments of epithelial cells where ferritin is located . Therefore , C . albicans co-regulates morphology , invasion , tissue damage and an iron acquisition system . This view may explain why iron acquisition from ferritin is a hyphal-specific property and does not occur with the normally non-invasive yeast cells . A second potential link exists between the external pH , hyphal formation and iron acquisition . It is well known that the external pH influences hyphal formation [100] , [101] and we recently reported that pH-dependent hyphal formation is crucial for liver invasion [34] . During liver invasion C . albicans cells are exposed to a neutral or alkaline pH and iron limited conditions as reflected by transcriptional profiles [34] . Availability of iron for fungal cells within a human host is even more difficult in neutral or slight alkaline pH conditions such as those found in the liver tissue ( pH 7 . 4 ) because the balance between the soluble Fe2+ ion and the insoluble ferric form Fe3+ shifts towards the insoluble form [41] . Therefore , the formation of hyphae and expression of Als3 in response to neutral pH may facilitate iron acquisition by C . albicans . Interestingly , the expression of ALS3 is not absolutely linked to the hyphal morphology in wild-type cells . Sosinska et al . [102] recently observed that hypoxic conditions and iron restriction in a vagina-simulative medium affected cell morphology and the cell wall proteome of C . albicans . One of the proteins found in yeast cells under these iron limited conditions was Als3 , which indicates that even proteins which are strictly hyphal-associated under most growth conditions , may be expressed in the yeast form . Similarly , White and co-workers recently showed that C . albicans expresses a number of hyphal-specific genes ( such as ECE1 ) in a murine gut model of commensalism , whilst growing in the yeast morphology [103] . The observation that yeast cells express Als3 under iron limited conditions may further support the view that this protein is involved in iron acquisition from the host . However , the expression of ALS3 is not directly linked to low iron conditions since two studies that analyzed the influence of iron on the genome wide gene expression of C . albicans [47] , [83] found that iron starvation did not increase the expression of ALS3 . The Als protein family of C . albicans encodes large cell-surface GPI-glycoproteins that were originally implicated in the process of adhesion to host surfaces [75] , [104] . Expression of Als3 , was shown to be hyphal-specific [36] and was observed in vivo during oral and systemic infection [33] , [34] . In addition to its adhesion properties , Als3 was recently shown to be an invasin that binds to cadherins and induces endocytosis by host cells [32] . In this study , we made the intriguing observation that Als3 has a third function in iron acquisition by binding to host ferritin , indicating that this single member of a protein family has multiple virulence-associated functions .
C . albicans were grown in liquid YPD medium ( 1% yeast extract [Merck , http://www . merck . de] , 2% bactopeptone [Difco , http://www . bdbiosciences . com] , and 2% D-glucose [Roth , http://www . carl-roth . de] ) in a shaking incubator at 30°C for 8 h . Subsequently , the cultures were diluted 1∶1000 in LIM0 medium [105] and incubated in a shaking incubator at 30°C overnight for iron starvation . For non-starved cells , precultures were incubated in YPD medium overnight at 30°C with shaking . The yeast cells were harvested by centrifugation , washed three times in filter sterilized ultra-pure water and counted using a hemacytometer . Strains of C . albicans and S . cerevisiae used in this study are listed in Table 3 and Table 4 , respectively . The epithelial cell line TR146 , derived from a squamous cell carcinoma of buccal mucosa [106] , was kindly provided by Cancer Research Technology ( http://www . cancertechnology . co . uk ) . TR146 cells were routinely grown in RPMI 1640 medium ( PAA , http://www . paa . com ) supplemented with 10% fetal bovine serum ( FBS; PAA ) . For experiments , epithelial cells were used between passages 10 to 20 . Monolayers with 70–90% confluent cells in 24 well plates were additionally incubated for 24 h in three different conditions: ( 1 ) RPMI 1640 with 50 µM bathophenanthrolinedisulfonic acid disodium salt ( BPS; iron chelator; Sigma-Aldrich , http://www . sigmaaldrich . com ) ; ( 2 ) RPMI 1640 with 10% FBS; ( 3 ) RPMI 1640 with 10% FBS and indicated concentrations of iron chloride ( FeCl3; Merck ) . After 24 h incubation , monolayers were washed twice with phosphate-buffered saline without calcium or magnesium ( PBS ) and serum-free RPMI 1640 medium was added . Each well was infected with ∼106 C . albicans cells and incubated for 8 h . Supernatants were removed for LDH measurements . All incubations were performed in a humidified incubator at 37°C in 5% CO2 . To monitor the ferritin content of cells , the uninfected monolayers were fixed with Roti®-Histofix 4% ( Roth ) and the ferritin content of the cells was visualized under the microscope using immunofluorescence . Briefly , fixed monolayers were permeabilized through incubation with 0 . 1% Triton X-100 ( Serva , http://www . serva . de ) for 15 min at room temperature and washed three times with PBS . Next , the samples were blocked using Image-iT™ FX signal enhancer ( Invitrogen , http://probes . invitrogen . com/products/ ) for 30 min at room temperature in a humidity chamber . Cells were again washed three times with PBS and incubated with rabbit anti-ferritin antibody ( Sigma-Aldrich ) coupled with dye DY-649 ( Dyomics , http://www . dyomics . com ) diluted 1∶1000 in PBS with 1% bovine serum albumin ( BSA , Sigma-Aldrich ) for 1 h at room temperature . Finally , cover-slips were washed three times with PBS , inverted and mounted on a microscope slide with ProLong® Gold Antifade Reagent with 4′ , 6-diamidino-2-phenylindole dihydrochloride ( DAPI ) ( Invitrogen ) . The samples were analyzed in duplicates using a Leica DM 5500B microscope ( Leica , http://www . leica-microsystems . com ) . The same exposure time and light intensity were used to analyze all samples , permiting comparisons . For every sample , 10 randomly chosen fields per cover-slip were photographed using a DFC 350 FX camera ( Leica ) . A representative picture of each condition was selected . Epithelial cell damage caused by C . albicans , was determined by the release of lactate dehydrogenase ( LDH ) into the medium using a Cytotoxicity Detection Kit–LDH ( Roche , http://www . roche . de ) . The assays were performed according to the manufacturer instructions and the measurements were performed in duplicates . To measure epithelial cell damage , the following calculation was used: 100× ( ECa−C1−C2 ) / ( 100L−C1 ) = relative cytotoxicity ( % ) . Absorbance measured at OD 490–600 directly correlates with LDH activity . ECa = epithelial cells infected with C . albicans; C1 = control 1–uninfected epithelial cells; C2 = control 2–only C . albicans; 100L = 100% lysis ( 0 . 2% Triton-X 100 , Serva ) . Controls 1 , 2 and 100% lysis were determined individually for each treatment . To investigate whether C . albicans was able to grow with ferritin as the sole source of iron , we added 350 µM BPS to the SD agar medium ( 6 . 7 g/l yeast nitrogen base , YNB [Difco]; 20 g/l D-glucose; 20 g/l purified agar [Oxoid , http://www . oxoid . com] ) . Additionally , HEPES buffer ( Sigma-Aldrich ) was added to the medium as indicated and the pH was adjust to 7 . 4 using a 5 M NaOH stock solution ( Roth ) . To prevent active acidification of the medium by the fungus , 20 g/l casamino acids ( Difco ) was used in place of D-glucose . The ferritin solution ( ferritin from horse spleen [Sigma-Aldrich] ) was diluted 1∶100 in a dilution buffer ( 5 mM HEPES; 0 , 1 M NaCl [Roth] ) and passed through a Microcon YM-100 Centrifugal Filter Unit ( Millipore , http://www . millipore . com ) . The retentate was collected in a fresh 1 . 5 ml microcentrifuge-tube and the original volume was adjusted with the dilution buffer . Afterwards , this ferritin solution was plated out on agar surfaces at indicated concentrations . To monitor the pH changes in the medium during C . albicans growth , the pH indicator bromocresol green ( Sigma-Aldrich ) was added to the medium at a concentration of 3 . 9 mg/l . C . albicans cells growing under iron limitation , as described above , were washed and enumerated . Approximately 5×105 cells were added per well in a 24 well plate ( TPP , http://www . tissue-cell-culture . com ) containing Poly-L-Lysine-coated ( Biochrom AG , http://www . biochrom . de ) 12-mm diameter glass cover-slips and 1 ml RPMI 1640 . The cells were incubated for 3 h at 37°C under 5% CO2 to induce hyphae . Afterwards , the cells were washed once with PBS and incubated for 1 h in 1 ml PBS with 1% bovine serum albumin ( BSA ) and 100 µg/ml ferritin . Subsequently , the cells were washed three times with PBS to remove non-bound ferritin and fixed with 500 µl Roti®-Histofix 4% . To test if viability is necessary for ferritin binding , C . albicans hyphae ( 3 h in RPMI 1640 at 37°C and 5% CO2 ) were killed using two different approaches: either 1 . 5 h incubation at room temperature with 0 . 05% Thimerosal ( Sigma-Aldrich ) or 2 times exposition to 0 . 5 J/cm2 UV light in a UV-crosslinker with a 254 nm low pressure mercury-vapor lamp ( Vilber-Loumart , http://www . vilber . de ) . Complete killing without residual viability of cells was checked by plating the cells on YPD agar plates . After killing , the cells were incubated with ferritin and fixed as described above . The fixed cells were washed three times with PBS and incubated with rabbit anti-ferritin antibody coupled with dye DY-649 diluted 1∶2000 in PBS with 1% BSA for 1 h at room temperature . Next , the cover-slips were inverted and mounted on a microscope slide with ProLong® Gold Antifade Reagent ( Invitrogen ) and cells were visualized using a Leica DM 5500B microscope ( Leica ) . Photomicrographs were taken using a DFC 350 FX camera ( Leica ) . To quantify how many C . albicans cells bound ferritin , at least 100 cells per cover-slip were counted and percent binding was calculated by counting the total number of cells and the number of cells displaying fluorescent signal . All binding assays were performed in duplicates . Cells incubated without ferritin were used as a negative control . Because S . cerevisiae cells were detached during the washing steps described above , a different approach was used . The use of a fluorophore-coupled ferritin reduced the number of washing steps in the staining procedure and consequently left more cells on the coverslip for observation by fluorescent microscopy . Briefly , 5×105 cells were added per well in a 24 well plate containing Poly-L-Lysine-coated 12-mm diameter glass cover-slips in 1 ml RPMI 1640 . The cells were incubated for 1 h at 30°C . Afterwards , the medium was removed and 250 µl PBS with 1% BSA and 25 µg/ml ferritin coupled with dye DY-649 was added . After 15 min at 30°C , the cells were washed once with PBS , fixed , mounted and visualized under the microscope as described above for C . albicans cells . C . albicans wild-type cells ( SC5314 ) were grown on poly-L-lysine-coated cover-slips ( 0 . 5 mm in diameter ) in the presence or absence of 100 µg/ml ferritin for 6 h in RPMI 1640 . Afterwards , the cells were washed with PBS four times to remove non-bound ferritin and then immersed in fixative ( 4% formaldehyde , prepared from para-formaldehyde [Roth] and 0 . 1% glutaraldehyde [Roth] in 0 . 05 M HEPES ) at room temperature . After three min the fixative was replaced with fresh fixative and stored at 4°C overnight . The samples were dehydrated in ethanol ( Roth ) by progressively lowering the temperature to −35°C and infiltrated with Lowicryl K4M resin ( Polysciences , http://www . polysciences . com ) at −35°C [107] . The resin polymerization was carried out under UV light at −35°C for 24 h and for 10 h at 0°C . Ultra thin sections ( 60–80 nm thick ) were produced with an Ultracut S ( Leica ) and a diamond knife . Sections were collected on formvar filmed copper slot grids . Bright-field transmission electron microscopy was performed with an EM902 ( ZEISS , http://www . zeiss . de ) at 80 kV . Images were recorded with a 1 k CCD camera ( Proscan , http://www . proscan . de ) . Flow cytometry was used to quantify the binding of ferritin on the surface of C . albicans hyphal cells . C . albicans cells were grown under iron limitation , as described above , washed and counted . Approximately 106 cells in 1 ml RPMI 1640 medium were added to poly-L-lysine-coated ( Biochrom ) 12-mm diameter glass cover-slips in a 24 well tissue-culture plate ( TPP ) . The cells were incubated for 2 h at 37°C in 5% CO2 to induce hyphae . Next , the cells were washed once with PBS and incubated for 1 h in 0 . 5 ml PBS with 1% bovine serum albumin ( BSA ) and 100 µg/ml ferritin . Subsequently , the cells were washed three times with PBS to remove non-bound ferritin and fixed with 500 µl Roti®-Histofix 4% . The fixed cells were washed three times with PBS and incubated with rabbit anti-ferritin antibody ( Sigma-Aldrich ) diluted 1∶500 in PBS with 1% BSA for 1 h at room temperature . After washing , the cells were incubated with a goat anti-rabbit secondary antibody conjugated with Alexa 488 ( Invitrogen ) diluted 1∶500 . Finally , the cells were detached from the cover-slips using a pipet point and resuspended in 0 . 5 ml PBS . The fluorescent intensity of the hyphae was measured using a LSRII flow cytometer ( Becton Dickinson , http://www . bd . com ) . Fluorescence data for 10 , 000 cells of each strain were collected . Ferritin enriched epithelial cell monolayers ( described above ) were washed twice with PBS and infected with ∼105 C . albicans cells in serum-free RPMI 1640 medium for 6 h . Next , the samples were washed twice with PBS and fixed with 500 µl Roti®-Histofix 4% . C . albicans cells and TR146 cells were incubated separately and used as controls . All incubation times were performed in a humidified incubator at 37°C in 5% CO2 . To stain C . albicans cells localized only outside epithelial cells , before permeabilization , the samples were incubated with 12 . 5 µg Concanavalin A–fluorescein conjugate ( Invitrogen ) in PBS for 45 min at room temperature . After washing , the cells were permeabilized by incubation with 0 . 1% Triton X-100 for 15 min at room temperature . After washing three times with PBS , the samples were blocked using Image-iT™ FX signal enhancer ( Invitrogen ) for 30 min at room temperature in a humidity chamber . After washing three times with PBS , the cells were incubated with rabbit anti-ferritin antibody coupled with dye DY-649 diluted 1∶1000 in PBS with 1% BSA for 1 h at room temperature . To stain C . albicans cells localized outside and inside epithelial cells , the samples were incubated with 10 µg/ml Calcofluor White ( Sigma ) in 0 , 1 M Tris-hydrochloride ( pH 9 . 0 [Roth] ) for 20 min at room temperature . Finally , cover-slips were washed three times with ultra pure water , inverted and mounted on a microscope slide with ProLong® Gold Antifade Reagent . At least two experiments in duplicates were analyzed using a Leica microscope and 10 randomly chosen fields per cover-slip were photographed . A representative picture of each strain was selected . Ferritin depleted or enriched oral epithelial cell monolayers ( as described above ) were washed twice with PBS and infected with ∼105 iron starved C . albicans cells in serum-free RPMI 1640 medium for 3 h . The samples were washed twice with PBS and fixed with 500 µl Roti®-Histofix 4% . C . albicans cells alone were incubated separately and used as control . All samples were incubated in a humidified incubator at 37°C and 5% CO2 . The samples were stained to distinguish invading from non-invading fungal cells as described above . At least 100 randomly selected organisms were analyzed and the percentage of organisms that had invaded the epithelial cells was calculated . C . albicans cells growing under iron limitation , as described above , were washed and enumerated . Approximately 2×106 cells were added per well in a 24 well plate containing Poly-L-Lysine -coated 12-mm diameter glass cover-slips in 1 ml RPMI 1640 with 100 µg/ml ferritin . The strains used were CAI4 carrying CIp10; Δhgc1 carrying CIp10 and Δras1 carrying CIp10 . The plasmid CIp10 was used to reconstitute URA3 into the RP10 locus of each strain [108] . After 1 . 5 h incubation at 37°C under 5% CO2 , the medium was removed and 100 µl peqGOLD RNAPure ( PeqLab , http://www . peqlab . de ) was added per well . The cells were immediately removed from the cover-slips using a pipette point . For each strain , cells from 12 wells were pooled in a microcentrifuge tube and immediately shock frozen in liquid nitrogen . To verify that ferritin was bound to C . albicans hyphae as observed before , additional cover-slips for each strain were fixed and ferritin was stained as described . For transcriptional profiling , C . albicans microarrays ( Eurogentec ) were used as previously described [109] . RNA was co-hybridized with a common reference ( RNA from SC5314 grown in YPD medium , mid-log phase , 37°C ) . Slides were hybridized , washed and scanned as described [109] . Data normalization ( LOWESS ) and analysis were performed in Gene-Spring 7 . 2 software ( Agilent Technologies ) . Reliable expression of genes was defined as normalized expression of present genes that did not vary more than 1 . 5 standard deviations within replicate arrays . Genes were defined as differentially expressed if their expression was at least 2 times stronger or 2 times weaker in at least one strain compared to the common reference . Using the Benjamini and Hochberg false discovery test , a p-value<0 . 05 was considered as significant . Microarray data from four independent experiments ( two of them with dye swap ) were used . To identify genes involved in ferritin binding , genes were selected that were up-regulated ( ≥2 . 5 increase in expression compared to the common reference ) in wild-type and Δhgc1 cells , but unaltered or down-regulated ( ≤1 . 5 of the common reference expression ) in the Δras1 mutant . Raw data have been deposited in NCBIs Gene Expression Omnibus ( GEO , http://www . ncbi . nlm . nih . gov/geo/ ) and are accessible through GEO series accession number GSE11490 . Statistical significances ( p-values ) were calculated with the Student's two-tailed t-test function in Microsoft Excel , with exception of the microarray analysis described above . | Iron is an essential nutrient for all microbes . Many human pathogenic microbes have developed sophisticated strategies to acquire iron from the host as most compartments in the body contain little free iron . For example , in oral epithelial cells intracellular iron is bound to ferritin , a protein that is highly resistant to microbial attack . In fact , no microorganism has so far been shown to directly exploit ferritin as an iron source during interaction with host cells . This study demonstrates that the pathogenic fungus Candida albicans can use ferritin as the sole source of iron . Most intriguingly , C . albicans binds ferritin via a receptor that is only exposed on invasive hyphae . This receptor is Als3 , which is a member of the Als-protein family . Als3 was previously demonstrated to be an adhesin with invasin-like properties . Mutants lacking Als3 failed to bind ferritin , grew poorly with ferritin as an iron source and were unable to damage epithelial cells . Strains of the baker's yeast expressing C . albicans Als3 , but not two closely related proteins , Als1 or Als5 , were able to bind ferritin . Therefore , C . albicans uses an additional morphology specific and unique iron uptake strategy based on ferritin while invading into host cells where ferritin is located . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"infectious",
"diseases/fungal",
"infections",
"genetics",
"and",
"genomics/gene",
"function",
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis",
"genetics",
"and",
"genomics/gene",
"expression"
] | 2008 | The Hyphal-Associated Adhesin and Invasin Als3 of Candida albicans Mediates Iron Acquisition from Host Ferritin |
Glycosyltransferases are a class of enzymes that catalyse the posttranslational modification of proteins to produce a large number of glycoconjugate acceptors from a limited number of nucleotide-sugar donors . The products of one glycosyltransferase can be the substrates of several other enzymes , causing a combinatorial explosion in the number of possible glycan products . The kinetic behaviour of systems where multiple acceptor substrates compete for a single enzyme is presented , and the case in which high concentrations of an acceptor substrate are inhibitory as a result of abortive complex formation , is shown to result in non-Michaelian kinetics that can lead to bistability in an open system . A kinetic mechanism is proposed that is consistent with the available experimental evidence and provides a possible explanation for conflicting observations on the β-1 , 4-galactosyltransferases . Abrupt switching between steady states in networks of glycosyltransferase-catalysed reactions may account for the observed changes in glycosyl-epitopes in cancer cells .
With the ready availability both of computing power and software tools for numerical simulation , the mathematical modelling of metabolic systems has become a core component of cell biology . Models of classical metabolic pathways , such as glycolysis [1–3] , the citric-acid cycle [4] , the urea cycle [5] and biosynthetic pathways such as N-linked and O-linked glycosylation [6 , 7] , have been developed as a way to understand how such processes are regulated . Online repositories of such models , such as the BioModels database [8] , allow many of these models to be examined without the need for programming ability on the part of the user . Software such as E-Cell [9] have enabled more complex models to be constructed at the cellular or organelle level . This paper examines a particular class of metabolic model , in which one or more enzymes can act on multiple substrates . To this class belong the cytochrome P450 enzymes that are involved in detoxifying multiple xenobiotics [10] , ribonuclease P [11] and also the enzymes of N-linked glycosylation [12–15] . Such enzymes recognise multiple substrates , and the products of the reactions can themselves become substrates , thus introducing a form of competitive inhibition with catalysis . It is known that , in the case of two substrates acted upon by the same enzyme the Michaelis constant of the kinetic rate law will be modified to include the effects of competing substrates upon one another [16 , 17] . In the first part of this paper , a general form of the Michaelis-Menten equation for n competing substrates is derived , and extended to an ordered-sequential mechanism involving a donor molecule held in common by all reactions . In the second part , we model galactosyltransferase acting on an initial acceptor glycoprotein to form two products , each of which are substrates for the same enzyme . Here we propose a possible mechanism for such behaviour and apply it to the glycosyltransferase model , demonstrating the switching between stable steady states over a range of parameter values .
Consider the case of a general two-substrate enzyme mechanism , in which a donor molecule , Ax , transfers the x moiety to an acceptor , B , Ax + B → A + Bx , a reaction type that is common to the transferases . We consider the situation in which there are n acceptor substrates , B1 … Bn . For random-order binding of donor and acceptor ( Fig 1A ) , an expression for the initial rate of appearance of the jth acceptor product , Bxj , is v j = V j [ Ax ] [ B j ] K s Ax K m B j ( 1 + s B j ′ ) + K m B j [ Ax ] ( 1 + s B j ) + K m j Ax [ B j ] + [ Ax ] [ B j ] ( 1 ) where Vj = kj[E0] is the maximal velocity obtained at saturating levels of Ax and Bj , K m j Ax = K s Ax K m B j / K s B j , s B j = ∑ i ≠ j [ B i ] / K m B i and s B j ′ = ∑ i ≠ j [ B i ] / K s B i . The derivation of this equation under rapid-equilibrium conditions is given in the S1 Appendix . In this model the K s Ax and K s B i are , respectively , the individual dissociation constants of Ax and Bi from the E⋅Ax and E⋅Bi enzyme-substrate complexes , while the Michaelis constants of these species , K m j Ax and K m B i are the corresponding dissociation constants of the E⋅Ax⋅Bi complex . The s B j and s B j ′ terms are sums of dimensionless acceptor substrate concentrations representing the degree to which the enzyme is competitively inhibited by substrates other than Bj itself . In the absence of substrate competition , s B j = 0 , and Eq ( 1 ) reduces to the standard form of a bisubstrate enzyme mechanism . In the limit , as [Ax] → ∞ , ( 1 ) becomes v j = V j [ B j ] K m B j ( 1 + s B j ) + [ B j ] , ( 2 ) an equation that is similar in form to that obtained in other studies [14 , 18 , 19] . Although the s B j symbolism is a convenience in order to show which terms of the rate law are affected by competitor concentrations , a representation that is more useful in computer simulations is the sum of concentrations of all its substrates , each weighted by its K m B j or K s B j: A E = ∑ i = 1 n [ B i ] K m B i , A E ′ = ∑ i = 1 n [ B i ] K s B i . ( 3 ) Substituting into Eq ( 1 ) , v j = V j [ Ax ] [ B j ] K s Ax K m B j ( 1 + A E ′ ) + K m B j [ Ax ] ( 1 + A E ) . ( 4 ) Whereas a rapid-equilibrium random-order mechanism is a feature of polypeptide N-acetylgalactosaminyltransferase [20] , sulfotransferases [21] , fucosyltransferases [22] and sialyltransferases [23] , with other glycosyltransferases , such as those of the N-acetylglucosaminyltransferase and galactosyltransferase families , the enzyme must bind the donor first , before catalysis can occur [24] . Under quasi-steady-state conditions ( Fig 1B ) , the rate law for the compulsory order binding is ( Eq S3 in S1 Appendix ) : v j = V j [ Ax ] [ B j ] K s Ax K m B j ( 1 + s B j ) + K m B j [ Ax ] + K m j Ax [ B j ] + [ Ax ] [ B j ] ( 5 ) In such a case , the inhibitory effect of multi-substrate competition will lessen as the concentration of the donor is increased towards saturating levels .
Thus far , the possibility of abortive ( dead-end ) ternary enzyme complexes has not been considered , which in random-order mechanisms are likely to occur [25] . Experimental evidence for the existence such complexes can be the appearance of inhibition at high substrate concentrations; in the case of glycosyltransferases , the inhibition is usually that of the acceptor [26–29] , but can also be that of the donor [30] . If we consider only the acceptor , an examination of the mechanism ( Fig 1A ) reveals that four additional binding events can occur , with the E⋅Ax , E⋅Bj , E⋅A and E⋅Bxj complexes . We consider binding of Bj to the second of these complexes , E⋅Bj , to provide a possible explanation for substrate inhibition with increasing acceptor concentration . Not only will Bj bind , but so will any competitive acceptor-substrate Bi , i = 1 , … , n . The oligosaccharides attached to glycoproteins ( glycans ) can be multivalent , meaning that the same acceptor has more than one recognition domain . By way of illustration , the enzyme β-N-acetylglucosaminylglycopeptide β-1 , 4-galactosyltransferase ( GalT; EC 2 . 4 . 1 . 38 ) , catalyses the transfer of d-galactose ( Gal ) residue to a terminal N-acetylglucosaminyl ( GlcNAc ) residue on a glycoprotein , glycopeptide or polysaccharide , with the general reaction: UDP-α -D-Gal + β -D-GlcNAc-R → UDP + Gal-β 1 , 4 -D-β-D-GlcNAc-R A theoretical system , similar to that studied experimentally by Paquêt and co-workers [31] , is shown in Fig 2 , in which galactose is incorporated into glycopeptide in four steps , starting with the initial acceptor B1 , to form the final product with two terminal galactoses ( B4 ) . Hence , the products B2 and B3 are also substrates of the enzyme , since both contain a terminal GlcNAc on which it can act . All three substrates are therefore competitive inhibitors in the earlier sense , and can form a ternary complex with E⋅Bj , the free terminal β-d-galactose in the acceptor competing with the donor , UDP-Gal [33] . Before continuing , we make the parenthetic observation that reaction networks such as those in Fig 2 follow a binomial distribution pattern in the number of acceptors at each step . If the initial substrate has m sites on which an enzyme can act , then the m immediate acceptor-products of that substrate will each have m − 1 available sites . There will be a reaction hierarchy based on the combinatorial filling of available sites until the final product is reached at m = 0 , with the number of substrates at the kth step following the familiar mCk pattern , m C k = m ! k ! ( m - k ) ! . After k steps , a glycan substrate originally with m sites will have m − k sites remaining . The resulting network of all possible reactions , for a single acceptor possessing m sites at which an enzyme can act , will have N ( m ) nodes and E ( m ) edges , given by N ( m ) = ∑ k = 0 m m C k and E ( m ) = ∑ k = 0 m m C k ( m - k ) . Every node , whether substrate or product , will have degree m , with the in-degree of a node at the kth step being k and its out-degree being m − k . The number of possible pathways from initial substrate to final product will be P ( m ) = ∑ k = 0 m m C k k ( m - k ) . Every glycan will have up to m of each type of dissociation constant , for the enzyme of which it is a substrate , product or inhibitor . Extending the derivation of the rapid-equilibrium random equation in the S1 Appendix , an additional term will be required in the denominator to represent the abortive complex ( es ) . Since there are n substrates , there will be n2 ways in which to form E⋅Bk⋅Bi . A double summation over the indices i and k will be required , giving the additional term ∑ i = 1 n ∑ k = 1 n [ E · B k · B i ] = [ E · Ax ] K s Ax [ Ax ] ∑ i = 1 n ∑ k = 1 n [ B k ] K I B k [ B i ] K s B i where K I B k is the dissociation constant of the kth acceptor from complex E⋅Bk⋅Bi . The rate of appearance of the jth product will then be v j = V j [ Ax ] [ B j ] K s Ax K m B j ( 1 + s B j ′ + s I ) + K m B j [ Ax ] ( 1 + s B j ) + K m j Ax [ B j ] + [ Ax ] [ B j ] ( 6 ) with s I = ∑ i = 1 n ∑ k = 1 n [ B k ] K I B k [ B i ] K s B i = ∑ k = 1 n [ B k ] K I B k ∑ i = 1 n [ B i ] K s B i . When n = 1 , this reduces to v = V max [ Ax ] [ B ] K s Ax K m B + K m B [ Ax ] + K m Ax [ B ] + K m Ax K I B [ B ] 2 + [ Ax ] [ B ] ( 7 ) The equation for the compulsory order mechanism will be identical , and the more computationally efficient representation , equivalent to Eq ( 4 ) , is v j = V j [ Ax ] [ B j ] K s Ax K m B j ( 1 + A E ′ + s I ) + K m B j [ Ax ] ( 1 + A E ) ( 8 ) with two summation terms , A E = ∑ i = 1 n [ B i ] / K m B i and A E ′ = ∑ i = 1 n [ B i ] / K s B i . A general scheme for the formation of ternary enzyme-acceptor complexes is given in Fig 3A . This scheme does dual service , in illustrating both the formation of n2 inhibitory complexes in an n-substrate environment , but also the two catalytic mechanisms involving compulsory-order and random-order binding of substrates , which in the latter case only occurs when j = k , and for substrate inhibition at high concentrations , when j = k = i . The scheme illustrates two aspects of multi-substrate competition: productive , in which catalysis occurs , and non-productive , where there is inhibition as a result of abortive complex formation at higher acceptor concentrations . In the productive case , the n acceptors compete with each other for the E⋅Ax complex , in either random-order or compulsory-order binding mechanisms . In the non-productive case , higher acceptor concentrations compete with the donor for binding to the free enzyme , as well as with each other , for an enzyme-acceptor complex , resulting in non-productive multi-substrate inhibition in compulsory-order mechanisms . In the case of a random-order mechanism , the acceptor may bind to either the free enzyme or to the E⋅Ax complex in pathways leading to the productive ternary ( E⋅Ax⋅Bj ) complex . Therefore high substrate inhibition may result from the mis-oriented binding of acceptor to the free enzyme or binding of a second B to the E⋅B resulting in an abortive ternary complex . The binding site at which competition occurs may differ , depending on the enzyme mechanism involved . Fig 3B displays three curves of v vs [acceptor] , showing the relief of substrate inhibition that occurs as the donor concentration is increased , and in Fig 3C , the velocity-substrate surface defined by two K I B values , for n = 2 . The situation is more complicated when multiple binding sites exist on each molecule of acceptor . According to Fig 2 , B1 is a substrate , but B4 is not , while B2 and B3 can bind as substrate inhibitors , though B1 cannot because it does not have a terminal GlcNAc . B4 acts as a competitive ( product ) inhibitor of UDP-Gal , with two possible inhibition constants , K I , 1 B 4 and K I , 2 B 4 . The effective value of n is the number of edges , E ( m ) , in the network of a substrate with m recognition sites , as defined in the previous section , which gives 16 summands in sI . For the network in Fig 2 , therefore , s I = [ B 4 ] K I , 1 B 4 [ B 1 ] K s , 1 B 1 + [ B 4 ] K I , 2 B 4 [ B 1 ] K s , 1 B 1 + [ B 3 ] K I , 1 B 3 [ B 1 ] K s , 1 B 1 + [ B 2 ] K I , 1 B 2 [ B 1 ] K s , 1 B 1 + [ B 4 ] K I , 1 B 4 [ B 1 ] K s , 2 B 1 + [ B 4 ] K I , 2 B 4 [ B 1 ] K s , 2 B 1 + [ B 3 ] K I , 1 B 3 [ B 1 ] K s , 2 B 1 + [ B 2 ] K I , 1 B 2 [ B 1 ] K s , 2 B 1 + [ B 4 ] K I , 1 B 4 [ B 2 ] K s , 1 B 2 + [ B 4 ] K I , 2 B 4 [ B 2 ] K s , 1 B 2 + [ B 3 ] K I , 1 B 3 [ B 2 ] K s , 1 B 2 + [ B 2 ] K I , 1 B 2 [ B 2 ] K s , 1 B 2 + [ B 4 ] K I , 1 B 4 [ B 3 ] K s , 1 B 3 + [ B 4 ] K I , 2 B 4 [ B 3 ] K s , 1 B 3 + [ B 3 ] K I , 1 B 3 [ B 3 ] K s , 1 B 3 + [ B 2 ] K I , 1 B 2 [ B 3 ] K s , 1 B 3 , ( 9 ) in which K X , i B k denotes the ith dissociation constant of the kth acceptor , where X is either s ( dissociation from E⋅Bi ) or I ( dissociation from an abortive ternary complex ) . It has been observed that bistability can arise when an enzyme is inhibited by one of its substrates in an open system [34] , in which substrate enters at a zero-order rate , and exits at a rate that is first-order in the concentration of that substrate . If the substrate can diffuse into the reaction medium according to v diff = K ( [ B ] 0 - [ B ] ) , ( 10 ) where [B]0 is the concentration of exogenous substrate , then multiple steady-state solutions for the concentration of substrate can coexist for venz = vdiff . This is illustrated in Fig 3D , where the number of points of intersection of the line ( 10 ) with the curve described by Eq ( 7 ) will depend on the values of [B]0 and the diffusion constant , K . Bistability can be demonstrated through numerical simulation of the one-dimensional ODE system: d b d t = K ( b 0 - b ) - V max a b K a K b + K b a + K a b + K a K s b 2 + a b , ( 11 ) where a and b are the concentrations of the donor and acceptor , respectively . It is assumed that the donor concentration is constant , while the external concentration of b is chosen as the parameter to vary . The numerical continuation software AUTO , part of the ODE solver XPPAUT [35] , was used to calculate the steady-state level of b for increasing b0 . For the parameters a = 0 . 6 , K = 0 . 075 , Kb = 0 . 1 , Ks = 0 . 05 , Vmax = 1 and Ka = 0 . 6 , bistability is obtained for 1 . 518665 < b0 < 2 . 325853 ( Fig 4 ) . Within this range two stable steady states of acceptor concentration can coexist , as shown by upper and lower branches in b–b0 space . This can be confirmed by solving dvdiff/db = −K for b , using the parameters of Fig 3D , and computing the ordinate-axis intercept for vdiff at these two concentrations , which will be points of tangency of the two lines described by Eq ( 10 ) with the velocity–substrate curve . The values of b , computed in Mathematica ( version 11 . 0 . 1; Wolfram Research , Inc . ) , are b* = {0 . 10755 , 0 . 693546} . Substituting into ( 10 ) , we evaluate b* + vdiff ( b* ) /K = b0 , obtaining the corresponding solutions b0 = {1 . 51866 , 2 . 32585} . The reaction scheme shown in Fig 2 is modelled with five differential equations , d b 1 d t = K ( b 0 - b 1 ) - ( v 1 + v 2 ) ( 12 ) d b 2 d t = K ( b 0 - b 2 ) + v 1 - v 3 ( 13 ) d b 3 d t = K ( b 0 - b 3 ) + v 2 - v 4 ( 14 ) d b 4 d t = K ( b 0 - b 4 ) + v 3 + v 4 ( 15 ) d a d t = K ( a 0 - a ) - ( v 1 + v 2 + v 3 + v 4 ) ( 16 ) where the bi represent the acceptor concentrations [Bi] , i = 1 … 3 , a is the concentration of UDP-Gal , and the enzyme velocities v1 … v4 are described by Eq ( 8 ) . As before , the model assumes free diffusion of substrates into the medium in which enzyme is active [36] . There will be additional terms in s B j and sI , since there will be two sets of constants for the initial oligosaccharide substrate B1 , one set for each recognition site . The total enzymic rate of removal of B1 , for saturating levels of Ax , will be v 1 + v 2 = V 1 b 1 K m 1 + b 1 + V 2 b 1 K m 2 + b 1 . ( 17 ) Assuming that the maximal velocities of each of v1 and v2 are the same , we can solve for substrate concentration at half-maximal velocity , to obtain apparent Km as the geometric mean of the individual Michaelis constants , K m app = K m 1 K m 2 . Under the same assumption , for a substrate with m recognition domains , the apparent Km will be the solution to 1 = ∑ i = 1 m K m app K m j + K m app . ( 18 ) Numerical simulation of the model also displayed bistability ( Fig 5 ) . Using a two-parameter continuation , the region of a0–b0 space under which bistability exists was determined ( Fig 5B ) . The values of the external concentrations at the point of the cusp were found to be ( b0 , a0 ) = ( 0 . 07094 , 0 . 5959 ) . Bistability was also obtained by varying the diffusion constant , K ( Fig 5C ) ; a two-parameter continuation in a0–K space revealed a closed region of bistability ( Fig 5D ) .
These results demonstrate that the complex interplay of enzyme and substrate can give rise to nonlinear behaviour in systems of reactions held far from thermodynamic equilibrium . The significance of the present study is that small changes in one condition , such as the amount of available sugar-nucleotide donor [56] , might incur large and abrupt changes in the amount of product formed . Since GalT action influences the number of sites available for sialylation , such changes should have important implications for cancer progression and metastasis , which have been shown to be related to these processes [57] , and for biotechnology , such as in the production of therapeutic antibodies [58] , which can be influenced through control of metabolic flux [59] . More generally , the occurrence of bistability in metabolism could provide the basis for cellular long-term memory [60] . The commonly occurring pattern of substrate inhibition in transferases should complement the already known behaviours of models based on sigmoidal functions . For instance , it is known that different glycosylation enzymes associate , and co-locate with the Golgi , according to the ‘kin recognition’ model [61] , and may therefore display cooperativity . Whether a combination of cooperativity and substrate inhibition could lead to higher order dynamic behaviour , such as oscillations in acceptor concentration , is an open question that deserves further study . | While enzymes tend to have a narrow substrate specificity , there are a number of enzymes that are promiscuous , acting on a wide range of substrates . In this article we derive expressions for general multi-substrate competitive inhibition for the class of transferases , with particular emphasis on glycosylation . By extending the enzyme reaction mechanism to include inhibition by high substrate concentrations , we show that switching behaviour ( bistability ) is possible within a thermodynamically open systems of glycosylation enzymes . The biological implication of this finding is that small changes to a predictor variable may induce abrupt changes in the secreted products . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"biochemistry",
"competitive",
"inhibitors",
"post-translational",
"modification",
"proteins",
"enzyme",
"metabolism",
"enzymes",
"transferases",
"biology",
"and",
"life",
"sciences",
"glycosylation",
"enzymology",
"enzyme",
"kinetics",
"enzyme",
"inhibitors",
"enzyme",
"ch... | 2018 | A mechanism for bistability in glycosylation |
Human T lymphotropic virus type I ( HTLV-I ) infection is largely latent in infected persons . How HTLV-1 establishes latency and reactivates is unclear . Here we show that most HTLV-1-infected HeLa cells become senescent . By contrast , when NF-κB activity is blocked , senescence is averted , and infected cells continue to divide and chronically produce viral proteins . A small population of infected NF-κB-normal HeLa cells expresses low but detectable levels of Tax and Rex , albeit not Gag or Env . In these “latently” infected cells , HTLV-1 LTR trans-activation by Tax persists , but NF-κB trans-activation is attenuated due to inhibition by HBZ , the HTLV-1 antisense protein . Furthermore , Gag-Pol mRNA localizes primarily in the nuclei of these cells . Importantly , HBZ was found to inhibit Rex-mediated export of intron-containing mRNAs . Over-expression of Rex or shRNA-mediated silencing of HBZ led to viral reactivation . Importantly , strong NF-κB inhibition also reactivates HTLV-1 . Hence , during HTLV-1 infection , when Tax/Rex expression is robust and dominant over HBZ , productive infection ensues with expression of structural proteins and NF-κB hyper-activation , which induces senescence . When Tax/Rex expression is muted and HBZ is dominant , latent infection is established with expression of regulatory ( Tax/Rex/HBZ ) but not structural proteins . HBZ maintains viral latency by down-regulating Tax-induced NF-κB activation and senescence , and by inhibiting Rex-mediated expression of viral structural proteins .
Human T-lymphotropic virus type 1 ( HTLV-1 ) is a complex human retrovirus that infect approximately 10–20 million people worldwide [1] . In 3–5% of infected individuals a malignancy of CD4+ T cells known as adult T-cell leukemia/lymphoma ( ATL ) develops over a course of several decades [2] , [3] . Other diseases caused by HTLV-1 include HTLV-1-associated myelopathy/tropical spastic paraparesis ( HAM/TSP ) , HTLV-1 uveitis , and other inflammatory diseases . Most HTLV-1-infected individuals become asymptomatic virus carriers . The prevailing view of HTLV-1 infection is that it is rather inactive and integrated HTLV-1 proviral DNA replicates largely through mitotic expansion of host cells . This view is based on three lines of evidence ( see [4] for comments ) : ( i ) undetectable viral structural mRNA or protein expression in most infected PBMCs; ( ii ) undetectable cell-free viral particles in the plasma; and ( iii ) genetically stable viral genome due to very limited de novo infection through error-prone reverse transcription . However , longitudinal studies of HTLV-1 carriers indicate that the patterns of proviral DNA integration in PBMCs continue to evolve over time [5] , suggesting that de novo infection of naïve cells occurs constantly in virus carriers ( see [4] for a review ) . In infected individuals , there is also a robust CTL response against Tax and HBZ [6]–[8] , implicating immune activation via persistent expression of viral antigens . Whether and how HTLV-1 establishes latency and reactivates is not understood . HTLV-1 viral trans-activator Tax is a potent activator of viral mRNA transcription and the NF-κB pathway [3] , [9] . We have shown previously that hyper-activation of NF-κB by Tax induces cellular senescence [10] . Remarkably , HBZ , a regulatory protein encoded by the HTLV-1 anti-sense transcript [11] , dampens NF-κB activation [12] and thereby mitigates Tax-induced senescence [10] . These results raise the possibility that HTLV-1 infection may lead to two alternative outcomes dictated by the levels of Tax and HBZ [10] . When expressed at high levels , Tax drives robust viral replication , hyper-activates NF-κB , and triggers a senescence checkpoint response . Low levels of Tax and higher levels of HBZ , by contrast , result in moderation of NF-κB activation , prevention of senescence , and survival and persistence of HTLV-1-infected cells . Our previous studies have shown that most HeLa and SupT1 cells infected by HTLV-1 in culture become senescent or arrested in cell cycle progression [13] . Here we demonstrate that HTLV-1 infection indeed can lead to productive infection with expression of all viral proteins , NF-κB activation , and senescence; or latent infection with expression of regulatory but not structural proteins . HTLV-1 latency is regulated by HBZ , which dampens LTR and NF-κB activation by Tax . The latter activity of HBZ prevents senescence induction and allows latently infected cells to proliferate . Interestingly , HBZ further inhibits Rex-mediated export of intron-containing viral mRNAs , thereby shutting off Gag , Gag-Pol , and Env expression , and virus production . The latent provirus can be reactivated by over-expressing Rex or down-regulating HBZ . Thus , the “latency state” of HTLV-1 in the cell culture system resembles that established by γ-herpesviruses such as EBV and KSHV , which express a handful of potentially oncogenic latency-associated viral proteins and RNAs that stimulate mitotic expansion of latently infected cells . We speculate that the persistent expression of Tax and HBZ during the early stage of HTLV-1 latency propels infected cells to proliferate .
We have previously derived a cell line known as HeLa-G that harbors a reporter cassette , 18×21-EGFP , consisting of the enhanced green fluorescence protein ( EGFP ) gene under the transcriptional control of 18 copies of the Tax-responsive 21-base-pair enhancer element [14] . HeLa-G cells express abundant GFP as a function of Tax expression either after HTLV-1 infection or transduction of the tax gene . We infected them by co-culture with mitotically inactivated HTLV-1-producing T-cell line , MT2 ( Fig . 1A Top , see Materials and Methods ) , and found that most infected cells ( 98% ) became senescent ( Fig . 1A left panel ) . However , a careful examination of infected HeLa-G cells revealed that a small population ( 2% ) continued to proliferate ( Fig . 1A middle panel ) . Since our recent data indicate that hyper-activation of NF-κB by Tax is responsible for inducing cellular senescence [10] , [15] , we also tested the effect of HTLV-1 infection on a HeLa-G-derived cell line , HeLa-G/ΔN-IκBα in which the transcriptional activity of NF-κB is shut off by the stable expression of a degradation-resistant form of IκBα ( ΔN-IκBα ) . As anticipated , HTLV-1-infected HeLa-G/ΔN-IκBα cells continued to proliferate after infection ( Fig . 1A right panel ) . We also monitored infected HeLa-G and HeLa-G/ΔN-IκBα cells by immunofluorescence for p65/RelA , the capsid protein p24 , and Rex at 48 hours after infection . As shown in Fig . 1B , NF-κB is activated in HTLV-1 infected HeLa-G cells as revealed by the localization of p65/RelA to the nucleus . Most of these cells became senescent as described above . By contrast , in HTLV-1 infected HeLa-G/ΔN-IκBα cells , p65/RelA is localized in the cytoplasm as a result of inhibition by ΔN-IκBα . Doublets of infected GFP-positive HeLa-G/ΔN-IκBα cells could be seen , indicative of cell proliferation . Both types of infected cells express Tax , Rex , and p24 as might be expected ( Fig . 1B ) . Proliferating HeLa-G/HTLV-1 and HeLa-G/ΔN-IκBα/HTLV-1 clones were isolated by cell sorting based on GFP expression , expanded in cell culture , and characterized . Intriguingly , while the HeLa-G/ΔAN-IκBα/HTLV-1 clones expressed robust levels of p24 ( Fig . 1B ΔN-IκBα/HTLV-1 clones 1–3 ) , no p24 expression was detectable in HeLa-G/HTLV-1 clones ( Fig . 1B HeLa-G/HTLV-1 clones 1 to 5 ) . PCR analyses showed the chromosomal DNA of isolated clones to be positive for integrated HTLV-1 proviral DNA . Results from 3 ( G1-3 ) and 2 ( ΔN1-2 ) representative clones of each group are shown ( supplementary Fig . S1 ) . While all PCR products of ΔN1 and ΔN2 clones were of correct sizes as might be expected , G1-3 clones lacked ( G1 and G2 ) or yielded a smaller env PCR product ( G3 ) in the region that spans nucleotides 5318-5784 of the HTLV-1 genome ( see supplementary Fig . S1 and Table S1 ) , indicating gene deletions in this env region . As indicated by immunoblotting , most if not all HeLa-G/ΔN-IκBα/HTLV-1 cells were productively infected by HTLV-1 and abundantly expressed Gag ( p24 and p19 matrix protein , abbreviated as p19 ) , Env ( gp46 ) , Tax , and Rex ( Fig . 2A ) . Together with the results described above ( Fig . 1A left panel ) , these data indicate that cells productively infected by HTLV-1 usually undergo senescence as a result of chronic NF-κB activation by Tax . However , when NF-κB activity is blocked , the senescence response is prevented and the productively infected cell ( PIC ) population can grow and divide , and be established as individual virus-producing clones . We next examined in depth the 3 cell lines derived from the minor population of proliferating HTLV-1-infected HeLa-G ( HeLa-G/HTLV-1 ) cells whose NF-κB activity was unaltered . Interestingly , all three expressed Tax and Rex , albeit at lower levels ( Fig . 2A right panels ) , but showed no detectable expression of p19 , p24 , or gp46 ( Env ) ( Fig . 2A ) . The absence of Env from these clones correlated with env mutations detected by PCR ( Fig . S1 ) . Since only GFP+ cells were sorted and isolated , the positive detection of Tax expression is perhaps not surprising . These clones are designated as latently infected cells ( LICs ) for reasons that will become obvious later . It should be pointed out that Tax/Rex and HBZ expression is very low for LIC clone 3 ( LIC3 ) , where Tax expression is only detectable by the 18×21-EGFP reporter and Rex can only be seen by immunoblotting occasionally ( see below ) . Activation of viral transcription by Tax is intact in LICs and PICs ( albeit at a very low level for LIC3 ) as revealed by significant luciferase activities after transfection with an LTR-Luc reporter ( Fig . 2B ) . This suggests that for LTR activation , the levels of Tax expressed in LICs and PICs are not limiting , with the exception of LIC3 . This is as might be expected since each LTR has only 3 Tax-responsive 21-bp repeat elements ( TxREs ) , and as the Tax/CREB complex recruited to the TxREs is known to have a high affinity for them , the effective concentrations of Tax necessary to drive LTR transcription need not be high . The lower levels of viral expression in LICs may be related to the chromosomal environments of proviral DNAs that dampen Tax-mediated trans-activation . As anticipated , the NF-κB activity in PICs is profoundly inhibited by the IκBα super-repressor ( ΔN-IκBα ) despite Tax expression . This is confirmed by the absence of detectable luciferase activity in them after transfection of an NF-κB reporter plasmid , E-selectin-Luc ( Fig . 2C PIC lanes ) . Despite detectable Tax expression , NF-κB activity was not significantly induced in LICs ( Fig . 2C ) . We think this is due to the expression of HBZ , which is known to down-modulate NF-κB activity , albeit not as drastically as ΔN-IκBα [10] , [12] . Indeed , unspliced , but not spliced HBZ mRNA was readily detected and its levels are similar in both PICs and LICs ( again , with the exception of LIC3 ) as determined by RT-PCR ( Fig . 3A ) , consistent with the notion that its expression is independently regulated . The abundance of unspliced versus spliced HBZ mRNA most likely depends on the availability of splicing factors and can vary from cells to cells . Finally , it should be pointed out that although existing HBZ antibody can detect HBZ after DNA transfection , its sensitivity is insufficient for detecting HBZ during infection . While the Tax/Rex mRNA ( pXIII ) levels in LICs were lower than those in PICs as indicated by mRNA quantitation ( Fig . 3B , LIC1/PIC1 and LIC1/PIC2 about 1/3 and 1/4 respectively ) , greater differences were seen for Gag-Pol mRNAs ( Fig . 3B , LIC1/PIC1 and LIC1/PIC2 about 1/5 and 1/8 respectively ) , Thus , more viral mRNAs of PICs are in the unspliced ( Gag-Pol ) form , and less so for LICs . The Env mRNAs in LICs were much lower ( Fig . 3B ) . We think this is due to nonsense-mediated degradation caused by env mutations ( Fig . S1 ) . As mentioned earlier , the reduced Gag-Pol and Tax/Rex mRNA expression in LICs is most likely associated with the chromosomal sites of integration . The mRNA stabilization by higher levels of Rex in PICs [16] , [17] likely also influences viral mRNA expression . Finally , inhibition of Rex-mediated nuclear export of intron-containing viral mRNAs in LICs contributes to the altered gene expression profile as elaborated below . Since the level of Rex is lower in LICs , we tested the possibility that their lack of Gag expression might be caused by a block in the nuclear export of unspliced viral mRNAs . Nuclear and cytoplasmic RNAs were fractionated for the LIC clones 1–3 and PIC clones 1 and 2 , and subjected to qRT-PCR to quantify the nuclear and cytoplasmic levels of Gag-Pol , pX-III , and the control β-actin mRNAs . As indicated in Fig . 4A , there is a block in nuclear export of Gag-Pol mRNA in LICs with nuclear to cytoplasmic ( N/C ) ratios of approximately 30–40 . By contrast , N/C ratios of Gag-Pol mRNA in PICs were approximately 1–2 . The N/C ratios of the doubly spliced pX-III mRNA range from 0 . 6 to 2 in both cell types . Importantly , even though the levels of Rex in LICs were modest compared to those in PICs ( Fig . 2A Rex panel on the right ) , it was detectable by immunoblotting , and was expected to export at least some unspliced Gag-Pol mRNAs . Intriguingly , Rex appeared altogether inactive in LICs . We next asked if p24 expression could be reactivated by Tax or Rex in the LICs . Contrary to conventional wisdom , over-expression of Tax in LICs via transfection of an expression vector , Bc12-Tax , had very little impact on inducing p24 expression except in LIC clone 3 ( Fig . 4B right 3 lanes ) . By contrast , when a Rex-expression plasmid was transfected , p24 expression was readily induced in all three LIC lines ( Fig . 4B compare middle 3 and left 3 lanes ) . As expected , more robust reactivation of LIC3 could be achieved with co-transfection of both Rex and Tax ( supplemental Fig . S2 ) . These results again suggest that Tax is not a limiting factor for viral gene expression in many LICs . Importantly , they also suggest that the endogenous Rex in LICs may be defective or inhibited; and the defect or block can be complemented or overcome by over-expressing exogenous Rex . To determine if the activity of Rex was blocked or defective in LICs , we transfected both LICs and PICs with an HTLV-1 Rex reporter plasmid , pRxRE1-RLuc ( Fig . 5A upper panel; [18] ) . This reporter encodes an mRNA that contains the HTLV-1 Rex-response element ( RxRE1 ) in the 3′ end , and an intron that harbors the coding sequence for the Renilla luciferase ( RLuc ) . In the absence of Rex , the RLuc sequence is removed by splicing . The mRNA exported to the cytoplasm is therefore without Rluc , hence no luciferase activity is expressed . In the presence of Rex , however , the RxRE1-RLuc-intron-containing mRNA is exported to the cytoplasm and translated to yield Renilla luciferase . Indeed , pRxRE1-RLuc-transfected PICs readily produced Renilla luciferase activity ( Fig . 5B PIC 1 and 2 ) , consistent with their chronic production of viral structural proteins facilitated by higher levels of Rex . By contrast , RxR1E-RLuc-transfected LICs expressed little luciferase activity ( Fig . 5B LIC 1–3 ) , in agreement with the notion that Rex is either defective or inhibited in LICs . No luciferase activity is detectable in transfected control HeLa-G cells ( Fig . 5B leftmost lane ) . We next asked if the lack of Rex activity in LICs might be due to inhibition by a trans-acting viral factor . Of all viral proteins , we thought HBZ to be the most likely to have a role in inhibiting the nuclear export function of Rex . This is because low or no NF-κB activation was detected in LICs despite Tax expression , suggesting that HBZ was expressed in LICs and was inhibiting NF-κB activation and senescence induction as previously proposed [10] , [12] . Since HBZ is already playing a critical role in rendering possible the continuous proliferation of Tax-expressing cells [10] , it is logical that it might additionally prevent virus production so as to establish latency . To test if HBZ could block nuclear export of unspliced mRNA by Rex , we titrated Rex and pRxRE1-RLuc plasmids to determine the minimal amount of Rex DNA needed to achieve maximal reporter activity ( Fig . 5C lanes 1–5 ) . That amount of Rex ( 50 ng ) was then used in co-transfection with increasing amounts of an HBZ-expressing plasmid . Indeed , a dose-dependent reduction in Renilla luciferase activity was observed when Rex and pRxRE1-RLuc reporter were co-transfected with HBZ ( Fig . 5C lanes 6–8 ) , suggesting that HBZ blocked Rex-mediated export of RexRE1-RLuc-intron mRNA . This effect of HBZ is mediated by the HBZ protein and not mRNA , because an HBZ mutant with the ATG translational start codon mutated to TTG failed to block the activity of Rex ( Fig . 5C lanes 9 and 10 ) . To confirm that HBZ is indeed responsible for preventing Gag expression in LICs , we derived a puromycin-selectable lentiviral vector encoding an HBZ-targeting shRNA under the transcriptional control of the Pol III-dependent snRNA U6 promoter . LICs were transduced with the LV-HBZ-shRNA-SV-puro ( Fig . 6A & B , HBZ shRNA ) or the empty vector , and selected in puromycin-containing medium for seven days . Down-regulation of HBZ mRNA in the HBZ-shRNA-treated cells was confirmed by qRT-PCR ( Fig . 6A ) . As expected , after HBZ knockdown , an increase in Tax/Rex expression was observed ( Fig . 6B ) , consistent with the notion that HBZ down-regulates viral gene expression [19] . Importantly , p24 expression was significantly induced and readily detected for LIC clones 1 and 2 , indicating viral reactivation ( Fig . 6B ) . LIC clone 3 did not show appreciable p24 expression , most likely because its level of sense mRNA transcription was too low . These results demonstrate that HBZ is responsible for down-regulating Tax-mediated viral sense mRNA transcription , and blocking Rex-mediated nuclear export of intron-containing HTLV-1 mRNAs and expression of viral structural proteins in LICs . It is important to note that in order for HBZ to exert effective control over viral replication , low levels of Tax/Rex expression are needed . The levels of Rex and Tax in PICs are 3- to 4-fold higher than those in LICs ( Fig . 2A right panels ) . While one cause for this difference may be the sites of proviral integration , since the major difference between LICs and PICs is the profound NF-κB inhibition by ΔN-IκBα in the latter , we thought strong NF-κB repression might contribute to the increased expression of Rex and Tax , and thereby reactivated latent HTLV-1 genome . Indeed , stable expression of the IκBα super-repressor , ΔN-IκBα , in LIC clones up-regulated Rex and Tax expression , and reactivated the latent proviruses as indicated by the induction of p24 expression , especially for LIC clones 1 and 2 ( Fig . 6C ) . The mechanism by which NF-κB inhibition induces Rex and Tax expression is currently under investigation .
In this paper , we present evidence to demonstrate that HTLV-1 infection can lead to two alternative outcomes based on the relative expression of Tax/Rex and HBZ ( summarized in Fig . 7 ) . In most HTLV-1 infected cultured cells , Tax/Rex expression is robust and viral structural proteins are abundantly expressed . In this condition , IKK/NF-κB is hyper-activated by Tax , triggering a host senescence response . When senescence induction is prevented by inhibiting NF-κB , cell clones productively infected by HTLV-1 can be readily established ( Fig . 1 ) . On the relatively rare occasions when Tax/Rex expression is weak , HBZ moderates NF-κB activation by Tax [10] , [12] , thus averting the host senescence response and allowing infected cells to continue to proliferate . Importantly , HBZ additionally inhibits Rex-mediated nuclear export of intron-containing mRNAs , thereby shutting off Gag , Gag-Pol , and Env production and committing infected cells into latency ( Fig . 7 ) . In latently infected cells , viral regulatory proteins , Tax , Rex , and HBZ , but not structural proteins are persistently expressed . Reactivation of Gag , Gag-Pol , and Env expression is achieved through up-regulation of Rex or down-regulation of HBZ ( Fig . 7 ) . Most interestingly , strong inhibition of NF-κB increases Rex and Tax expression and reactivates HTLV-1 replication . We did not investigate the involvement of other HTLV-1 accessory proteins including p21Rex , p12I , p13II , and p30II in the present model . P30II is a nuclear and nucleolar protein thought to be a post-transcriptional modulator of viral replication [20] . Published data suggest that p30II retains the doubly-spliced Tax/Rex mRNA in the nucleus and thereby down-modulates viral gene expression by reducing the levels of Tax and Rex [20] . It has also been shown to interact with CBP/p300 and interfere with LTR trans-activation by Tax [21] . The continuous expression of Tax and Rex in LICs , albeit at low levels , indicates that the Tax/Rex mRNA is not sequestered . Importantly , the LICs described here were identified by virtue of Tax-driven GFP expression via the 18×21-EGFP reporter cassette . Cells that had no Tax/Rex expression would not have been scored in this system . Thus , a study of the latency state where Tax/Rex expression is completely silenced by p30II requires other approaches . An unexpected finding from the present study is the high frequency with which LIC clones were found to harbor env mutations ( Fig . S1 ) . Additional attempts to isolate LIC clones that contain fully functional proviral DNA were not successful . This contrasts with the PIC clones , most if not all of which readily express Tax/Rex , Gag and Env . Whether the profound NF-κB inhibition in HeLa-G/ΔN-IκBα cells is responsible for shutting off innate host defense mechanism ( s ) that target mutations to retroviral genomes remains to be investigated . The outcomes of HTLV-1 infection reported herein can adequately explain data from clinical and in vivo studies [4] , [22] . In this model , cells productively infected by HTLV-1 immediately enter into senescence ( Fig . 7; and ref . [10] ) and most likely become eliminated by cytotoxic T lymphocytes [22]–[24] . Removal of senescent cells by NK cells is also a likely possibility [25] , [26] . The latently infected cells , however , continue to express Tax , Rex and HBZ and can rely on the mitogenic activities of Tax , and HBZ protein and/or mRNA to drive cell proliferation . This is in accordance with the detection of Tax/Rex mRNA in a small population of infected cells reported previously [6] , and the robust CTL response against Tax and HBZ seen in infected individuals [6] , [8] , [27] . The dynamically evolving proviral integration patterns in asymptomatic HTLV-1 carriers can now be explained by viral reactivation and de novo infection of naïve cells brought about via up-regulation of Rex and/or down-regulation of HBZ expression . In this model , it is necessary for latently infected cells to express only muted levels of Tax/Rex such that their activities can be controlled by HBZ . Indeed , proviral DNA integration sites in asymptomatic carriers were found mostly to locate in transcriptionally silent regions of chromosomes [28] . Finally , the observation that strong NF-κB inhibition can increase Rex expression and viral reactivation has clinical implications . Bortezomib , a proteasome inhibitor that inhibits NF-κB by stabilizing IκBα , has been entered into clinical phaseI/II trials for ATLL . Although ATLL cells in general no longer replicate HTLV-1 , latently infected cells most likely persist in patients and may be reactivated by NF-κB inhibitors so as to influence the course of the disease . Based on present data , administration of antivirals to prevent HTLV-1 reactivation and spread may be advisable when bortezomib is used for ATLL treatment . Given the alternative outcomes of HTLV-1 infection , do ATL cells emerge from productively or latently infected T lymphocytes ? We think productive HTLV-1 infection of T lymphocytes whose senescence checkpoint response has been impaired is most likely the first step in ATL development . Such precancerous lymphocytes can express sufficient levels of Tax to overcome HBZ inhibition and achieve persistent IKK/NF-κB activation without inducing senescence . Through the loss of senescence response , the proliferative and survival advantages conferred by Tax-driven NF-κB activation , and the mitogenic activity of HBZ , such lymphocytes can readily undergo clonal expansion . This view agrees with recent high-throughput DNA sequencing data showing that most proviral integrations in ATL cells occur in transcriptionally active regions in the sense orientation [28] . As Tax is a primary CTL target , the loss of Tax and forward ( sense ) viral gene expression from precancerous T lymphocytes is selected and occurs through 5′ LTR DNA methylation , nonsense mutations , and deletions [29] . However , after Tax and Tax-dependent NF-κB activation are lost from pre-cancerous T lymphocytes , the impairment to the senescence checkpoint response remains , and can facilitate the evolution of Tax-independent NF-κB activation . As the expression of HBZ mRNA and protein is independently regulated , they can continue to exert mitogenic effect to propel the proliferation of cancerous cells [30]–[32] . Understanding how Rex , Tax , and HBZ expression is altered by cell signaling and cellular physiology to affect latency establishment and viral reactivation can facilitate the control of viral infection to prevent progression to disease .
HTLV-1 infections were performed in a 10 cm dish by co-culturing HeLa-G or HeLa-G/ΔN-IκBα cells ( 1–2×106 ) with HTLV-1-producing MT2 cells ( 3×106 ) that have been mitotically inactivated by mitomycin C ( MMC ) treatment ( 10 µg/ml for 2 hours ) . The co-culture was carried out in the presence of polybrene ( 8 µg/ml ) for 16 hours . MT2 cells were then removed by washing with phosphate buffered saline ( PBS ) . Fresh media was added and cells were grown for an additional 24 hours , and then harvested . GFP-positive cells were isolated using a cell sorter ( BD FACSAria ) housed in a lamella flow hood under aerosol-protection condition . Sorted cells were plated at low density on a 15-cm dish . Individual proliferating colonies were picked after a week into 96-well plates and further screened for the integrated HTLV-1 genome by PCR . HeLa-G or HeLa-G/ΔN-IκBα cells grown on chamber glass slides were infected with HTLV-1 as described above . Forty eight hours after infection , cells were fixed with 4% paraformaldehyde , permeabilized with 0 . 1% Triton X-100 , and immunostained overnight with the indicated primary antibodies followed by Alexa Fluor 568 secondary antibodies ( Invitrogen , Carlsbad , CA . ) Nuclei were counterstained using 4′ , 6′-diamidino-2-phenylindole ( DAPI ) . The slides were then mounted in a fluorescence mounting medium ( Dako ) . Images were captured using a Zeiss Pascal inverted confocal microscope . Standard procedures were used for immunoblotting . Typically , 30–50 µg of total proteins was used in each sample . The HTLV-1 Tax hybridoma monoclonal antibody 4C5 had been described previously [15] . The rabbit polyclonal antibody against Rex was a generous gift of Dr . Gisela Heidecker . HTLV-1 p24 antibody was purchased from Advanced Bioscience , HTLV-1 p19 and HTLV-1-gp46 ( Env ) antibodies were from Zeptometrix , IκBα , β-actin , goat anti-mouse , and goat anti-rabbit HRP conjugated secondary antibodies were from Santa Cruz . HTLV-1-infected cells generated as above were harvested , washed , and dissolved in lysis buffer ( 50 mM Tris , pH 8 . 0 , 10 mM EDTA , 100 mM NaCl , 0 . 5% sarcosyl , 0 . 5 mg/ml proteinase K ) . DNA was then precipitated with isopropanol . To screen for the full-length HTLV-1 provirus , sequence-tagged site polymerase chain reaction was carried out as described [33] using primers specified for distinct regions/genes of the HTLV-1 genome ( supplementary Table S1 ) . PCR products were resolved on 2% agarose gels . Total mRNA , nuclear and cytoplasmic mRNAs from HTLV-1-infected cell clones were isolated using the PARIS kit ( Ambion ) according to manufacturer's instructions . Contaminating genomic DNA was removed using the turbo DNA-free kit ( Ambion ) . Complementary DNA ( cDNA ) was synthesized from 500 ng of RNA in a total volume of 10 µl with iScript reverse transcription super mix ( Biorad ) . The cDNA used for HBZ mRNA quantitation was prepared using a gene-specific antisense primer , HBZ-R2 ( see supplementary Table S 2 ) , to avoid contamination from HTLV-1 sense strand cDNA . Real-time PCR was performed using 2 µl of the cDNA as template in a 20 µl reaction , using gene-specific primers ( see supplementary Table S2 for sequences ) and LightCycler DNA SYBR Green I master mix ( Roche applied science ) in a LightCycler thermal cycler ( Roche Diagnostics ) . The mRNA level in each sample was normalized to that of the β-actin mRNA . Relative mRNA levels were calculated using the 2−ΔCt method [34] . To determine the nuclear-to-cytoplasmic ratio of a given viral mRNA species , 150 ng of nuclear or 300 ng of cytoplasmic mRNA was used for cDNA synthesis . Complementary DNA from each fraction was quantified for the level of Gag-Pol , pXIII , and β-actin mRNA transcripts respectively by PCR , the relative abundance of each viral mRNA in the nuclear or cytoplasmic compartment was determined by normalizing the level of a given viral mRNA against that of the β-actin mRNA in the same mRNA preparation . The nuclear-to-cytoplasmic ratio of viral mRNAs was calculated based on the relative abundance measurements , and then plotted . Cells ( 3×105 ) were seeded into a 24-well plate overnight . After 16 hours , DNA transfections were performed using Fugene HD reagent ( Promega ) . Two hundred nanograms of each reporter plasmid HTLV-1 LTR-Luc , E-selectin-Luc , or RxRE-RLuc were used in the respective luciferase reporter assay . The RxRE1-RLuc contains the HTLV-1 Rex-response element downstream of the Renilla luciferase reporter ( RLuc ) gene located within an intron ( kindly provided by Dr . Jaqueline Dudley ) . The amounts of Rex or HBZ expression plasmid used range from 0 to 200 ng . All transfections were performed in duplicates . The total DNA amount ( 500 ng ) was kept constant in all transfections using an empty vector plasmid , pcDNA3 . 1 . Twenty nanograms per well of control luciferase plasmid pGL3-Luc ( firefly ) or pRL-TK ( renilla ) were also included in each transfection . After 48 hours , cells are harvested , and luciferase activity was measured using the Dual-Luciferase Reporter Assay System ( Promega ) according to the manufacturer's instructions . Transfection efficiencies were normalized using either TK-renilla or PGL3-Luc . Data are mean ± s . d . from at least three independent experiments . The lentivirus expressing the degradation resistant mutant of IκBα , LV-ΔN-IκBα-SV-puro , has been described earlier [10] . For delivery of the anti-HBZ short hairpin RNA ( shRNA ) , A small hairpin RNA ( shRNA ) expression cassette containing the HBZ shRNA sequence [30] downstream of the mouse U6 promoter was amplified by PCR and cloned into a self-inactivating lentiviral vector , SMPU [14] , engineered to contain the SV-puro ( the puromycin resistance gene placed under the control of the SV40 early promoter ) . Lentiviral vectors were prepared as previously reported [15] . HeLa-G cells were transduced with the lentiviral vector in DMEM supplemented with 10% fetal bovine serum and selected in the same medium containing puromycin ( 1 µg/ml ) . | Most HTLV-1-infected individuals are asymptomatic . It is thought that the proviral DNA is transcriptionally inert and HTLV-1 replicates through mitotic expansion of host cells . The evolving provirus integration patterns in HTLV-1 carriers , however , suggest new infection occurs continuously . Whether or how HTLV-1 establishes latency and reactivates is unclear . We show that HTLV-1 infection in culture can lead to two alternative outcomes — productive infection accompanied by senescence or latent infection followed by clonal expansion — based on the relative expression of regulatory proteins: Tax , Rex , and HBZ . HTLV-1 latency is established by HBZ , and reactivation is achieved by Rex through regulating nuclear export of viral mRNAs . Elucidating mechanisms underlying HTLV-1 latency and reactivation can facilitate virus control to prevent progression to disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"hiv",
"immunopathogenesis",
"clinical",
"immunology",
"virology",
"biology",
"and",
"life",
"sciences",
"immunology",
"microbiology",
"viral",
"diseases"
] | 2014 | Regulation of Human T-Lymphotropic Virus Type I Latency and Reactivation by HBZ and Rex |
Nipah virus ( NiV ) is an emerging disease that causes severe encephalitis and respiratory illness in humans . Pigs were identified as an intermediate host for NiV transmission in Malaysia . In Bangladesh , NiV has caused recognized human outbreaks since 2001 and three outbreak investigations identified an epidemiological association between close contact with sick or dead animals and human illness . We examined cattle and goats reared around Pteropus bat roosts in human NiV outbreak areas . We also tested pig sera collected under another study focused on Japanese encephalitis . We detected antibodies against NiV glycoprotein in 26 ( 6 . 5% ) cattle , 17 ( 4 . 3% ) goats and 138 ( 44 . 2% ) pigs by a Luminex-based multiplexed microsphere assay; however , these antibodies did not neutralize NiV . Cattle and goats with NiVsG antibodies were more likely to have a history of feeding on fruits partially eaten by bats or birds ( PR = 3 . 1 , 95% CI 1 . 6–5 . 7 ) and drinking palmyra palm juice ( PR = 3 . 9 , 95% CI 1 . 5–10 . 2 ) . This difference in test results may be due to the exposure of animals to one or more novel viruses with antigenic similarity to NiV . Further research may identify a novel organism of public health importance .
Nipah virus ( NiV ) is a zoonotic paramyxovirus whose reservoir host is fruit bats of the genus Pteropus [1]–[3] . NiV was first recognized in a large outbreak in Malaysia where pigs were an intermediate host for the transmission of NiV infection in humans [4] , [5] . Outbreak investigators speculated that pigs were infected with NiV by ingesting partially eaten saliva-contaminated fruit dropped by Pteropus bats [6] . Pig farmers were more likely to be infected with NiV suggesting infected pigs transmitted NiV to humans through close contact [7] . Between 2001 and 2013 NiV has caused 227 recognized human infections in Bangladesh with a case fatality of over 75% [8]–[15] . Although there is no serological or microbiological confirmation of NiV infection in domestic animals in Bangladesh , three outbreak investigations have identified suggestive associations between domestic animals and human infection . In the 2001 outbreak in Meherpur , Bangladesh , human Nipah cases were 7 . 9 times more likely than controls to have contact with a sick cow ( odds ratio[OR] 7 . 9 , 95% confidence interval [CI] 2 . 2–27 . 7 ) [8] . In a 2004 outbreak , a NiV-infected child had a close contact history with two sick goats and in a 2003 human Nipah outbreak at Naogaon , Bangladesh , cases were more likely than controls to have had contact with a nomadic pig herd ( OR 6 . 1 , 95% CI 1 . 3–27 . 8 ) [16] , [17] . Bats frequently visited date palm trees and licked shaved surfaces of the trees to drink sap at night [18] . Date palm sap spoiled by bat feces is occasionally fed to cattle in Bangladesh [19] . Domestic animal infection with NiV may represent an immediate risk to human infection as well as a risk for further evolution of the virus for adaptation to mammals other than bats . We conducted a cross-sectional study to look for evidence of NiV antibodies in domestic livestock , including cattle , goats and pigs , and to identify exposures associated with NiV antibodies .
Field staff obtained written consent from the animal owners for data and sample collection . icddr , b's Research Review Committee , Ethical Review Committee and Animal Experimentation Ethics Committee reviewed and approved the study protocols . The protocol numbers are PR-10015 for the henipavirus study and 2008–063 for the Japanese encephalitis study . For assessing NiV exposure in cattle and goats , we selected Faridpur , Rajbari , Meherpur , Tangail and Naogaon districts as study sites because they had previous human NiV outbreaks . We identified the nearest Pteropus bat roost from the human index case's household for each of the five sites . We enrolled cattle and goats living within a 1000 meter radius of the fruit bat roost in each site . If an insufficient number of cattle and goats were identified , we extended this area up to 5000 meters in increments of 1000 meters . We enrolled the pig samples from a population based survey done in pigs in 3 adjacent Northwestern districts ( Naogaon , Rajshahi and Nawabganj ) of Bangladesh during May-September 2009 as part of a separate study on Japanese encephalitis [20] . Those three districts were chosen for pig sampling because of higher number of Japanese encephalitis cases reported from these areas [21] . For cattle and goat enrollment , we generated random latitude/longitude coordinates within a 1000 meter radius of each of the five selected Pteropus bat roosts using global positioning system ( GPS ) coordinates . From each GPS location , we identified the nearest household . For selecting subsequent households , we chose the nearest front door of every second household . We enrolled a maximum of three animals , either cattle or goats or both , that were either healthy or sick from each household . We selected animals aged >2 months or when they were weaned from the dam's milk and could feed on grass or other foods in the environment that may be contaminated with henipaviruses . For pig specimens the study team conducted a census of the pig population at Naogaon , Rajshahi and Nawabganj districts relying on the pig raisers' social network [22] . The primary objective of the pig sampling was a separate study exploring prevalence of infection with Japanese encephalitis virus , and as a result the field team did not collect the same information on fruit bat exposure as was collected for cattle and goats . Field workers visited the areas to collect data on demographics and management of pigs and sampled 312 pigs . The study team selected pigs over 6 months of age for sample collection because of their exposure to Japanese encephalitis virus for longer period . Field staff interviewed animal owners to collect information on their animal characteristics , management , ecological and environmental data using a structured questionnaire . The management data included rearing systems and feeding practices . We categorized feeding practices for cattle and goats as intensive ( animals are kept in pens and supplied feed entirely from outside ) , semi-intensive ( sometimes grazing and sometimes supplied feed in pens ) and extensive ( only grazing without supplementation ) . For pigs , field staff collected rearing system data on two categories including backyard ( pigs were allowed to graze in the nearby pasture ) and nomadic ( pigs were allowed to move from one area to another for scavenging feed ) . We collected five to eight ml of blood for preparing serum from each selected cattle , goat or pig using aseptic sterile equipment . All animal sera were tested at the Australian Animal Health Laboratory ( AAHL ) using a Luminex-based multiplexed microsphere assay that specifically detects antibodies to the soluble attachment glycoproteins ( sG ) of henipaviruses ( NiV and Hendra virus ( HeV ) ) [23] . Beads coated with either NiVsG or HeV sG were mixed with sera at a dilution of 1∶100 . Biotinylated Protein A/G and Streptavidin-PE were then used to detect bound antibody . Beads were interrogated by lasers in a BioRad BioPlex machine and the results recorded as the Median Fluorescent Intensity ( MFI ) of 100 beads . Bayesian mixture models were used to characterize the bimodal distribution of microsphere assay outputs to classify individuals as seropositive or seronegative , following methods described in Peel et al . [24] . In contrast to Peel et al ( 2013 ) , where similar results were obtained whether mixture models were fitted to data from different age groups within the one species simultaneously or independently , for the data from different species described here , optimal fitting was observed when each species was fitted independently . Conservative species-specific cutoffs were determined so that individuals with MFI values above this cutoff were >99% likely to be seropositive ( MFI = 300 for cattle and goats and MFI = 650 for pigs ) . Full details of the method , assumptions and results are provided in the Supporting Information to this manuscript . Cattle , goat and pig sera showing higher MFI values were further analyzed by western blot ( WB ) , enzyme-linked immunosorbent assay ( ELISA ) and serum neutralization test ( SNT ) . The WB test was used to detect non-neutralizing antibodies against recombinant N protein of henipaviruses [25] . A subset of NiVsG positive sera were also tested against Cedar virus ( CedV ) sG in the Luminex assay . Laboratory personnel at the Viral Special Pathogens Branch , Centers for Disease Control and Prevention tested all NiVsG positive sera , along with a randomly selected a subset of negative sera using their in-house enzyme-linked immunosorbent assay ( ELISA ) . Gamma-irradiated lysates from NiV-infected and mock-infected Vero E6 cells were used as antigens and Protein A/G used for detection of bound antibodies [26] . SNT was performed at AAHL under biosafety level ( BSL ) 4 conditions . Briefly , sera diluted 1∶10 was mixed with 200 TCID50 NiV in 96-well tissue culture plates , incubated for 30 minutes at 37°C and 100 ul containing 2×104 vero cells in suspension added . The cells were incubated for 3 days and then observed for viral CPE . We calculated the prevalence of antibodies separately for cattle , goats and pigs by dividing Luminex-positive animals by the total number of animals of that species tested . We calculated the prevalence ratio ( PR ) to identify the association between Luminex results and exposure variables by bivariate analysis . Before examining the independence of multiple explanatory variables , we framed a causal diagram to identify causal associations between variables of interest and to identify confounders as described [27] , [28] . Exposure variables having a prevalence ratio >1 in bivariate analysis and selected variables from the causal diagram were entered to construct the final model of multivariate logistic regression analysis . We adjusted all confidence limits for geographical clustering in both bivariate and multivariate logistic regression model to minimize clustering effect during animal enrollment . Based on geographical position of enrolled households , district wise cluster was formed with unique code . Confounding variables were also entered in the multivariate logistic regression model for adjustment during analysis . All statistical analysis was done by using STATA 10 . 0 .
We enrolled 400 cattle , 400 goats and 312 pigs between May 2009 and January 2011 . Among all enrolled cattle and goats , 798 ( 99% ) were reared in backyard farms , 587 ( 73% ) cattle and goats were fed using semi-intensive practices , 150 ( 19% ) were fed using intensive practices , and 63 ( 8% ) were fed using extensive practices . The median age of sampled cattle was 33 months; 67% were female and 46% were a local breed . The mean age of sampled goats was 21 months; 69% were female and 94% were Black Bengal breed . The study team identified 5 , 450 households rearing a total of 11 , 364 pigs throughout Rajshahi ( 34% ) , Nawabgonj ( 13% ) and Naogaon ( 53% ) districts . More than 60% ( n = 6 , 963 ) of pigs were over 12 months of age and half of the total pig population were female . Of the 312 sampled pigs , 49% were female and all were a local breed . The mean age of sampled pigs was 23 months ( range 5–60 ) . Of the tested animals , 26 cattle ( 6 . 5% , 95% CI 4 . 3–9 . 4 ) , 17 goats ( 4 . 3% , 95% CI 2 . 5–6 . 7 ) and138 pigs ( 44 . 2% , 95% CI38 . 6–49 . 9 ) had antibodies against NiV soluble attachment glycoproteins ( NiVsG ) in the Luminex assay ( Table 1 ) . The NiVsG positive sera had a range of MFI values between 306 and 20 , 975 ( Figure 1 ) . A total of 39 NiVsG positive sera ( 9cattle , 2 goats and 28 pigs ) showing the highest MFI in Luminex assay were further tested by serum neutralization test against NiV . No neutralizing antibodies were detected . We also tested NiVsG positive sera from 3 cattle , 1 goat and 21 pig sera that reacted most strongly in Luminex assay by western blot . Antibodies against NiV N protein were detected in two cattle sera , one with an MFI value of 7365 and one with an MFI of 2537 and two pig sera ( Figure 2 ) . NiVsG positive sera along with 140 NiVsG negative sera ( 9 cattle , 13 goats and 118 pigs ) were tested for NiV antibodies using CDC's in-house ELISA . All specimens were negative for NiV antibodies by ELISA . A total of 25 NiVsG positive sera were tested for CedV antibodies in the Luminex assay . None showed significant binding for CedV . We identified NiV Luminex antibody positive animals from all study sites ( Table 2 and Figure 3 ) . The majority of NiV antibody positive cattle ( 92% ) and goats ( 94% ) were female ( Table 3 ) . During sample collection , 99% of animals were observed to be apparently healthy and all antibody positive animals had no apparent clinical signs of illness . In bivariate analyses , cattle and goats with NiVsG antibody levels above the chosen cutoffs were more likely to have a history of being fed partially bat and/or bird eaten-fruits ( PR = 3 . 9 , 95% CI 2–7 . 2 , p<0 . 001 ) , drinking raw juice prepared from bat and/or bird-eaten Asian Palmyra palm fruits ( Borassus flabellifer ) ( PR = 9 . 5 , 95% CI 5 . 2–17 . 4 , p<0 . 001 ) , grazing in areas exposed to roaming pig herds ( PR = 1 . 7 , 95% CI 0 . 6–4 . 3 , p = 0 . 3 ) , and living in fruit orchard areas ( PR = 1 . 7 , 95% CI 0 . 8–3 . 8 , p = 0 . 2 ) ( Table 4 ) . However , in multivariate analysis the two exposures that were independently associated were having a history of feeding on fruits partially eaten by bats or birds ( PR = 3 . 1 , 95% CI 1 . 6–5 . 7 , p = 0 . 001 ) and drinking of raw palmyra palm juice ( PR = 3 . 9 , 95% CI 1 . 5–10 . 2 , p = 0 . 004 ) ( Table 5 ) . Out of 800 cattle and goats , 2% ( n = 16 ) of animals were fed juice prepared from partially bats and/or birds-eaten Asian Palmyra palm fruit by their owners . There was no significant difference in pig NiVsG seroprevalence between backyard and nomadic rearing systems ( 20% in backyard vs . 15% in nomadic herds , p = 0 . 4 ) .
This study identified antibodies against NiVsG in 26 cattle , 17 goats and 138 pigs; however these antibodies did not neutralize NiV , and did not react against NiV antigens in an ELISA , though 2 cattle and 2 pig sera reacted with NiV N protein by WB . Animals that were fed fruit that had been partially eaten by bats or birds were >3 times more likely to have antibodies against NiVsG compared with animals not fed partially eaten fruit . The serological response in these domestic animals suggests they were likely infected with a henipavirus . The positive test results on two different diagnostic platforms targeting two different NiV proteins ( sG and N ) , but negative SNT results and the association with bat bitten fruit suggests that the animals were likely infected with a non-Nipah henipavirus . Cedar virus ( CedV ) is the only non-Nipah non-Hendra henipavirus to have been isolated and fully described [29] , yet there is evidence of considerable diversity of henipaviruses . Samples from 6 bat species in 5 different African countries identified RNA sequence of paramyxovirus L gene suggestive of 19 novel non-Nipah non-Hendra henipaviruses [30] . Three additional novel henipaviruses have been identified by sequencing nucleic acid of the paramyxovirus large gene from Pteropus giganteus , the putative bat reservoir of NiV in Bangladesh [31] . The virus ( or viruses ) detected here appear to be more closely related to NiV than HeV , as measured by cross-reactive antibodies specific for NiVsG . Phylogenetic analysis of NiV isolates from Malaysia and Bangladesh suggest that strains of NiV transmitted from bats to humans were genetically diverse , however all isolated viruses from animals and humans in these two countries show full cross-neutralizing antibodies [32]–[34] . While studies on African bats have showed antigen-antibody reactions to henipaviruses in the Luminex assay , and cross-neutralization of HeV and NiV in serum neutralization tests [35] , [36] , studies in Vietnam on bats and in Ghana on pigs showed similar types of antigen-antibody reactions of henipaviruses in the Luminex assay without cross neutralization , similar to what we identified in domestic animals in Bangladesh [37] , [38] . Cedar virus , detected in Australian fruit bats , is also not cross-neutralizing with HeV or NiV and has limited cross-reactivity in the Luminex sG binding assays [29] . Finally , in India some individual Pteropus bats have shown antibodies that cross-neutralized Nipah and Hendra virus [39] . Taken together these observations suggests that there is a spectrum of henipavirus strains circulating , with differing levels of antibody cross-reactivity . Challenges associated with assessing serological responses to an uncharacterized virus were mitigated here by using a Bayesian mixture model approach , which enables the assay output to be assessed in its own right , without the need to compare it to an alternative assay [24] . These analyses strongly supported cutoffs of MFI = 300 for cattle and goats and MFI = 650 for pigs as being very conservative ( individuals >99% likely to be seropositive ) ( Details in the supporting material ) . Fruit bats can contaminate fruits , grasses or other plants with henipaviruses through their excretions and secretions . Epidemiological findings from multiple HeV outbreaks in Australia suggested that the horse index cases were likely to have been exposed via feeding in paddocks containing fruit trees frequented by fruit bats and thereby contaminated with HeV [40] . In our study , animal owners reared animals mainly in the backyard ( ≈99% ) and 73% of these animals were fed with a semi-intensive feeding system . Pteropid bats visit fruit trees as part of their nightly foraging activities , and sometimes drop partially eaten fruits to the ground [41] , [42] . Nipah virus RNA has been detected from urine and throat swab samples collected from P . giganteus in Bangladesh [31] from fruit partially eaten by P . hypomelanus and P . vampyrus in Malaysia [41] . As the domestic animals in this study were scavenging for a portion of their daily feeding time , they could have been exposed to dropped fruits or an environment contaminated with bat excreta , which might increase the risk of henipa-like virus transmission from bats to these animals . In this study , animal owners reported that sometimes they offered dropped fruits as foods to their animals . A few animal owners also reported that they prepared fresh juice from intact Asian Palmyra palm fruit for themselves and they used Palmyra palm fruits partially eaten by bats and/or birds for their animals . The association between exposure to bat-contaminated feeding exposure and presence of antibodies detected by Luminex assay against NiVsG proteins in livestock animals suggests that P . giganteus bats , the reservoir species of NiV , or a related frugivorous bat species such as Cynopterus sphinx or Rousettus leischenaulti – both common in Bangladesh and observed to have similar foraging patterns with P . giganteus [43] , could be the source of infection that resulted in the generation of these antibodies . Henipaviruses can infect a wide variety of animal species including humans [4] , [44]–[48] . This is consistent with the ability of the virus to infect a wide range of mammals by exploiting the very well conserved ephrin B2 and ephrin B3 receptor [49]–[51] . In Malaysia , antibodies against NiV were detected in goats , dogs , cats and horses during a human Nipah outbreak that suggests a wide range of animal species were exposed and infected with NiV [44]–[46] . Pigs were identified as the most frequently infected domestic animal hosts and they transmitted infection from bats to humans as an intermediate host [4] , [5] . In this study , our data also suggest pigs were more likely to be exposed to henipaviruses than cattle and goats . The high rate of seropositivity in pigs could be due to the frequent exposure and/or their high susceptibility to henipavirus infection . Alternatively , this may represent a henipavirus that has adapted to and developed a reservoir in swine . Swine in Malaysia and in Ghana have evidence of susceptibility to henipavirus infection [5] , [38] . We don't know whether other henipaviruses are infecting human populations , but further investigation in bats , domestic animals and people may further clarify henipavirus ecology in Bangladesh and globally . This serological study of healthy animals provides little insight on the clinical consequences of these infections . All antibody positive animals were apparently healthy during sample collection , but they may have had signs of disease earlier . Moreover , animals with severe illness may have died before sampling . Further studies in sick animals would be necessary to evaluate the association of these non-Nipah henipaviruses with illness . Laboratory findings suggest cattle , goats and pigs were exposed to a novel virus or viruses with antigenic similarity to NiV . The association of antibody positive findings by Luminex assay in cattle and goats with exposures to potentially bat-contaminated foods suggests that the source of this virus is likely frugivorous bats . Further research should be undertaken to characterize the range of henipaviruses spilling over from bats to domestic animals because of their potential animal health and human health importance . | Nipah virus ( NiV ) , is an emerging disease that causes severe encephalitis and respiratory illness in humans . Pigs were identified as an intermediate host for NiV transmission in Malaysia , and in Bangladesh three NiV outbreak investigations since 2001 identified an epidemiological association between close contact with sick or dead animals and human illness . We collected samples from cattle and goats reared around Pteropus bat roosts in human NiV outbreak areas in Bangladesh , and tested pig sera collected for a Japanese encephalitis study . We detected antibodies against NiV glycoprotein in 26 ( 6 . 5% ) cattle , 17 ( 4 . 3% ) goats and 138 ( 44 . 2% ) pigs by a Luminex-based multiplexed microsphere assay , but none were virus neutralizing . There may have been exposure of Luminex positive animals to one or more novel viruses with antigenic similarity to NiV . Further research may identify a novel organism of public health importance . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"public",
"and",
"occupational",
"health",
"research",
"design",
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"animal",
"studies",
"epidemiology",
"tropical",
"diseases",
"research",
"and",
"analysis",
"methods"
] | 2014 | Serological Evidence of Henipavirus Exposure in Cattle, Goats and Pigs in Bangladesh |
Echinococcosis is a neglected zoonotic disease affecting over 1 million people worldwide at any given time . It is the leading cause of hospital admissions for parasitic diseases in Chile . We conducted a retrospective investigation of hospitalized cases to describe the epidemiological trends of echinococcosis in Chile . We also examined the potential environmental risk factors for echinococcosis hospitalization rates . Through nation-wide hospital discharge data , a total of 11 , 516 hospitalized patients with cystic echinococcosis were identified between January 2001 and December 2012 . The mean age of hospitalization was 40 years , with notable gender difference in pediatric patients . The hospitalization rate was found to be overall steadily decreasing from 2001 ( 7 . 02 per 100 , 000 ) to 2012 ( 4 . 53 per 100 , 000 ) with a 5% decrease per year ( rate ratio = 0 . 95 [95% CI: 0 . 94 , 0 . 96] ) . The hospitalization rate was higher in the south of Chile compared to the north . Goat density and intermediate precipitation were found to be significantly positively associated with the hospitalization rate while annual average temperature was found to be significantly negatively associated with the hospitalization rate . Findings of this study indicate that echinococcosis is still an important public health burden in Chile related to interaction with livestock and climate . Efforts should be placed on targeted prevention measures for farmers and raising awareness of echinococcosis among health care workers .
Echinococcosis is a neglected zoonotic disease , which affects over 1 million people worldwide at any given time and the loss of 1 to 3 million disability-adjusted life years annually [1] . It is caused by parasitic tapeworms of the genus Echinococcus . The two species of clinical and public health importance are E . granulosus , related to cystic echinococcosis ( CE ) , and E . multilocularis , related to alveolar echinococcosis ( AE ) [1–2] . CE affects livestock production , and it is estimated that CE treatment in humans and losses in livestock cost the global economy 3 billion USD every year [1] . Although the cyst of the parasite is slow-growing in the human body , infected people may face debilitating and life-threatening symptoms such as abdominal or chest pain , coughing , vomiting or allergic reactions which require complicated treatment with poor prognosis ( post-operative death rate of 2 . 2% and post-treatment relapse of 6 . 5% ) [1–2] . CE is common in Latin America , where over 2 , 000 cases are reported annually in Argentina , Brazil , Chile , Uruguay , and Peru [2] . In Chile , echinococcosis has been classified as a mandatory notifiable disease since 2000 [3] and is the leading cause of hospital admissions for parasitic diseases [4–6] . The estimated hospital discharge rate associated with echinococcosis at the national level is estimated at 4 . 7–5 . 0 cases per 100 , 000 population , and approximately 2 cases per 100 , 000 inhabitants require surgical treatment , mainly affecting the working age group [7–8] . To date , AE has never been reported in Chile [6] . The lifecycle of E . granulosus is maintained in a dog-livestock-dog cycle . Humans are accidental hosts and are infected via ingestion of embryonated eggs through the environment , food or direct contact with animals [1–2] . Survival of the eggs in the environment , and thus human transmission , is highly impacted by a number of environmental and anthropogenic factors , including climate [9–10] . The presence of large number of dogs harboring the parasite , allowing dogs to feed on uncooked innards or entrails from sheep or cattle , inadequate facilities for slaughter and home slaughtering and consumption have all also been shown to be risk factors for persistence or emergence of CE [7 , 11–12] . There are currently no official national programs for the control of E . granulosus in Chile , despite efforts in previous years [8] . Chile first established a CE control program in 1979 , which was carried out by the Livestock and Agricultural Service . This program entailed routine deworming of domestic dogs with praziquantel eight times a year in the administrative regions of Aisén and Magallanes ( S1 Fig ) . However , due to implementation costs , the frequency of administrating praziquantel was reduced to twice a year . Overall , the program led to a 60–70% reduction in CE prevalence in sheep and dogs over a 27-year period ( 60% to 0 . 7% and 71% to 0 . 5% from 1978 to 2004 , respectively ) [7–8 , 13–14] . The program was dismantled in 2004 . Subsequently , deworming has been voluntarily carried out by farmers , or through a few local campaigns . Since 2015 , four different control programs have been initiated again in the regions of Coquimbo , Bío Bío , Araucanía and Aisén ( S1 Fig ) [7–8 , 13–14] . In the absence of adequate control programs , there is a risk for re-emergence of CE . Since the national program was terminated in 2004 , the change in hospitalization rate for echinococcosis has not been documented . In addition , as Chile starts implementing vaccination programs for CE in animals , surveillance data for CE is important to conclude the impact of its vaccination program in human populations . The objective of our study was to describe the epidemiological trends of hospitalized echinococcosis cases between 2001 and 2012 , and investigate the impact of potential risk factors on differences in hospitalization rates between provinces .
This study involved human subjects and no consent was given . Hospitalization records were deidentified and analyzed anonymously . Patient data could not be linked directly or indirectly to identifiable individuals . A retrospective study was conducted nation-wide from hospitalized cases diagnosed with CE and reported to the Ministry of Health from January 1st 2001 to December 31st 2012 . There are 15 administrative regions in Chile , and hospital discharge data were collected from public and private hospitals in each region ( S1 Fig ) . Patients were diagnosed with CE if they had at least one positive imaging test ( ultrasound , NMR , CAT or X-ray ) and one positive serologic test ( in-house IgG ELISA as screening test and in-house IgG , IgM and IgA Western Blot as confirmatory test ) . Samples were tested at the Public Health Institute of Chile . Cases were coded with the International Classification of Diseases 10th revision ( ICD-10 ) which identified the causative species of echinococcosis and the location of lesions in the body ( B670 to B679 ) . Patient information such as age , sex , municipality/region of residence , date of admission , date of discharge , duration of hospitalization and state of discharge were also provided . Livestock , demographic and climate data were obtained from Chile census data [15] . Annual regional population figures were obtained from population census data of Chile , defined as the populations of June 30th of each year [15] . Demographic and clinical variables were described using proportions for categorical variables and mean or median for continuous variables . Differences in proportions , mean and median were tested with the Chi-squared test , student t-test and Wilcoxon rank test , respectively . Annual regional hospitalization rates ( per 100 , 000 ) were calculated as the number of CE-related hospitalizations over the total population in the region in a given year . Crude and adjusted annual hospitalization rates were calculated , where rates were sex and age adjusted through direct standardization using the population demographics from 2007 as reference . Livestock density was defined as the number of heads per square kilometer . Annual average temperature was defined as the average temperature during a one-year period in degree Celsius , and total annual precipitation was defined as the total precipitation during a one-year period in millimeters . Trends in hospitalization rates were analyzed by negative binomial regression for the whole study population . Negative binomial regressions were also used to identify correlates of differences in hospitalization rates between provinces for the year 2007 to allow for more input data . Candidate variables were livestock density and climate data from the year 2007 ( year at which the livestock census was done ) gathered at the provincial level . A total of 11 variables related to climate and animals were evaluated for their association with CE at one point in time . Precipitation was best fit to the data using a generalized linear model with cubic splines to capture non-linear effects . The Akaike Information Criterion ( AIC ) was calculated for each regression and each combination of input variables ( S1 Table ) . We identified 10 models with an AIC score within a difference of two units from the lowest AIC score , indicating similar fits to the model . Ultimately , the most parsimonious model was chosen and included the following three variables: goat density , annual average temperature and annual average precipitation . A log-likelihood ratio test was performed to evaluate the inclusion of precipitation in the final model and revealed statistically significant ( p = 0 . 042 ) . This model was then adjusted for age and sex categories , and the variables remained significant . These variables were used to fit a log-linear regression , to explain the difference in hospitalization rates between the two 2001 and 2007 data points . This time , livestock and climate data from 1997 and 2007 were used . All analyses were conducted using R software version 3 . 1 . 3 [16] .
A total of 11 , 516 hospitalized patients with CE were discharged between 2001 and 2012 ( Table 1 ) . Slightly more than half of patients were male ( 52 . 95% , 6098/11516 ) . The mean age of cases was 40 . 1 years ( standard deviation 21 . 1 ) . The duration of hospital stay varied between 1 and 292 days with a median of 8 ( interquartile range 4–16 ) days . When classifying the data into age groups based on likely workforce participation , 15 . 2% ( 1746/11516 ) of cases were 0–14 years-old , 65 . 9% ( 7586/11516 ) of cases were 15–59 years-old and 19 . 0% ( 2184/11516 ) were 60 years and above . Demographic data and clinical data are presented in Table 1 . When categorizing the locations of cysts by age group , a higher proportion of patients aged 15 years and older had liver cysts ( 46 to 47% ) , whereas 34% of patients aged 0–14 years had liver cysts ( p<0 . 01 ) . Conversely , a higher proportion of pediatric cases had lung cysts ( 12% ) compared to patients aged 15 years or older ( 6 to 7% ) ( p<0 . 01 ) ( S2 Table ) . The annual hospitalization rate for CE over 12 years was 5 . 84 per 100 , 000 populations ( Table 2 ) . It has been found to be overall steadily decreasing from 2001 ( HR = 7 . 02 per 100 , 000 ) to 2012 ( HR = 4 . 53 per 100 , 000 ) with a 5% decrease ( rate ratio = 0 . 95 [95% CI: 0 . 94 , 0 . 96] ) calculated by negative binomial model ) . It has been decreasing for both sex and in all age groups ( Table 2 ) . The hospitalization rate was consistently higher in the southern regions of Chile ( Aisén and Magallanes ) and central Chile ( Coquimbo ) compared to the northern regions of Chile ( Fig 1 ) . The hospitalization rate for CE in northern Chile was low over the years , although higher hospitalization rate was observed in some years in the Arica y Parinacota and Los Ríos regions . Temporal and spatial patterns of hospitalization rates are described in Table 2 and in Fig 1 . Goat density was associated with higher hospitalization rates ( rate ratio per 10 goats/sqkm = 3 . 40 [1 . 58 , 7 . 97] ) while annual average temperature was associated with lower hospitalization rates ( rate ratio = 0 . 70 [0 . 60 , 0 . 82] ) . Among our observation points , precipitation was significantly associated with higher hospitalization rates for intermediate precipitation levels . The curve of predicted incidence as a function of precipitation is presented in Fig 2 . Changes in these variables were not predictive of long-term trends in hospitalization rates between 2001 and 2007 . Regression results are presented in Table 3 . When adjusting for age and sex categories , the impact of each variable on the hospitalization rate stayed similar and did not change our findings ( rate ratio per 10 goats/sqkm = 4 . 11 [2 . 34 , 7 . 42] , rate ratio for annual average temperature = 0 . 71 [0 . 64 , 0 . 79] and the predictive curve for annual precipitation was similar and overall statistically significant ) .
This study is the first to model echinococcosis in Chile , and one of the firsts in Latin America . Although reports on CE prevalence and incidence have previously been reported in Latin American countries , this comprehensive analysis linking human CE cases , host species and environmental factors will provide a basis for approaching CE prevention and control holistically . This study demonstrated that the hospitalization rates for CE remain high in Chile compared to previous estimates , despite a general decrease over 12 years [7–8] . The majority of cases were found in people of working age , which is consistent with previous findings [5–6] . Although the overall gender difference was small , pediatric cases aged 0 to 14 years were more likely to be male than female ( p<0 . 01 ) . This may be due to behavioral differences between boys and girls , where boys have more exposure to sources of infection such as soil and carcasses . Furthermore , the relationship between patients’ age group and location of cysts suggested that there was a significant difference between the location of the cyst in pediatric and adult cases . Inconsistent findings from previous literature suggest that pediatric cases are more likely to have liver or lung cysts compared to adults , depending on the study [17–18] . In our study , pediatric cases had a high proportion of pulmonary cysts compared to adults , which may be attributed to the slower development rate of liver cysts compared to lung cysts , delaying health seeking behavior for patients with liver cysts in pediatric years [19–20] . The highest hospitalization rates were observed in the southern regions , the highest being in Aisén . This geographical difference between the north and south may be attributed to agricultural characteristics . Southern Chile is highly agricultural , specializing in cattle and sheep husbandry , whereas horticulture and raising camelids are common in the north . It is plausible that southern areas with higher caprine production have higher rates of hospitalization . Another possibility is that hospitalization rates appear higher in rural areas with smaller populations , as one new case may dramatically elevate the rate in the region . In addition to this , we observed a somewhat high hospitalization rate in the Coquimbo region in the central-north part of Chile . One possible explanation is that goat production is common in Coquimbo . This is further supported by our regression model , where higher goat density was significantly associated with a higher hospitalization rate for CE . We found that the density of goats , temperature and precipitation are important factors that are associated with changes in CE hospitalization rates in Chile . We considered the density of livestock in the model because high density of goats , for example , indicate more intensive goat husbandry , increased risk for dogs to be in contact with infected goat’s viscera , and increased risk for people to be infected by E . granulosus eggs from dog feces . Goats appear to be an important host for the maintenance and transmission of Echinococcus granulosus in Chile . This is somewhat unexpected as sheep are known to be the main risk factor for CE transmission [21–22] , but might be explained by the fact that goats are host to more strains of E . granulosus [23] . Further strain typing data of Echinococcus cysts in humans and animals would augment our findings in elucidating the epidemiological link between species . In addition , slaughterhouse data from 2014 indicate a higher cumulative incidence of CE in goats compared to cattle , sheep and horses ( 305 cases per 1 , 000 goats , 200 cases per 1 , 000 cattle , 31 cases per 1 , 000 sheep; 12 cases per 1 , 000 horses ) [8] . A lower occurrence of CE in sheep and cattle compared to goats may explain why sheep and cattle were not found to be important risk factors in our regression model . Given that people do not become infected with echinococcosis by simply handling or manipulating infected goats or sheep , it is possible that improper handling of goat viscera during and after slaughter increases the risk of dogs , the definitive host , to become infected . This will , in turn , increase people’s risk to become infected through contact with feces of E . granulosus infected dogs . Although intervention programs in Chile did not target goats in the past , future control measures may be more effective if goats were also targeted . In addition , lower temperatures and medium range precipitation in Chile are risk factors for CE hospitalizations according to our model . This is consistent with the high hospitalization rate observed in southern Chile , which is characterized with an oceanic climate . It is also supported by previous reports on the survival of the E . granulosus eggs in the environment [9–10 , 24] . Low temperatures and humid soil allow for the eggs to survive longer in the environment . Moreover , E . granulosus eggs are sensitive to high temperatures and low humidity as it leads to loss of infectivity [25] . These are critical factors that increase the risk of infection of livestock and humans and thus the number of hospitalizations . It is also worth noting that areas in Chile with extremely low temperature and low precipitation may also be regions with fewer sheep and goats , since these animals are not able to survive under extreme climates . Changes in climate variables were not predictive of long term hospitalization trends likely due to the time lag between time of infection , presentation of clinical signs and health-care seeking behavior . The association between CE hospitalization rates and the variables that were found to be significant in our model are ecological in nature and cannot be causally interpreted . The overall decrease in the rate of hospitalizations related to CE could be attributed to national prevention measures and heightened awareness of echinococcosis in Chile since 1979 [13–14] . While large-scale programs at the national level are ideal , resources should be allocated to local , targeted programs in order to decrease transmission . We recommend that the Ministry of Health focus their prevention efforts on farm workers in regions with high hospitalization rates through workshops and other methods of education . Focusing on children’s education in school may be effective in guiding the population to interact with their domestic animals appropriately , including washing hands after grooming them , avoiding kissing them on the mouth , and deworming their dogs every 2 to 3 months . Furthermore , our finding that temperature and precipitation factors are associated with CE hospitalization could be utilized to reinforce targeted prevention measures in areas where climatic conditions are conducive to E . granulosus survival and transmission or husbandry of intermediate hosts such as sheep and goats . In Chile , this translates to implementing adapted deworming in areas with high concentration of echinococcosis cases , especially in southern Chile and Coquimbo . Our findings should be interpreted in light of some limitations . First , the outcome that was used to measure CE was the number of hospitalizations . Trend in hospitalization rates does not necessarily reflect trend in disease incidence , but rater hospital seeking behavior in people who are infected with CE , which could be affected by various factors such as severe climate and livestock health . It is also possible for one person to be counted more than once if the patient was hospitalized multiple times for CE . In addition , hospitalization rates may have decreased due to better outpatient care , while the incidence itself stayed constant . However , the ICD-10 coded hospitalization records are the most accurate and consistent data currently available in Chile , and other measures of disease are deemed inaccurate and inconsistent by Chilean public health officials due to significant underreporting . Second , hospitalization rates analyzed by each administrative region and provinces may not perfectly reflect the true incidence of the region . This is because rural dwellers may seek health care in larger cities and the true incidence in rural regions may be higher . Additionally , echinococcosis is a disease developing over the course of several years implying a lag between infection and hospitalization that our model could not take into account . As no data is available on incubation period , we chose to use livestock and climate data from the same hospitalization year as a proxy . Our guess is that the association would be even stronger if we had used data from several years before hospitalization . In order to evaluate the risk of re-emergence of echinococcosis in Chile , targeted studies on high-risk populations such as agricultural workers and veterinarians and their interaction with host animals are required . Studies could investigate the lag between extreme climatic events and epidemics of echinococcosis . Furthermore , examining risk factors in a cohort or case-control study will be beneficial to validate findings from cross-sectional studies on echinococcosis in Latin America . Finally , to confirm the role of livestock and climate variables in hospitalization rates , studies at a smaller geographical scale are recommended for risk analysis . Improved surveillance will lead to a better understanding of the trends and transmission dynamics of echinococcosis , which will allow for effective prevention and control measures . Although CE is a notifiable disease in Chile , the number of notified cases to the Ministry of Health was much lower than the number of hospitalized cases from 2001 to 2012 ( 4 , 288 versus 11 , 673 cases ) . The number of hospitalized case is only partially represented by the number of notified cases , which demonstrates the need for collecting data and encouraging laboratories , clinicians and hospitals to report any suspected case of E . granulosus infection . Raising awareness of echinococcosis among health care workers may be key to increased reporting in the future . Moreover , we recommend that the diagnostic method also be collected as part of patients’ records , since it is currently unavailable and could be useful for future surveillance . In conclusion , echinococcosis is a zoonosis with a significant public health burden in Chile , related to farming and interaction with livestock , and also environmental conditions such as temperature and precipitation . Long-term efforts should be placed on CE in Chile , ensuring that cases are notified to the Ministry of Health and targeted intervention programs are implemented to reduce the hospitalization rate of the disease . | Humans are infected by many types of parasites that originate from animals . Cystic echinococcosis ( CE ) is a tapeworm infection that affects over 1 million people over the world at any given time , with symptoms such as abdominal pain , chest pain , vomiting , and allergic reactions . The lifecycle of this parasite is maintained between dogs and livestock , and humans are ‘accidentally’ infected through ingesting parasite eggs . This makes the control of CE difficult for public health practitioners . Attempts have been made to better understand which animals or environmental factors increase or decrease the number of human CE infections , but mainly with data from only one point in time . In this study , we studied the trends of this disease in Chile , a country with many cases of CE , over a 12-year period . Then , using a statistical model , we examined the risk factors related to animals and the environment that contribute to a higher rate of CE hospitalizations . We found that higher goat density , intermediate precipitation level and low temperatures are important factors associated with an increase in hospitalization rates related to CE in Chile . This new knowledge contributes to a better overall understanding of the trends and risk factors of CE , which will help to implement targeted prevention and control strategies in regions where the disease is common . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"hospitalizations",
"livestock",
"medicine",
"and",
"health",
"sciences",
"chile",
"(country)",
"ruminants",
"tropical",
"diseases",
"geographical",
"locations",
"vertebrates",
"parasitic",
"diseases",
"animals",
"mammals",
"health",
"care",
"neglected",
"tropical",
"dise... | 2017 | Trends and correlates of cystic echinococcosis in Chile: 2001–2012 |
Lytic herpes simplex virus 1 ( HSV-1 ) infection triggers disruption of transcription termination ( DoTT ) of most cellular genes , resulting in extensive intergenic transcription . Similarly , cellular stress responses lead to gene-specific transcription downstream of genes ( DoG ) . In this study , we performed a detailed comparison of DoTT/DoG transcription between HSV-1 infection , salt and heat stress in primary human fibroblasts using 4sU-seq and ATAC-seq . Although DoTT at late times of HSV-1 infection was substantially more prominent than DoG transcription in salt and heat stress , poly ( A ) read-through due to DoTT/DoG transcription and affected genes were significantly correlated between all three conditions , in particular at earlier times of infection . We speculate that HSV-1 either directly usurps a cellular stress response or disrupts the transcription termination machinery in other ways but with similar consequences . In contrast to previous reports , we found that inhibition of Ca2+ signaling by BAPTA-AM did not specifically inhibit DoG transcription but globally impaired transcription . Most importantly , HSV-1-induced DoTT , but not stress-induced DoG transcription , was accompanied by a strong increase in open chromatin downstream of the affected poly ( A ) sites . In its extent and kinetics , downstream open chromatin essentially matched the poly ( A ) read-through transcription . We show that this does not cause but rather requires DoTT as well as high levels of transcription into the genomic regions downstream of genes . This raises intriguing new questions regarding the role of histone repositioning in the wake of RNA Polymerase II passage downstream of impaired poly ( A ) site recognition .
Transcription termination is an essential process in gene expression that is coupled to all parts of RNA metabolism including transcription initiation , splicing , nuclear export and translation ( reviewed in [1 , 2] ) . It results in the release of RNA polymerase II ( Pol II ) and the nascent transcript from the chromatin , determines the general fate of individual transcripts and plays a crucial role in limiting the extent of pervasive transcription of the genome . Herpes simplex virus 1 ( HSV-1 ) efficiently modulates cellular RNA metabolism and both cellular and viral gene expression to facilitate lytic infection [3–9] . Using 4-thiouridine- ( 4sU ) -tagging followed by sequencing ( 4sU-seq ) , we recently reported that lytic HSV-1 infection results in the disruption of transcription termination ( DoTT ) of the majority but not all cellular genes [10] . This was dependent on de novo protein synthesis and already became broadly detectable by 2-3h of infection , which is before the release of the first newly generated virus particles at around 4h post infection ( p . i . ) . At 7-8h p . i . , about 50% of all 4sU-seq sequencing reads mapping to the human genome originated from intergenic regions ( compared to <10% in uninfected cells ) . Previously , we referred to transcription beyond poly ( A ) sites due to DoTT as ‘read-out’ . As this term has led to confusion , we now use the term ‘read-through’ to refer to transcription that extends beyond poly ( A ) sites . Transcription into a downstream gene arising from read-through from an upstream gene is referred to as ‘read-in’ . For more than half of expressed cellular genes , poly ( A ) read-through affected >35% of their transcription . Read-in transcription into downstream genes was responsible for the seeming induction of about 1 , 100 cellular protein-coding and non-coding genes late in infection . In addition , it resulted in chimeric transcripts spanning two or more genes as evidenced by intergenic splicing events that connect exons of neighboring cellular genes . Subsequently , two other studies reported on the disruption of transcription termination in cellular stress responses and cancer [11 , 12] . Transcription downstream of genes ( DoG ) was observed in the osmotic stress response in human neuroblastoma cells , which was independent of de novo protein synthesis but appeared to at least partially rely on inositol-1 , 4 , 5-trisphosphate receptor ( IP3R ) activation and calcium signaling [11] . In addition , pervasive transcription read-through was identified in renal cell carcinoma [12] . This was dependent on the loss of histone methyltransferase SETD2 , consistent with the role of epigenetic factors in RNA processing . Similar to HSV-1 infection , novel RNA chimeras were observed . Invasion of oncogenes by polymerases that initiated at upstream genes indicated a novel link between aberrant expression of oncogenes and chimeric transcripts prevalent in cancer . Taken together , these findings raise important questions regarding the underlying molecular mechanisms and functional roles of DoTT/DoG transcription in HSV-1 infection , cellular stress responses and cancer . DoG transcription during osmotic stress was identified by Vilborg et al . upon exposure to 80mM KCl for 1h ( from now on referred to as ‘salt stress’ ) in a human neuroblastoma cell line ( SK-N-BE ( 2 ) C ) by RNA-seq on nuclear , RiboMinus-treated RNA [11] . This revealed about 2 , 000 human genes to be affected . In addition , DoG transcription was also observed following heat stress ( 44°C ) [11] . Recently , Vilborg et al . also reported on DoG transcription upon oxidative stress and found significant similarities but also clear stress-specific differences between the three stressors [13] . In our primary study , we analyzed newly transcribed RNA purified using 4sU-seq in one hour intervals of the first 8h of lytic HSV-1 infection of primary human foreskin fibroblasts ( HFF ) ( Fig 1A ) . Under these conditions , the HSV-1 infected cells only start to lyse around 16 to 24h of infection . This allowed us to directly assess and quantify the relative frequency of transcripts experiencing DoTT as well as the extent of read-through transcription occurring within one hour intervals during the first eight hours of infection [10] . Throughout this manuscript , we refer to HSV-1-induced disruption of transcription termination as ‘DoTT’ to differentiate it from stress-induced DoG transcription . It is important to note here that transcription in intergenic regions downstream of genes was almost exclusively observed on the sense strand in relation to the upstream gene . This clearly distinguishes read-through from the recently reported activation of antisense transcription of the host genome during lytic HSV-1 infection [14] . Although DoTT was much more prominent at late times ( 7-8h p . i . ) of HSV-1 infection than in salt or heat stress , we wondered whether the two phenomena might reflect the same cellular mechanism . We thus performed a detailed comparison and characterization of HSV-1-induced DoTT and DoG transcription triggered by salt and heat stress using 4sU-seq in the same cell type , namely HFF . This showed clear similarities in read-through between HSV-1 infection and the different stresses but also clear context- and condition-specific differences . Furthermore , we performed ATAC-seq ( transposase-accessible chromatin using sequencing [15] ) to compare chromatin accessibility before and during HSV-1 infection and stress . Strikingly , HSV-1-induced DoTT was accompanied by a strong increase in chromatin accessibility downstream of the affected poly ( A ) sites , which essentially matched the region of read-through transcription . This did not cause but rather required DoTT as well as a high level of transcriptional activity into downstream genomic regions . Interestingly , this effect was specific to HSV-1 and not observed in salt or heat stress ( up to 2h ) indicating that other mechanisms by which HSV-1 perturbs RNA processing contribute to this unexpected gene-specific alteration in the host chromatin landscape .
To directly compare HSV-1-induced DoTT with DoG transcription during cellular stress responses , we performed 4sU-seq analysis ( 60min 4sU-tagging followed by RNA sequencing ) of HFF exposed to either salt ( 80mM KCl ) or heat stress ( 44°C ) for 1 and 2h ( see Fig 1B ) . Two biological replicates of each condition as well as 2 untreated samples for each stressor were analyzed . 4sU-seq data for the first 8h of HSV-1 infection in HFF were obtained from our previous study [10] . A visual inspection of mapped reads for marker genes with either strong ( SRSF3 , SRSF6 ) or no ( GAPDH , ACTB ) DoTT/DoG transcription already indicated a striking similarity between presence or absence of DoTT/DoG transcription in the three conditions ( Fig 1C and 1D; Fig A in S3 File , links to UCSC genome browser sessions showing read coverage for all cellular genes and samples separately for both replicates can be found at www . bio . ifi . lmu . de/HSV-1 ) . As previously reported for HSV-1 infection [10] ( Fig 1E ) , the percentage of reads mapping to intergenic regions downstream of gene 3’ ends increased substantially during salt and heat stress in HFF ( Fig 1F ) . Intergenic read counts were highest directly downstream of gene 3’ ends and gradually decreased with increasing distance to gene 3’ ends . Furthermore , downstream intergenic transcription occurred almost exclusively in the same orientation as the upstream gene in all conditions ( Fig B in S3 File ) . The low levels of antisense reads downstream of genes increased with increasing distance from gene 3’ ends as a consequence of read-through transcription for genes expressed from the opposite DNA strand outside of the 100kb downstream window considered . The gradual decrease in read levels downstream of genes was not due to differences in the length of read-through between genes , but was also observed at the level of individual genes ( Fig B in S3 File and Fig C in S3 File ) . It could be approximated reasonably well by a linear fit at least late in HSV-1 infection and at 2h salt and heat stress , but the slope of the linear fit differed between genes ( Fig C in S3 File ) . As a consequence of this gradual decrease and in contrast to regular mRNAs , 3’ ends of poly ( A ) read-through transcripts are not clearly defined [10 , 11] . As the extent of read-through for individual genes gradually increased throughout infection , read-through transcripts extended further and further downstream of the gene . To compare the extent of DoTT/DoG transcription between the three conditions , we focused on the 9 , 404 protein-coding and lincRNA ( long intergenic non-coding RNA ) genes whose expression was well detectable ( fragments per kilobase of exons per million mapped reads ( FPKM ) ≥1 ) in all uninfected/untreated 4sU-seq samples . We then applied our previous approach [10] of dividing expression in the 5kb downstream of genes by the gene expression ( FPKM ) value ( see methods ) . This measure ( denoted as percentage of downstream transcription ) is independent of any normalization to sequencing depth , which is canceled out in the division . As 4sU-tagging provides newly transcribed RNA from defined intervals of infection and stress , the obtained ratios quantify the percentage of transcripts newly transcribed in this interval that experience poly ( A ) read-through . To avoid confounding effects due to transcription of neighboring genes , we only included genes separated from neighboring genes on the same strand by at least 5kb on either side ( 5 , 928 genes ) . Although the restriction to the first 5kb downstream of a gene is relatively arbitrary , using a larger window of e . g . 10kb resulted in highly correlated values of downstream transcription ( Spearman correlation Rs>0 . 95 ) but would exclude an additional 737 genes ( 12 . 4% ) from the analysis . To account for small levels of downstream transcription in uninfected and untreated cells ( mean = 4 . 2% and 0 . 06% , respectively ) , we calculated read-through as the difference in the percentage of downstream transcription between infected/stressed and uninfected/untreated samples ( see methods ) . Read-in was quantified in the same way by first quantifying transcription in the 5kb upstream of genes relative to gene expression and then subtracting levels in uninfected/untreated samples . Since our previous study indicated that genes with read-in were more prone to read-through , we only used genes for the comparative analysis with at most 10% read-in in both HSV-1 infection and salt and heat stress ( 3 , 682 genes , Table A in S1 File ) . With the exception of the first three hours of HSV-1 infection where DoTT was hardly detectable , read-through values were highly correlated between replicates ( Fig D in S3 File; Rs≥0 . 85 ) . The induction of DoG transcription upon salt and heat stress was reflected in median read-through levels of 6 to 15% ( Fig 2A; Fig D in S3 File for individual replicates ) . Consistent with the recent report by Vilborg et al . [13] , global read-through levels peaked at 1h of salt stress , but required 2h to reach comparable levels in heat stress . At the highest level , read-through in both salt and heat stress was comparable to read-through at 4-5h post HSV-1 infection , but considerably lower ( ~3-fold ) than at the end of our HSV-1 infection time-course ( 7-8h p . i . ) . Median read-through levels in all conditions were highly correlated ( Rs = 0 . 99 ) to the overall perturbation of gene expression ( measured as standard deviation of FPKM log2 fold-changes; Fig 2B ) . Here , results for salt and heat stress fitted very well to a curve estimated from our HSV-1 time-course . At single gene level , however , read-through showed only a weak positive correlation with fold-changes in gene expression for HSV-1 infection ( after the first 3h ) , salt and heat stress ( Fig E in S3 File; Rs≤ 0 . 37 ) . Vilborg et al . [13] also only found weak correlations between fold-changes in DoG transcription and fold-changes in expression of the respective genes ( Rs = 0 . 12 ) . The even lower correlations observed by Vilborg et al . may be explained by their use of nuclear RNA , which also contains RNA produced before stress . This underestimates gene expression changes for genes with low basal RNA turnover [16] . It should be noted that gene expression fold-changes estimated from RNA-seq data ( even after normalization to house-keeping genes as performed here ) only indicate changes in the relative , but not absolute , abundance among all expressed genes . As the overall transcription levels decline during lytic HSV-1 infection [17] , positive fold-changes do not necessarily indicate actual transcriptional induction but only less down-regulation compared to other genes . In our previous study , we reported that DoTT-induced read-through was increased for genes without the canonical AAUAAA poly ( A ) -signal upstream of the gene 3’end . Similarly , Vilborg et al . found several 6-mers to be depleted ( including AAUAAA ) or enriched downstream of genes with pan-stress DoG transcription . However , their analysis focused on the total frequency of the 6-mers downstream of all pan-stress DoG genes instead of the frequency for individual genes . We now aimed to identify 6-mers whose abundance in the 100nt up- or downstream of individual gene 3’ends was significantly correlated to read-through ( FDR adjusted p-value <0 . 0001 for at least one condition or time-point , see methods ) . Strikingly , AAUAAA was the only 6-mer whose abundance upstream of gene 3’ends was significantly correlated with read-through in both stresses and HSV-1 infection ( Fig 2C ) and its absence upstream of gene 3’ ends was associated with significantly higher read-through ( Wilcoxon rank sum test , p<0 . 0001; Fig 2D ) . Other 6-mers were only significantly correlated to read-through in HSV-1 infection and showed no significant differences in read-through in salt or heat stress ( Fig F in S3 File ) . Upstream of gene 3’ends , negative correlations were found for a 6-mer overlapping the AAUAAA sequence as well as two C-rich motifs . Downstream of gene 3’ends , this included a number of G-rich motifs . Only one motif downstream of genes was positively correlated to read-through ( AUUUUU ) , but only in HSV-1 infection . This sequence resembles binding motifs of a number of RNA binding proteins [18 , 19] , including HNRNPC ( Heterogeneous Nuclear Ribonucleoprotein C ) , which has been shown to influence poly ( A ) site usage . To directly compare HSV-1-induced DoTT to DoG transcription , we calculated Spearman rank correlations of read-through values between each pair of conditions and time-points . This compares the ranking of genes with regard to read-through , i . e . whether top- and lowest-ranked genes tend to be the same between samples . Read-through mostly correlated extremely well ( Rs>0 . 8 ) between adjacent time-points for the same condition apart from the first three hours of HSV-1 infection where DoTT was hardly noticeable ( Fig 3A ) . Moderate but comparable correlations were observed between salt stress and either heat stress or HSV-1 infection at 4-5h p . i . ( Rs = 0 . 45-0 . 51 ) . In contrast , read-through in heat stress was slightly better correlated to salt stress than to HSV-1 infection ( Rs = 0 . 4 ) . Since we observed a weak correlation between read-through and gene expression fold-changes in all conditions , we also calculated correlations after excluding genes with highest fold-changes ( ≥2 in any sample ) . This aimed to exclude genes for which differences in read-through between conditions might be explained by changes in transcriptional activity . However , correlations for the remaining 2 , 601 genes did not increase , which is probably explained by the observation that gene expression fold-changes were also well correlated ( Fig G in S3 File ) . Thus , differences between conditions in DoTT/DoG transcription cannot be explained by differential alterations in transcriptional activity . Next , we performed hierarchical clustering of genes based on read-through ( average of replicates ) for each condition ( Fig 3B ) . This identified a large cluster of 1 , 368 genes ( 37% ) with read-through in all conditions ( marked in blue ) as well as a number of clusters with differences between conditions . It furthermore highlighted the prevalence of DoTT/DoG transcription with only 102 genes ( 3% ) showing no DoTT/DoG transcription ( defined as ≤5% read-through ) in any infected/stress sample . Overrepresentation analysis for Gene Ontology ( GO ) terms using DAVID [20] found an enrichment of genes with extracellular regions ( 25 genes ) and heparin binding ( 6 genes ) among these 102 genes . However , no functional categories were overrepresented for the 1 , 368 genes with read-through in all conditions . Interestingly , the only gene experiencing ≥75% read-through already after 2-3h p . i . HSV-1 infection and in all stress conditions was interferon regulatory factor 1 ( IRF1 ) ( Fig H in S3 File ) . IRF1 is an important mediator of both type I and II interferon signaling and studies with IRF1-deficient mice have shown an important role for IRF1 in the immune response against viruses [21–23] . Furthermore , even a relatively small reduction in IRF1 expression , e . g . mediated by cellular miR-23a , is sufficient to measurably augment HSV-1 replication in cell culture [24] . Notably , ribosome profiling data from our previous study revealed a >4-fold drop in IRF1 translation during HSV-1 infection despite an >1 . 8-fold increase in total RNA at 8h p . i . [10] . This presumably reflects the negative effects of DoTT on IRF1 translation and suggests that HSV-1 exploits DoTT to evade the host immune response . A striking characteristic of HSV-1-induced DoTT was the associated increase in aberrant splicing [10] . In particular , this comprised novel intragenic and intergenic splicing events as well as splicing associated with nonsense-mediated decay ( NMD ) . Intergenic splicing joins known exons of neighboring genes and confirms transcription of large chimeric transcripts spanning two or more cellular genes . It can be observed as early as 3-4h p . i . in HSV-1 infection . One of the most prominent examples connects SRSF2 and JMJD6 . We also observed intergenic splicing in the two stress conditions , but the few examples did not cluster with intergenic splicing events in HSV-1 infection ( Fig 3C ) . Analysis of induced splicing events upstream of gene 3’ ends , however , showed similar characteristics in both HSV-1 infection and salt and heat stress . In all three conditions , induced intragenic splice junctions were enriched for novel splice junctions and junctions found only in processed transcripts ( containing no ORF but not classified as long or short non-coding RNAs ) or in NMD-associated transcripts ( Fig 3D; examples in Fig I in S3 File ) . Genes with induced intragenic splicing events showed increased read-through in all three conditions ( Fig I in S3 File ) , but read-through was also observed in genes without induced splicing events . Thus , aberrant splicing upstream of gene 3’ ends more likely resulted from , rather than is responsible for DoTT/DoG transcription . One possible explanation for the association of aberrant splicing with DoTT/DoG transcription may be that all serine and arginine rich splicing factor ( SRSF ) genes included in our analysis ( SRSF2 , SRSF3 , SRSF5 , SRSF6 , SRSF7 , SRSF10 , SRSF11 ) showed DoTT/DoG transcription in at least two , but mostly all three conditions . All of these SRSF genes showed a >2-fold greater drop in translation at 8h p . i . HSV-1 infection in the ribosome profiling data than expected from the changes in their total RNA levels . Vilborg et al . reported that salt stress-induced DoG transcription in SK-N-BE ( 2 ) C cells depends on IP3R activation , Ca2+ release from intracellular stores and downstream kinases [11] . HSV-1 entry into cells is dependent on the activation of Ca2+ signaling pathways and triggers Ca2+ release from intracellular stores [25 , 26] . In addition , HSV-1 infection results in an increasing loss of stable , resting Ca2+ at late times of infection indicating a bimodal role of Ca2+ signaling in HSV-1 infection [27] . Before assessing the effect of Ca2+ signaling inhibitors on DoTT in HSV-1 infection of HFF , we first aimed to reproduce the results by Vilborg et al . in salt stress . HFF were exposed to 80mM KCl for 1h in presence of ( i ) an inhibitor of IP3R signaling ( 2-APB ) , ( ii ) the membrane permeable Ca2+ chelator BAPTA-AM , or ( iii ) inhibitors of the downstream kinases Ca2+/calmodulin-dependent protein kinase II ( CaMKII ) and protein kinase C/protein kinase D ( PKC/D ) ( KN93 and Gö6976 , respectively ) . DoG transcription was first quantified by qRT-PCR on total RNA for DDX18 , which shows strong read-through in HSV-1 infection as well as salt and heat stress . Consistent with the previous report , BAPTA-AM prevented DoG transcription while the other inhibitors resulted only in a moderate ( 25–65% ) reduction ( Fig 4A ) . We thus aimed to assess the effect of BAPTA-AM on DoTT in HSV-1 infection . To avoid the described detrimental effects of BAPTA-AM on virus entry and the onset of productive infection [25 , 26] , we only added BAPTA-AM to the cell culture media of HFF at 1h p . i . ( MOI = 10 ) when viral gene expression is already well initiated . To first determine its effect on viral gene expression , we quantified immediate-early ( ICP0 ) , early ( ICP8 ) and true late ( ICP5 ) gene expression at 8h p . i . by qRT-PCR . Strikingly , BAPTA-AM treatment was highly detrimental to viral gene expression of all three kinetic classes resulting in a >1 , 000-fold drop in viral mRNA levels ( Fig 4B ) . Considering this strong reduction in viral gene expression , we hypothesized that depletion of intracellular Ca2+ by BAPTA-AM in HFF might globally impair Pol II activity rather than specifically interfere with DoTT/DoG transcription . We thus analyzed the effect of 1h of BAPTA-AM treatment of uninfected cells on transcriptional activity of three cellular genes ( SRSF3 , IRF1 and DDX18 ) . For this purpose , we labeled newly transcribed RNA by adding 500μM 4sU to the cell culture medium for 1h . Following isolation and purification of the 4sU-labeled newly transcribed RNA ( 4sU-RNA ) from a fixed amount of biotinylated total RNA per condition ( 60μg ) , transcriptional activity of these genes was quantified using qRT-PCR on 4sU-RNA . BAPTA-AM indeed induced a drop in transcriptional activity that was at least as strong as observed upon inhibition of Pol II using actinomycin D ( Act-D; Fig 4C ) . In addition , global 4sU incorporation rates into total cellular RNA were substantially reduced upon BAPTA-AM treatment ( Fig 4D ) . This indicated that BAPTA-AM might not only interfere with Pol II but also with rRNA synthesis ( Pol I and III transcription ) , which contributes about 50–60% of 4sU-RNA in HFF as estimated from our RNA-seq data [10] . We thus quantified transcription rates from 4sU-RNA for a Pol I transcript ( 18S rRNA ) , a Pol III transcript ( 5S rRNA ) in addition to four genes transcribed by Pol II ( GAPDH , SRSF3 , IRF1 and DDX18 ) upon exposure of HFF to 80mM KCl for 1 and 2h and BAPTA-AM ( Fig 4E ) . In addition , we tested whether the combined exposure of cells to Gö6976 and KN93 , which also diminished salt stress-induced DoG transcription in total RNA , also globally affected transcriptional activity . While salt stress alone already resulted in a drop in transcription rates for Pol I ( ≈1 . 5-fold ) , II ( 3- to 5-fold ) and III ( ≈1 . 4-fold ) transcripts , BAPTA-AM impaired transcriptional activity of all three polymerases . This suggests that global inhibition of cellular RNA polymerases by BAPTA-AM rather than a specific effect on transcription termination is responsible for the loss of salt stress-induced DoG transcripts . As BAPTA and its derivatives share a high selectivity for Ca2+ over Mg2+ ( >105 stronger binding ) , the observed effects did not result from the co-depletion of intracellular Mg2+ [28] . Interestingly , combined Gö6976/KN93 treatment also globally impaired Pol I , II and III transcription , albeit to a lesser degree ( 2- to 10-fold ) , thereby explaining the slight reduction in DoG levels in total RNA ( Fig 4A ) . In contrast , 2-ABP treatment , which had shown no effect on DoG transcription when analyzing total cellular RNA , did not impair polymerase activity . Finally , we quantified read-through transcription for the three DoG genes SRSF3 , IRF1 and DDX18 in 4sU-RNA ( Fig 4F ) . Neither KN93/Gö6976 nor 2-ABP treatment had any effect on the induction of the respective DoG transcripts . Unfortunately , BAPTA-AM treatment did not allow to reliably measure read-through transcription due to the impaired transcription ( very low copy numbers or even negative PCR results ) . We conclude that the reduced levels of DoG transcripts upon inhibition of Ca2+ signaling do not result from direct effects on DoG transcription but from global inhibitory effects on cellular transcription in general . To our knowledge , this strong inhibitory effect of BAPTA-AM treatment on RNA polymerase activity has not been appreciated so far and should be considered when interpreting results obtained using BAPTA-AM to inhibit calcium signaling . Vilborg et al . initially reported that DoG transcripts ( DoGs ) were strongly enriched at the chromatin [11] and that one of the more abundant DoGs , doSERBP1 ( downstream of SERBP1 ) , remained at the site of synthesis . However , they subsequently also observed DoGs in the nucleoplasma of cells when searching for them by confocal microscopy with increased sensitivity [13] . To assess the fate of the transcripts arising from DoTT in HSV-1 infection , we separated cell lysates ( uninfected cells and 8h p . i . ) into cytoplasmic , nucleoplasmic and chromatin-associated fractions [29 , 30] and analyzed all three fractions as well as total cellular RNA by RNA-seq ( 2 replicates ) . The efficient separation of the cytoplasmic from the nuclear RNA fraction was confirmed by the enrichment of well-described nuclear lincRNAs ( MALAT1 , NEAT1 , MEG3; Fig J in S3 File ) in nucleoplasmic and chromatin-associated RNA as well as cytoplasmic enrichment of reported cytoplasmic lincRNAs ( LINC00657 , VTRNA2-1; Fig J in S3 File ) . In addition , overrepresentation of intronic reads in chromatin-associated RNA compared to nucleoplasmic RNA ( >5-fold higher ) demonstrated the efficient separation of these two RNA fractions ( Fig J in S3 File ) In uninfected cells , only chromatin-associated RNA showed notable levels of downstream transcription ( median 7 . 2%; Fig 5A ) , consistent with the standard model of transcription termination in eukaryotic cells [1] . At 8h p . i . , substantial read-through was observed in all fractions except for cytoplasmic RNA ( Fig 5B , Table B in S1 File ) , indicating that read-through transcripts are not efficiently exported to the cytoplasm . When we grouped genes according to their extent of read-through in 7-8h p . i . 4sU-RNA , we observed a strong increase during infection in the enrichment of the respective mRNAs ( counting only the exonic regions upstream of gene 3’ ends ) in both nucleoplasmic ( Fig 5C ) and chromatin-associated RNA ( Fig J in S3 File ) depending on the extent of read-through . While no change in nuclear enrichment was observed for genes without read-through , genes with >75% read-through were on average >2 . 5-fold more enriched at 8h p . i . than in uninfected cells . In particular , IRF1 was >6 and 4-fold more enriched in nucleoplasmic and chromatin-associated RNA , respectively , at 8h p . i . than in uninfected cells . Further evidence for an inefficient export of read-through transcripts is provided by intergenic splicing events , which are mostly absent in cytoplasmic RNA at 8h p . i . despite their considerable abundance in the other subcellular RNA fractions ( Fig 5D ) . This also explains our previous observation based on ribosome profiling that RNA chimeras and consequently genes induced by read-in transcription arising from DoTT are not , or only poorly translated [10] . We conclude that DoTT leads to nuclear retention of the respective read-through transcripts and thereby notably contributes to HSV-1 induced host shut-off . The similar overall level and high gene-specific correlation ( Rs = 0 . 8 ) of read-through in nucleoplasmic and chromatin-associated RNA indicates that transcripts resulting from HSV-1-induced DoTT are generally released from the chromatin , i . e . the site of synthesis , into the nucleoplasm ( see e . g . Fig 5F ) . Nevertheless , we identified 18 genes ( Table C in S1 File ) for which these transcripts appeared to remain at the chromatin ( ≤5% read-through in nucleoplasmic and cytoplasmic RNA , but ≥25% in chromatin-associated RNA; examples in Fig K in S3 File ) . Interestingly , there was a modest correlation ( Rs = 0 . 32-0 . 53 ) between the percentage of downstream transcription observed in chromatin-associated RNA of uninfected/unstressed cells and read-through upon stress or HSV-1 infection ( Fig 5E ) . This suggests that genes with a relatively high extent of downstream transcription in uninfected/unstressed cells might be predisposed for DoTT/DoG transcription . To exclude that this was an artifact of read-through being calculated from downstream transcription , we calculated ‘mock’ read-through values from the two biological replicates for the same time-point ( see methods ) . For mock read-through , the correlation was much lower at only ~0 . 13 . This suggests a link between downstream transcription detectable in chromatin-associated RNA in uninfected/untreated cells and read-through in stress/infection . A possible explanation might be that the respective poly ( A ) sites are weaker and thus more prone to further disruption by HSV-1 or stress-related mechanisms . Fig 5F illustrates this for IRF1 , for which downstream transcription in chromatin-associated RNA of uninfected cells was 14% and covered ~5kb . Interestingly , the correlation between downstream transcription in chromatin-associated RNA in uninfected cells and read-through during infection was highest at early time-points , i . e . at 1h for salt/heat stress and 2-3h p . i . for HSV-1 infection ( Fig 5E and 5G ) . At late stages of HSV-1 infection , even cellular genes with very little downstream transcription in chromatin-associated RNA from uninfected cells showed read-through transcription ( Fig J in S3 File ) . Based on publicly available DNase hypersensitive and ATAC-seq data for unstressed murine fibroblasts , Vilborg et al . recently reported that , even prior to stress , pan-DoG genes are already characterized by a chromatin signature indicative of an open chromatin state . However , due to the lack of respective data following salt or heat stress , they could not assess the consequences of read-through on cellular chromatin . We thus performed ATAC-seq in HFF at 0 , 1 , 2 , 4 , 6 and 8h of HSV-1 infection and 1 and 2h of salt and heat stress ( n = 2 ) . For all ATAC-seq samples , open chromatin regions ( OCRs ) were enriched around promoters , thereby confirming the high quality of the data ( Fig L in S3 File ) . Both length and score of OCRs at gene promoters correlated with gene expression in uninfected cells ( Rs = 0 . 42 and 0 . 4 , respectively; Fig L in S3 File ) . In contrast to the findings by Vilborg et al . , we did not observe a positive correlation between DoTT/DoG transcription and the presence of OCRs in the 5kb downstream of genes in unstressed/uninfected cells ( Fig M in S3 File ) . However , we noted a weak positive correlation ( Rs≤0 . 25 ) between the presence of downstream OCRs ( dOCRs ) and the expression level of the corresponding genes ( Fig M in S3 File ) . Notably , the highly expressed genes GAPDH and ACTB , which were not affected by DoTT/DoG transcription ( Fig 1D; Fig A in S3 File ) , were characterized by open chromatin downstream of their 3’ends already in uninfected cells ( Fig M in S3 File ) . In summary , our data argues against genes being predisposed for DoTT/DoG transcription by open chromatin downstream of their 3’ ends . We next analyzed the impact of HSV-1-induced DoTT on chromatin accessibility . To our surprise , we observed a substantial increase in open chromatin downstream of individual genes with HSV-1-induced DoTT ( Fig 6A , Fig N in S3 File ) . Here , downstream regions were often covered by OCRs for tens-of-thousands of nucleotides , similar to the pattern of read-through transcription in these downstream regions . This already became detectable at 4h p . i and resulted in a substantial increase in the number of long OCRs ( Fig 6B ) , which were specifically enriched downstream of genes ( Fig O in S3 File ) . Thus , these do not result from global effects of HSV-1 infection on cell viability ( e . g . due to enhanced chromatin accessibility in a subpopulation of dying cells ) . To quantify the total extent of open chromatin downstream of individual genes , we assigned dOCRs to genes if they were either close to the gene 3’ end or another dOCR that had already been assigned to the respective gene ( see methods ) and then calculated the total genomic length covered by dOCRs ( = dOCR length ) . This revealed a specific increase of dOCR length throughout infection for genes with high read-through ( Fig 6C ) . For 174 of the 681 genes ( 26% ) with >80% read-through at 7-8h p . i . , dOCR length exceeded 5kb at 6h p . i . , while only 26 of 326 ( 8% ) genes with ≤5% read-through at 7-8h p . i . had a dOCR length ≥5kb at 6h p . i . ( Fisher’s exact test p = 6 . 71×10-12 ) . For 11 of these 26 genes ( 42% ) , this was likely due to a close-by downstream gene with DoTT on the opposite strand ( see Fig 6D for FBN2 , >60kb dOCR matches the read-through of the SLC12A2 gene on the opposite strand ) . These 11 genes showed no DoTT despite long dOCRs ( originating from DoTT for genes with convergent transcription on the opposite strand ) and strong expression at 7-8h p . i . ( 10 with FPKM >1 , 6 with FPKM >3 ) . This indicates that the increase in downstream open chromatin during HSV-1 infection is not responsible for DoTT but rather that the formation of dOCRs requires DoTT . Furthermore , induction of long dOCRs for genes with read-through was dependent on the transcription rates of the respective genes . Genes with >80% read-through and long dOCRs were much higher expressed at 7-8h p . i . than read-through genes without long dOCRs ( Fig 6E ) . Accordingly , when dOCR length was compared to read-through for the 1 , 273 most highly expressed genes ( FPKM ≥2 ) at 7-8h p . i . , the difference in dOCR length between genes with different read-through levels was much more pronounced ( Fig 6F ) . Finally , strong increases in OCRs within gene bodies or promoter regions were only observed for genes with read-in transcription but not upstream of the poly ( A ) read-through . This explains the smaller , but nevertheless notable global increase in long OCRs in gene bodies ( Fig O in S3 File ) . Given the striking increase in dOCR length for well-expressed genes affected by HSV-1 induced DoTT , we also expected to see an increase in chromatin accessibility for salt and heat stress . However , there was no general increase in the number of long OCRs during salt or heat stress and no increase in dOCR length for individual genes in contrast to HSV-1 infection ( Fig P in S3 File ) . Accordingly , dOCR length did not increase for genes with high levels of stress-induced DoG transcription ( Fig P in S3 File ) , not even for highly expressed genes ( Fig 6A , Fig N in S3 File ) . Since read-through at 2h salt and heat stress was comparable to 4-5h p . i . HSV-1 infection and extensive dOCRs were clearly detectable at 4h p . i . , stress-induced DoG transcription does not appear to lead to open chromatin downstream of genes . Thus , only HSV-1 induced DoTT , but not DoG transcription in salt or heat stress , results in this striking increase in the accessibility of genomic regions downstream of affected genes . In addition to enhanced chromatin accessibility downstream of pan-stress DoGs , Vilborg et al . also found an enrichment of several histone marks typically found at actively transcribed genes ( H3K36me3 , H3K79me2 ) and at enhancers ( H3K4me1 , H3K27ac ) based on ENCODE data from unstressed murine NIH-3T3 fibroblasts [13] . Considering the discrepancy of our findings regarding open chromatin to their findings , we also analyzed ChIP-seq data from ENCODE for histone marks in uninfected/unstressed HFF . Significant positive correlations ( FDR adjusted p-value <0 . 01 ) between read-through in stress conditions and the presence of histone marks in the 5kb downstream of genes were only observed for the elongation marker H3K36me3 and DoG transcription in heat stress ( Fig M in S3 File ) . However , weak positive but not significant correlations were also observed in salt stress for H3K36me3 . The same was true in both stresses for two markers of accessible regulatory chromatin , H3K27ac and H3K4me1 . Interestingly , for H3K36me3 , positive correlations were also observed to read-through already detectable in the first two hours of HSV-1 infection . However , at later times of infection , this shifted to highly significant negative correlations between read-through and the presence of H3K27ac , H3K27me3 , H3K4me1 and H3K4me3 marks in the ENCODE data for uninfected cells . While this highlights important differences between DoTT and DoG transcription , the biological significance of the presence of certain histone marks in cells prior to stress or infection remains unclear . ChIP-seq data from time-course experiments of both HSV-1 infection and stress will be required to resolve these conflicting observations . In collaboration with Wyler et al . , we recently reported on the activation of antisense transcription in the human genome during lytic HSV-1 infection [14] . To assess whether this antisense transcription was also associated with the formation of OCRs , we investigated the 11 antisense transcripts that had been extensively validated by RT-qPCR and Nanostring nCounter assays ( Fig Q in S3 File ) . Interestingly , induction of antisense transcripts was clearly accompanied by an induction of corresponding long OCRs in three of these cases ( BBCas , EFNB1as , ING1as ) . These represented 3 of the 4 ( together with C1orf159as ) most highly expressed antisense transcripts at 7-8h p . i . , consistent with a role of transcription in the formation of long OCRs . For another four cases , an effect on open chromatin was visible but less clear ( NFKB2as , IFFO2as , FOXO3as , C1orf159as ) . Moreover , similar to transcripts of DoTT-affected genes , the length of the 11 antisense transcripts gradually increased quite substantially during HSV-1 infection , indicating that they are also affected by DoTT . To exclude that long OCRs during HSV-1 infection are an artifact of or are directly related to the induced antisense transcription , we determined the fraction of long OCRs ( ≥5kb; <80 long OCRs per replicate in uninfected cells , >500 per replicate at 6 and 8 p . i . ) that overlapped ( ≥25% of OCR in antisense transcript ) any of the 3 , 098 antisense transcripts identified by Wyler et al . ( Fig R in S3 File ) . In uninfected cells , ≥40% of the few long OCRs overlapped with an antisense transcript . With increasing duration of infection , this fraction decreased and only ~13% of long OCRs overlapped an antisense transcript at 8h p . i . , but often also a region of read-through transcription on the opposite strand . This supports a model in which HSV-1-induced dOCRs originate from read-through transcription while antisense transcripts also experience DoTT and consequently show the associated long OCRs if transcribed at a sufficient rate .
HSV-1 infection , cellular stress responses and cancer result in extensive transcriptional activity downstream of a subset of cellular genes [10–12] , but the relationships between the underlying molecular mechanisms remained unclear . By directly comparing HSV-1-induced DoTT with DoG transcription in salt and heat stress in the same experimental setting , we show significant overlaps between the genes affected by DoTT/DoG transcription but also clear context- and condition-specific differences . Importantly , differences were not only observed between DoTT and DoG transcription but also for DoG transcription between the two different stresses . Notably , the gene-specific correlation of read-through between salt stress and heat stress essentially equaled the correlation between salt stress and HSV-1 infection at 4-5h p . i . Multiple cis- and trans-regulatory factors are known to determine both splicing and poly ( A ) site usage [31] and even promoter elements have been shown to shape RNA processing by influencing Pol II processivity [32] . Thus , variability in DoG transcription upon different stressors and HSV-1-induced DoTT may originate from differences in downstream responses , interactions with other signaling pathways activated upon the different stresses or infection or even activation of alternative pathways with similar molecular consequences on the transcription termination machinery . In any case , the striking similarities between DoTT and DoG transcription indicate that related mechanisms are at play during HSV-1 infection . While the extent of DoTT further increased at late times of infection , salt stress-induced DoG transcription already declined by 2h , presumably due to detrimental effects of prolonged exposure to enhanced extracellular K+ concentrations on the exposed cells . In this respect , the expression of viral proteins counteracting the consequences of detrimental stress responses such as translational arrest and apoptosis may enable the much more efficient disruption of transcription termination by HSV-1 . The results presented here and in our previous manuscript [10] could have been a result of transcriptional noise that becomes evident in the context of transcription inhibition or extensive degradation of actively transcribed mRNAs by the virion-associated host shut-off protein ( vhs ) [33] . Alternatively , it might result from de novo pervasive transcription initiation downstream of the respective genes . However , we disfavor these models . First , it is important to note that we analyzed newly transcribed rather than total RNA . Therefore , transcriptional activity downstream or upstream of genes is always directly compared to the transcriptional activity of the corresponding gene occurring during the same timeframe of infection . The global loss in Pol II activity should equally affect genomic regions within , downstream and upstream of genes . Furthermore , strong transcriptional down-regulation of hundreds of genes has been analyzed in a broad range of different conditions using 4sU-seq [34–36] , none of which showed any increase in transcriptional activity downstream of genes . In addition , infection with a vhs-null mutant , which does not trigger a notable decline in transcriptional activity until at least 12h of infection [17] , still resulted in a very similar extent of read-through transcription as wild-type HSV-1 infection [10] . The data obtained in this study provide further strong evidence that downstream transcriptional activity arises from DoTT . First , the high correlations between the extent of read-through in HSV-1 infection , salt and heat stress indicate that all three conditions involve a common mechanism , namely poly ( A ) read-through . Second , RNA-seq analysis of subcellular RNA fractions revealed a striking dependence of nuclear retention of exonic regions during infection on the extent of read-through observed for the respective genes . The most likely scenario is that DoTT and extensive poly ( A ) read-through transcription result in large aberrant transcripts that cannot be efficiently exported to the cytoplasm . Third , the induction of extensive dOCRs for genes experiencing DoTT , which depends on the transcription level of these genes , provides strong evidence for an increase in absolute transcriptional activity downstream of these genes during infection . Additional evidence against pervasive de novo transcription initiation downstream of genes is provided both by the intergenic splicing events between neighboring genes induced in HSV-1 infection and the strong strand-specificity of the downstream transcriptional activity . De novo transcription initiation would not be limited to the strand of the upstream gene but would be expected to occur on either strand . The strong strand-specificity also excludes that downstream transcriptional activity is an artifact of the reported activation of antisense transcription during infection [14] . Moreover , DoTT and read-through transcription is clearly much more prominent than this antisense transcription . We now even provide evidence that antisense transcripts are also affected by DoTT and show DoTT-associated dOCRs . Vilborg et al . reported that DoG transcription was associated with an open chromatin state downstream of genes prior to stress [11] . This observation was not confirmed in our ATAC-seq data from primary human fibroblasts . While we currently cannot fully explain the discrepancy between these findings , we hypothesize that the enrichment of accessibility marks observed by Vilborg et al . may result from a restriction to pan-DoG genes detectable in nuclear RNA . As this also includes RNA transcribed before stress , relative levels of DoG transcripts are lower than in newly transcribed RNA . Thus , their analysis may be biased towards more highly transcribed DoG transcripts , which are more readily detectable . When we analyzed histone mark ChIP-seq data from ENCODE for uninfected HFF , we could only reproduce the positive correlation reported by Vilborg et al . [13] between the presence of the transcription elongation mark H3K36me3 ( but not H3K4me1 and H3K27ac ) and read-through in heat stress and to a lesser degree in salt stress . Interestingly , this was also observed during the first two hours of HSV-1 infection , which nicely fits to our observation that genes with active downstream transcription in chromatin-associated RNA in uninfected cells are more prone to read-through . In addition , it indicates that read-through occurring very early in HSV-1 infection may reflect a cellular stress response to infection and thus essentially DoG transcription . Later in infection , however , the picture completely shifts to negative correlations between read-through and repressive ( H3K27me3 ) and general ( H3K4me3 ) promoter marks as well as accessible regulatory chromatin ( H3K27ac and H3K4me1 ) . While the correlation with both repressive promoter marks and activating histone marks late in HSV-1 infection is difficult to interpret and seems contradictory , it hints that at this point other mechanisms than a general stress response may come into play . It is important to note , however , that the respective ChIP-seq data were only obtained from uninfected/unstressed cells and thus do not reflect the changes in histone marks upon infection/stress . The most striking finding of our study is the extensive increase in genome accessibility downstream of well-expressed genes affected by DoTT during HSV-1 infection , which essentially matched the transcriptional read-through observed at the respective time of infection . Of note , the peak heights of extensive dOCRs were often similar to levels observed in gene promoters of the respective genes where histones are displaced by transcription factors binding to promoter elements . However , in DoTT-associated dOCRs , this was not restricted to a few hundred base pairs but extended for tens-of-thousands of nucleotides . Our data indicate that dOCRs are not the cause but rather the consequence of DoTT and their formation additionally requires high levels of transcriptional activity in the respective downstream genomic regions . Considering the high correlation between DoTT and DoG transcription , we were surprised not to observe any evidence of dOCRs for DoG transcription in salt or heat stress . As DoTT-associated dOCRs were already well detectable by 4h of infection when the overall extent of DoTT and DoG transcription was very similar , the lack of dOCRs in salt and heat stress is not merely due to quantitative differences between the three conditions . Progression of transcribing Pol II across a gene is accompanied by the displacement of nucleosomes , followed by their rapid co-transcriptional repositioning immediately behind Pol II ( reviewed in [37] ) . We hypothesize that dOCRs result from impaired histone repositioning in the wake of Pol II . The lack of dOCRs in salt and heat stress indicates that dOCRs do not merely arise when Pol II starts transcribing far into previously untranscribed regions of the genome . Furthermore , gene bodies upstream of poly ( A ) sites affected by DoTT showed no general induction of OCRs , suggesting that there is no general inhibition of histone repositioning during HSV-1 infection . However , induced OCRs were also observed in gene bodies following read-in transcription , which argues against a role of distinct histone modifications in intergenic regions . Interestingly , HSV-1 infection was found to mobilize histones including linker and core histones ( H1 , H2B , H3 . 1 and H4 ) as well as histone variants ( H3 . 3 ) [38 , 39] . This resulted in increases in the pools of “free” histones despite an inhibition of histone synthesis during infection [40 , 41] . Therefore , it is unlikely that the induction of dOCRs results from a deprivation of free histones . On the contrary , the reported histone mobilization may at least partly result from impaired histone repositioning downstream of genes and thus release of histones from the respective regions into the nucleoplasm following read-through transcription . A critical role in nucleosome reassembly is played by the histone chaperons Spt6 and the FACT ( FAcilitates Chromatin Transcription ) complex ( reviewed in [42] ) . Interestingly , recruitment of Spt6 to active cellular genes includes direct interactions with the C-terminal domain ( CTD ) of Pol II [43 , 44] . Here , specific post-translational modifications of the CTD , which depend on the position of the transcribing Pol II within a gene , govern the functional state and properties of Pol II and its interactions with other factors [45] . Recently , the HSV-1 ICP22 protein was found to relocate both Spt6 and FACT to viral replication compartments . This may limit their availability to Pol II when transcribing cellular genes in HSV-1 infection [46] , but does not explain the selective failure in nucleosome reassembly only downstream of genes with read-through . Follow-up studies on recruitment and disengagement of Spt6 and FACT from Pol II upon infection with wild-type HSV-1 and mutant viruses as well as the concurrent analysis of post-translational modifications of the Pol II CTD will provide important insights into the functional regulation of transcription by Pol II and its termination downstream of genes . In summary , our findings provide a much more detailed picture of the molecular processes involved in DoTT/DoG transcription and point the direction for further studies to elucidate the underlying molecular mechanisms .
Human fetal foreskin fibroblasts ( HFF ) were purchased from ECACC and cultured in DMEM with 10% FBS Mycoplex and 1% penicillin/streptomycin . HFF were utilized from passage 11 to 17 for all high-throughput experiments . This study was performed using wild-type HSV-1 strain 17 . Virus stocks were produced in baby hamster kidney ( BHK ) cells ( obtained from ATCC ) as described [10] . HFF were infected with HSV-1 24h after the last split for 15 min at 37°C using a multiplicity of infection ( MOI ) of 10 . Subsequently , the inoculum was removed and fresh media was applied to the cells . Salt stress was initiated by adding 80mM KCl to the tissue culture medium . Heat stress was started by replacing the cell culture medium with pre-warmed 44°C medium and culturing the cells for 1 or 2h at 44°C . Newly transcribed RNA was labeled for 1h using 500μM 4-thiouridine ( Carbosynth ) . Total RNA was isolated using Trizol and newly transcribed RNA was purified as described [10] . The IP3R inhibitor 2-APB ( 100μM , Sigma-Aldrich ) , the PKC/PKD inhibitor Gö6976 ( 10μM; Tocris ) , the CaMKII inhibitor KN-93 ( 10μM; Tocris ) and the calcium chelator BAPTA-AM ( 50μM; Cayman Chemical ) were applied as described [11] . Actinomycin D ( 2μg/ml , Sigma-Aldrich ) was applied at a final concentration of 2μM to inhibit Pol II . Reverse transcription was performed using All-in-One cDNA Synthesis Supermix ( Biotool ) including a mix of hexanucleotide random primers and poly-dT primers . qRT-PCR was performed using the SYBR Green 2x Mastermix ( Biotool ) ( qRT-PCR primer sequences in Table D in S1 File ) . Relative quantitation was performed using the ΔΔCT approach . Dot blot analysis was performed as described previously [47] with a few minor changes regarding the detection of 4sU-incorporation into total cellular RNA . Briefly , metabolic labeling of newly transcribed RNA was initiated by adding 500μM 4sU to the cell culture medium together with either 50μM BAPTA-AM , 2μg/ml Actinomycin D or mock ( DMSO ) . Total RNA was isolated using Trizol and thiol-specifically biotinylated using Biotin-HPDP . Following removal of the unincorporated Biotin-HPDP by Chloroform extraction and recovery of the biotinylated RNA by isopropanol/ethanol precipitation , 200ng down to 22ng of biotinylated RNA or 60ng to 0 . 6 ng of a biotinylated oligo ( 50bp ) were spotted on a positively charged Zeta membrane ( Biorad ) in alkaline buffer . The membrane was subsequently probed with a Streptavidin-DyLight-680 conjugate and visualized using a LI-COR imaging system . Subcellular RNA fractions ( cytoplasmic , nucleoplasmic and chromatin-associated RNA ) were prepared combining two previously published protocols [29 , 30] . For the detailed protocol see S2 File . The efficiency of the fractionations was controlled by qRT-PCRs for intron-exon junctions for ACTG1 ( chromatin-associated vs other three fractions ) and western blots for histone H3 ( nuclear vs cytoplasmic fraction ) . Fractionation efficiencies were furthermore confirmed on the RNA-seq data by comparing expression values of known nuclear and cytoplasmic RNAs as well as intron contributions ( Fig J in S3 File ) . Sequencing libraries were prepared using the TruSeq Stranded Total RNA kit ( Illumina ) . While rRNA depletion was performed for total RNA and all subcellular RNA fractions using Ribo-zero , no rRNA depletion was performed for the 4sU-RNA samples . Sequencing of 75bp paired-end reads was performed on a NextSeq 500 ( Illumina ) at the Cambridge Genomic Services and the Core Unit Systemmedizin ( Würzburg ) . HFF were infected for 8h with wild-type HSV-1 at an MOI of 10 or exposed to 1h or 2h of 80mM KCl or 44°C as described above . ATAC-seq was performed according to the original protocol starting with 1x105 cells per condition [15] . ATAC-seq libraries were quantified by Agilent Bioanalyser and sequenced by NextSeq 500 at the Cambridge Genomic Services ( 75bp paired-end reads ) . Sequencing adapters were trimmed from sequencing reads using cutadapt [48] . Trimmed sequencing reads were mapped against ( i ) the human genome ( GRCh37/hg19 ) , ( ii ) human rRNA sequences and ( iii ) the HSV-1 genome ( HSV-1 strain 17 , GenBank accession code: JN555585 , only for HSV-1 infection data ) using ContextMap v2 . 7 . 9 [49] ( using BWA as short read aligner [50] and allowing a maximum indel size of 3 and at most 5 mismatches ) . For the two repeat regions in the HSV-1 genome , only one copy each was retained , excluding nucleotides 1–9 , 213 and 145 , 590–152 , 222 . As ContextMap produces unique mappings for each read , no further filtering was performed and all reads mapped to the human genome were used for downstream analyses . Number of mapped sequencing reads per genome position ( = coverage , sum of 2 replicates ) were visualized using the Integrative Genomics Viewer ( IGV ) [51] . No normalization was performed for this purpose . Number of read fragments per gene were determined from the mapped 4sU-seq reads in a strand-specific manner using featureCounts [52] and gene annotations from Ensembl ( version 87 for GRCh37 ) . All read pairs ( = fragments ) overlapping exonic regions on the corresponding strand by ≥25bp were counted for the corresponding gene . HSV-1 gene annotations were obtained from GenBank ( GenBank accession code: JN555585 ) . Expression of cellular protein-coding and lincRNAs was quantified in terms of fragments per kilobase of exons per million mapped reads ( FPKM ) and averaged between replicates . Only reads mapped to the human genome were counted for the total number of mapped reads for FPKM calculation . Fold-changes in FPKM values were normalized by dividing by median fold-changes for housekeeping genes ( as defined in [53] ) to account for different levels of DoTT/DoG transcription and consequently different numbers of intergenic reads in different samples and conditions . The percentage of transcription downstream or upstream of a gene ( on the same strand ) were calculated separately for each replicate as: As both transcription downstream or upstream of the gene ( = FPKM in 5kb downstream or upstream of gene ) and transcription within the gene ( = gene FPKM ) are quantified in the same timeframe of infection using 4sU-seq , both should be affected to the same degree by a general decrease in transcription . Thus , calculation of this ratio cancels out the effect of any general decrease in transcription . %downstream and upstream transcription were averaged between replicates and transcription read-through and read-in were then calculated as: If this resulted in negative values for a gene , read-through or read-in were set to 0 . As calculation of %downstream transcription or %upstream transcription cancels out the effect of any overall decrease in transcription , calculation of read-through or read-in are independent of any such decrease . Mock read-through values were calculated from the two replicates for each time-point for each condition . Here , mock read-through ( x , r ) = %downstream transcription in replicate r for sample x - %downstream transcription in replicate r’ for sample x . Here , r , r'∈{1 , 2} and r'≠r . Occurrence numbers of all possible 6-mer nucleotide sequences were determined within 100nt up- or downstream of gene 3’ends . Spearman correlations between these counts for each gene and read-through values in each sample as well as significance of correlations were calculated using the cor . test function in R and adjusted for multiple testing for each sample using the method by Benjamini and Hochberg for controlling the false discovery rate ( FDR ) [54] . Splice junctions and read counts for splice junctions were determined from spliced 4sU-seq/RNA-seq read mappings . All predicted junctions were considered that used at least one annotated exon boundary and ended within the annotated 3’ and 5’ ends of the corresponding gene . Only reads were counted that included at least 5bp on either side of the splice junction . Regulation ( up- or downregulation ) of splice junctions was evaluated in terms of the odds-ratio: cj*coj*cjcoj . Here , cj* and cj are the junction counts in infected or treated and uninfected or untreated cells , respectively . coj* and coj are the counts for all other junctions of the same gene in infected or treated and uninfected or untreated cells , respectively . Odds-ratios and significance of odds-ratios were calculated from replicate data using the Mantel-Haenszel chi-squared test in R . Multiple testing correction was performed with the method by Benjamini and Hochberg [54] . Splicing events were considered significantly upregulated ( downregulated ) if the adjusted p-value was ≤ 0 . 01 and the odds-ratio ≥2 ( ≤0 . 5 ) . Adapter trimming and mapping to human and HSV-1 genomes was performed as described for the 4sU-seq data . BAM files with mapped reads were converted to BED format using BEDTools [55] and OCRs were determined from these BED files using F-Seq with default parameters [56] . No filtering of OCRs was performed . Assignment of OCRs to gene promoters was performed using ChIPseeker [57] . 5kb dOCR length for each gene was calculated as the number of nucleotides in the 5kb directly downstream of the gene 3’ end that overlap an OCR . dOCR length for a gene was calculated as the total genomic length of downstream OCRs ( including only the positions downstream of the gene 3’ end ) assigned to the gene in the following way . First , all OCRs overlapping with the 10kb downstream of a gene were assigned to this gene . Second , OCRs starting at most 5kb downstream of the so far most downstream OCR of a gene were also assigned to this gene . This was performed iteratively , until no more OCRs could be assigned . 5kb dOCR length and dOCR length were averaged between replicates . Empirical cumulative distribution functions for dOCR length were calculated with the ecdf function in R [58] . For this purpose , genes were grouped according to read-through at 7-8h p . i . HSV-1 infection or salt or heat stress . Thresholds were chosen such that genes without DoTT ( read-through ≤5% ) were in one group and the remaining genes were divided in equal-sized groups according to read-through . Narrow peaks for ChIP-seq data of histone modification marks ( H3K27ac , H3K27me3 , H3K36me3 , H3K4me1 , H3K4me3 , H3K9me3 ) in HFF were downloaded from ENCODE ( epigenome series ENCSR403RCR ) . Presence of histone modification marks downstream of each gene was evaluated by determining the number of nucleotides in the 5kb directly downstream of the gene 3’ end that overlap peaks for the corresponding histone marks ( denoted as downstream histone mark length ) . Spearman correlations between read-through in all conditions and downstream histone mark length and significance of correlations were calculated using the cor . test function in R and adjusted for multiple testing for each sample using the method by Benjamini and Hochberg [54] . The datasets generated and analyzed in the current study are available in the Gene Expression Omnibus ( GEO ) database under the following accession numbers: 4sU-seq data of HSV-1 infection: GSE59717 . 4sU-seq data of salt and heat stress: GSE100469 . RNA-seq of total , cytoplasmic , nucleoplasmic and chromatin-associated RNA: GSE100576 . ATAC-seq data for HSV-1 infected cells: GSE100611 . ATAC-seq data for salt and heat stress: GSE101731 . | Recently , we reported that productive herpes simplex virus 1 ( HSV-1 ) infection leads to disruption of transcription termination ( DoTT ) of most but not all cellular genes . This results in extensive transcription beyond poly ( A ) sites and into downstream genes . Subsequently , cellular stress responses were found to trigger transcription downstream of genes ( DoG ) for >10% of protein-coding genes . Here , we directly compared the two phenomena in HSV-1 infection , salt and heat stress and observed significant overlaps between the affected genes . We speculate that HSV-1 either directly usurps a cellular stress response or disrupts the transcription termination machinery in other ways with similar consequences . In addition , we show that inhibition of calcium signaling does not specifically inhibit stress-induced DoG transcription but globally impairs RNA polymerase I , II and III transcription . Finally , HSV-1-induced DoTT , but not stress-induced DoG transcription , was accompanied by a strong increase in chromatin accessibility downstream of affected poly ( A ) sites . In its kinetics and extent , this essentially matched poly ( A ) read-through transcription but does not cause but rather requires DoTT . We hypothesize that this results from impaired histone repositioning when RNA Polymerase II enters downstream intergenic regions of genes affected by DoTT . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cellular",
"stress",
"responses",
"classical",
"mechanics",
"gene",
"regulation",
"cell",
"processes",
"dna-binding",
"proteins",
"vertebrates",
"mechanical",
"stress",
"dogs",
"animals",
"mammals",
"dna",
"transcription",
"epigenetics",
"chromatin",
"chromosome",
"biolo... | 2018 | HSV-1-induced disruption of transcription termination resembles a cellular stress response but selectively increases chromatin accessibility downstream of genes |
In the current study , we used a mouse model and human blood samples to determine the effects of chronic alcohol consumption on immune responses during Mycobacterium tuberculosis ( Mtb ) infection . Alcohol increased the mortality of young mice but not old mice with Mtb infection . CD11b+Ly6G+ cells are the major source of IFN-α in the lungs of Mtb-infected alcohol-fed young mice , and IFN-α enhances macrophage necroptosis in the lungs . Treatment with an anti-IFNAR-1 antibody enhanced the survival of Mtb-infected alcohol-fed young mice . In response to Mtb , peripheral blood mononuclear cells ( PBMCs ) from alcoholic young healthy individuals with latent tuberculosis infection ( LTBI ) produced significantly higher amounts of IFN-α than those from non-alcoholic young healthy LTBI+ individuals and alcoholic and non-alcoholic old healthy LTBI+ individuals . Our study demonstrates that alcohol enhances IFN-α production by CD11b+Ly6G+ cells in the lungs of young Mtb-infected mice , which leads to macrophage necroptosis and increased mortality . Our findings also suggest that young alcoholic LTBI+ individuals have a higher risk of developing active TB infection .
It is estimated that more than two billion people worldwide are infected with Mycobacterium tuberculosis ( Mtb ) , but only 5–10% of these individuals develop TB during their lifetime [1 , 2] . The geriatric population represents a large reservoir of latent tuberculosis infection ( LTBI ) [3] . It is difficult to diagnose and treat tuberculosis in aged individuals [3 , 4] . Approximately 57% of tuberculosis deaths occur in the aged population ( above 50 ) , and this burden is high in developed countries [5] . Immunosuppressive conditions , such as HIV infection , diabetes mellitus and drug and alcohol abuse , are risk factors that increase the chances of tuberculosis ( TB ) reactivation in people with LTBI [6–9] . In addition , individuals with alcoholism show higher relapse rates and a higher probability of having multidrug-resistant TB [10] . Alcoholism leads to the development of liver cirrhosis , cancer , insulin resistance , epilepsy , hypertension , psoriasis , preterm birth complications , cardiovascular diseases and stroke [11 , 12] . Chronic alcohol consumption impairs the host immune response to cancer and infections [13] . Alcohol impairs monocyte phagocytic and antigen-presenting capacities and suppresses the alveolar macrophage production of monokines , such as IL-23 , in response to infection [14–16] . Alcohol-exposed dendritic cells produce more IL-10 and less IL-12 , suggesting an inhibitory effect on dendritic cell function [17 , 18] . In humans and experimental mice , chronic alcohol consumption makes neutrophils hypo-responsive to bacterial infections[19] . Prolonged alcohol consumption induces type I interferon ( IFN ) and tumor necrosis factor alpha ( TNF alpha ) production [20] . Alcohol impairs NK cell trafficking and inhibits NK cell cytotoxicity [21] . Chronic alcohol consumption impairs adaptive immune responses mediated by B and T-cells [22] . These immunosuppressive effects of alcohol are more severe in elderly individuals than in young individuals [23] . Chronic alcohol consumption makes the host susceptible to various bacterial infections , including TB [24] . Epidemiological and immunological evidence strongly suggest a link between alcoholism and the worsening of TB disease [25] . Chronic alcohol consumption impairs the immune responses of Mtb-infected mice [19] . Alcohol feeding before BCG vaccination reduces T cell responses , but there are no effects when BCG vaccination is delivered prior to alcohol feeding [26] . These studies were performed after the short term-feeding of mice with alcohol , and the mechanism ( s ) involved in host susceptibility remain unknown . The long-term effects of alcohol consumption on host defense mechanisms against Mtb infection are also unknown , particularly in old individuals . In this study , we determined the survival of alcohol-fed young and old mice infected with Mtb . We also determined the immune mechanisms responsible for the early death of alcohol diet-fed young mice infected with Mtb .
Young and old mice were fed alcohol or control diets for one month and then infected with Mtb H37Rv as detailed in the methods section . Alcohol or control diet feeding was continued until the death of the mice or the termination of the experiment . As shown in Fig 1 , eighty percent of Mtb-infected alcohol-fed young mice died within 6 months ( p<0 . 01 , Fig 1A ) ; there was a twenty-five percent death rate in Mtb-infected alcohol-fed old mice , a twenty-five percent death rate in Mtb-infected control diet-fed old mice and no deaths in the control diet-fed young mice . In these groups of mice , most of the deaths occurred after three months . The bacterial burden in the lungs of these mice was measured three months after Mtb infection . As shown in Fig 1B , there was a marginal increase in the bacterial burden in Mtb-infected control diet-fed old mice compared to that in the other groups , and there was a marginal but significant decrease in the bacterial burden in the lungs of alcohol-fed old mice . The above results demonstrate that there is no correlation between the bacterial burden and increased mortality in alcoholic mice infected with Mtb . Serum alcohol levels and liver alanine transaminase activity were similar among all groups of mice ( Figs 1A , 1B and S1 ) . We determined whether alcohol had any effect on the pro- and anti-inflammatory responses of young mice following Mtb infection . Young mice were fed control and alcohol diets and infected with Mtb as in Fig 1 . After three months , the levels of various cytokines and chemokines were measured in the lung homogenates by multiplex ( 23-plex ) ELISA . As shown in Fig 2A , at three months p . i . , various cytokines and chemokines were measured , but only IFN-α levels were increased significantly in Mtb-infected alcohol-fed young mice compared to those in uninfected alcohol-fed and Mtb-infected control diet-fed mice ( Fig 2A ) . There was a marginal but significant decrease in IL-1α levels in Mtb-infected alcohol-fed young mice compared to those in Mtb-infected control diet-fed mice ( Fig 2A ) . Histological analyses indicated that the number of lesions throughout the lungs was significantly higher in Mtb-infected alcohol diet-fed young mice than in Mtb-infected young control mice and uninfected young mice ( Fig 2B , 2C and 2D ) . To determine the cellular source of IFN-α in Mtb-infected alcohol diet-fed young mice , we first quantified the leukocyte populations by flow cytometry . As shown in Fig 3 and S2A and S2B Fig , at three months after Mtb infection , the number of CD11b+Ly6G+ cells in the lungs was significantly higher in Mtb-infected alcohol diet-fed young mice than in Mtb-infected control mice and uninfected alcohol diet-fed mice . We next determined the phenotype of IFN-α-producing pulmonary cells three months p . i . ; there were no significant differences in the absolute numbers of IFN-α-producing CD11c+ and F4/80 cells ( Fig 4A and 4B ) . However , the absolute number of IFN-α-producing CD11b+Ly6G+ cells in the lungs was significantly higher in Mtb-infected young alcoholic mice than in uninfected alcohol diet-fed mice and Mtb-infected control mice ( Fig 4C and 4D ) . To confirm our findings that IFN-α levels were increased in the lungs of Mtb-infected young alcohol diet-fed mice , mice were euthanized three months p . i . , and lung sections were examined for IFN-α+ cells by immunofluorescence staining . As shown in Fig 4F , the mean immunofluorescence intensity for IFN-α was significantly higher in Mtb-infected alcohol diet-fed young mice than in Mtb-infected control and uninfected alcoholic mice . We also found that Ly6G+ cells are the major source of IFN-α ( Fig 4E ) . We further characterized this cell population in the lung tissues of Mtb-infected young alcohol diet-fed mice . As shown in S6 Fig , Ly6G+IFN-α+ cells were positive for CD11b , CD200 and CD163 but negative for F4/80 , CD68 , CD115 , CD11c , and Ly6C . Type 1 interferons have a protective role in viral infections [27] . However , in Mtb infection , type 1 interferon signaling causes immunopathology and early mortality in the infected mice [28] . We investigated whether the increased mortality of Mtb-infected alcohol-fed young mice was due to enhanced IFN-α production . Young mice were fed control and alcohol diets and infected with Mtb as in Fig 1 . After three months , the mice were treated with either a neutralizing anti-IFNAR-1 mAb or an isotype-matched IgG1 control mAb . As shown in Fig 5A , 100% percent ( p = 0 . 001 ) of the Mtb-infected alcohol diet-fed young mice that received the isotype-matched control mAb died within 2 months . In contrast , all mice that received the anti-IFNAR-1 mAb survived . Histological analyses indicated that there were significantly fewer necrotic lesions throughout the lungs of the anti-IFNAR-1 mAb-treated Mtb-infected alcohol diet-fed young mice than in those of the isotype antibody-treated Mtb-infected alcohol diet-fed young mice ( Fig 5B , 5C and 5D ) . We further examined the lung lesions of Mtb-infected ( three months after infection ) alcohol diet-fed young mice using confocal microscopy . As shown in S3 Fig , cleaved caspase 3 expression was similar in Mtb-infected young alcoholic mice , Mtb-infected control and uninfected alcoholic mice , suggesting that there is no significant difference in lung cell apoptosis in these groups of mice . We next examined the expression of receptor-interacting serine/threonine-protein kinase ( RIP ) -1 and RIP-3 , which are known to be expressed by cells undergoing programmed necrotic cell death ( necroptosis ) [29] . Six months after Mtb infection , lungs were isolated from control and alcohol diet-fed young and old mice , and the gene expression levels of RIP-1 and RIP-3 were determined by real-time PCR . As shown in Fig 6A , the expression levels of RIP-1 and RIP-3 in the lungs were significantly higher in Mtb-infected alcohol diet-fed young mice than in Mtb-infected alcohol diet-fed old mice and control diet-fed young mice . Confocal microscopy examinations of the lung sections also indicated significantly higher RIP-1 and RIP-3 expression in the lung lesions of Mtb-infected alcohol diet-fed young mice than in the lungs of Mtb-infected control and uninfected alcoholic mice ( Fig 6B and S3B Fig ) . RIP-1 and RIP-3 are expressed by F4/80 macrophages but were not expressed by IFN-α-producing Ly6G+ cells ( Fig 6C and 6D ) . To determine whether IFN-α-producing Ly6G+ cells are involved in the enhanced RIP-1 and RIP-3 expression in lung macrophages , we first used confocal microscopy to examine the lung sections of Mtb-infected young alcoholic mice for Ly6G+ and RIP-1 and RIP-3-expressing F4/80+ cell interactions . At three months p . i . , the imaging results indicated that RIP-1 and RIP-3-expressing F4/80+ cells from Mtb-infected young alcoholic mice were in closer proximity to IFN-α-producing Ly6G+ cells than those from Mtb-infected control mice ( Fig 6D and S4 Fig ) . More importantly , our results indicate that the marked increase in IFN-α production was spatially defined at the region where both Ly6G+ and RIP-1 and RIP-3-expressing F4/80+ cells interact with each other ( Fig 6D ) . RIP-1 and RIP-3 expression levels in the lungs were significantly lower in anti-IFNAR-1 mAb-treated Mtb-infected young alcoholic mice than in isotype antibody-treated Mtb-infected young alcoholic mice ( Fig 6E ) . We compared IFN-α production in the lungs of alcohol-fed young and old mice following Mtb infection . Young and old mice were fed control and alcohol diets and infected with Mtb as in Fig 1 . After three months , the mice were euthanized , and the lung sections were examined for IFN-α by confocal microscopy . As shown in S5 Fig , the immunofluorescence intensity for IFN-α was significantly lower in Mtb-infected old alcoholic mice than in Mtb-infected young alcoholic mice . We also found that RIP-1 and RIP-3 expression levels were significantly lower in the Mtb-infected old alcoholic mice than in the Mtb-infected young alcoholic mice ( S5B , S5C and S5D Fig ) . To determine the relevance of the above findings to the clinical manifestation of human Mtb infection , we obtained blood samples from alcoholic and non-alcoholic LTBI+ individuals . First , on the basis of age , we characterized the LTBI+ individuals by age group: <45 years ( young ) and >50 years ( old ) . We cultured peripheral blood mononuclear cells ( PBMCs ) in the presence of 10 μg/ml γ-irradiated Mtb . After 72 hours , IFN-α levels were determined by ELISA as detailed in the methods section . As shown in Fig 7 , γ-irradiated Mtb significantly enhanced the IFN-α levels by 2-fold in PBMCs from young alcoholic LTBI+ individuals and compared with those from non-alcoholic young LTBI+ individuals and by 2 . 9-fold compared to those from old alcoholic LTBI+ individuals . The baseline IFN-α levels were also high in young alcoholic LTBI+ individuals compared with those of other groups ( Fig 7 ) .
Chronic alcohol consumption modulates host immune defense mechanism ( s ) and makes the host susceptible to various fungal , viral and bacterial infections , including Mtb [13 , 15 , 19] . However , limited information is available regarding the mechanisms involved in alcohol-mediated host susceptibility to Mtb and other intracellular bacterial infections . In the current study , we fed young and old mice control and alcohol diets and determined the mortality rates and the immune mechanisms involved in host susceptibility to Mtb infection . Approximately 80% of the Mtb-infected alcohol-fed young mice died within 5 months; however , only 25% of Mtb-infected alcohol-fed old mice and 25% of alcohol-fed uninfected young mice died during the same period . There were no significant differences in the bacterial lung burdens of control and alcohol diet-fed young mice and alcohol diet-fed old and young mice . IFN-α levels were significantly higher in the lungs of Mtb-infected alcohol-fed young mice , and treatment with an anti-IFNAR-1 antibody increased their survival . In the lungs of Mtb-infected alcohol-fed young mice , IFN-α enhanced the expression of RIP-1 and RIP-3 , which are known to be involved in necroptosis . Mtb-infected alcohol-fed old mice and Mtb-infected control diet-fed old and young mice did not express IFN-α , RIP-1 or RIP-3 in their lungs . In response to Mtb , PBMCs from alcoholic LTBI+ healthy individuals produced significantly higher amounts of IFN-α than PBMCs from non-alcoholic young LTBI+ individuals and alcoholic and non-alcoholic aged LTBI+ individuals . Our findings demonstrate that alcohol enhances Ly6G+ cell infiltration and IFN-α production and increases necroptosis in the lung macrophages of young mice infected with Mtb , which leads to enhanced mortality . Chronic alcohol consumption inhibits host protective immune responses to infections , including Mtb infection , and increases the mortality rates of young and aged individuals [30 , 31] . According to the Centers for Disease Control’s ( CDC ) estimations , one-third of binge drinkers are old individuals , and human studies have found that compared young individuals , old individuals are more susceptible to various diseases [32–34] . Old individuals are likely to take prescribed medications , and in some cases , malnourishment and alcohol may have different effects on these individuals [34–36] . No experimental animal studies have been performed to determine the effects of chronic alcohol feeding in aging and Mtb infection . In the current study , we found that alcohol diet-fed young mice ( 1–2 months ) are more susceptible to Mtb infection and have a higher mortality rate than alcohol diet-fed old mice ( 17–22 months ) ( Fig 1A ) . Our findings suggest that alcohol worsens the TB pathology in the early stage of life and leads to increased mortality . We found that IFN-α is responsible for the early death of alcoholic Mtb-infected young mice . IFN-α is a type 1 interferon that belongs to the interferon family , which regulates the immune responses to infection , cancer and autoimmune diseases [37 , 38] . Type 1 interferons have a protective role during viral infections , but during Mtb infection , they enhance the pathogenicity [39 , 40] . Mtb proteins induce the production of type 1 interferons by host myeloid cells [41] . Type 1 interferons inhibit IL-1β production and enhance Mtb growth in myeloid cells [42] . In the current study , we found that IFN-α produced by Ly6G+ cells was associated with macrophage necroptosis and fatal immunopathology in the lungs of young alcohol diet-fed mice IFNAR1 signaling is detrimental during Mtb infection and promotes excess inflammation [28] . Furthermore , in the current study , we found that Mtb-infected alcohol diet-fed mice survived for a longer period of time and had less bacterial burden in their lungs when type 1 IFN signaling pathways were blocked . Old mice express transient early resistance to pulmonary tuberculosis , and type 1 cytokines have no influence on this early resistance [43] . It is known that several signaling pathways are defective in old mice [44] . The transient resistance in Mtb-infected old mice is due to a population of memory CD8+ T cells that express several receptors for Th1 cytokines; in addition , in aged mice , lung macrophages secret more proinflammatory cytokines in response to Mtb [43] . We have not determined the CD8+ cell and macrophage responses in alcoholic old mice; however , our results demonstrate that alcohol-fed mice were unable to enhance IFN-α production in Mtb-infected old mice , and there were no effects on mortality compared to the non-alcoholic Mtb-infected old mice . IFN-α production in young alcoholic Mtb-infected mice significantly reduced their survival . Our current findings suggest that defective signaling pathways that are involved in the production of IFN-α in Mtb-infected old mice may be protecting them from alcohol-mediated lung cell necroptosis . In various experimental models , it was shown that immune cells , such as macrophages , dendritic cells , T cells and Ly6G+ cells , produce type 1 interferons , and plasmacytoid dendritic cells are the major source [45] . However , only less than 1% of leucocytes are dendritic cells , and fifty percent of blood leucocytes are Ly6G+ cells [46] . Necrosis and Ly6G+ cell infiltration in the lung granuloma are characteristic features of tuberculosis granulomas , and these properties are associated with increased mycobacterial load and exacerbated lung pathology in human and experimental animals [47] . Netting Ly6G+ cells are the major inducers of type I IFN production [48] . Whole blood transcript signatures for active TB patients and pathway analyses revealed that the TB signature is dominated by a neutrophil-driven interferon ( IFN ) -inducible gene profile that consists of both IFN-γ and type I IFNαβ signaling [49] . Alcohol consumption can reduce the recruitment of Ly6G+ cells to the site of infection [19 , 50] . We have further investigated whether IFN-α-producing CD11b+ Ly6G+ cells are neutrophils . We found that these cells express a unique phenotype , including some neutrophils markers ( positive for CD11b , CD200 and CD163 but negative for F4/80 , CD68 , CD115 , CD11c , and Ly6C ) , suggesting these cells are not neutrophils . In the current study , we found fewer CD11b+Ly6G+ cells in the lungs of alcohol-fed Mtb-infected old mice than in those of alcohol-fed Mtb-infected young mice . We have not determined the factors underlying the reduced CD11b+Ly6G+ infiltration in the lungs of alcohol-fed Mtb-infected old mice , but it is known that in old individuals , the neutrophil lifespan is decreased , neutrophil precursors in the bone marrow proliferate less , neutrophil recruitment at the site of inflammation is reduced , and CD11b+ Ly6G+ cells are less functional due to alterations in signaling pathways [51 , 52] . Our results suggest that alcohol consumption can enhance these defects in old mice and that Ly6G+ neutrophil- like cells are unable to migrate to the lungs of Mtb-infected mice , resulting in less IFN-α production and necroptosis and enhanced survival . We found that IFN-α produced by Ly6G+ cells in alcohol-fed Mtb-infected young mice induces necroptosis in lung macrophages . Necroptosis is programmed necrosis that differs from other death pathways ( apoptosis , autophagy and pyroptosis ) due to the requirement of a unique signaling pathway associated with the activation of receptor-interacting protein ( RIP ) kinases 1 and 3 [53 , 54] . Caspase 1 expression was similar among all groups of infected mice , but RIP-1 and RIP-3 expression was significantly higher in the lungs of alcohol-fed Mtb-infected young mice than in those of control diet-fed Mtb-infected young mice and alcohol diet-fed Mtb-infected old mice . We also found that RIP-1 and RIP-3 expression was restricted to macrophages and that IFN-α-producing Ly6G+ cells were colocalized around macrophages . These findings suggest that IFN-α-producing Ly6G+ cells in alcohol-fed Mtb-infected young mice enhance necroptosis in lung macrophages . Necroptosis exacerbates inflammatory responses to infection which contributes to tissue damage and pathology [55 , 56] . Our current findings demonstrate that alcohol enhances IFN-α mediated necroptotic death of lung macrophages in young Mtb-infected mice ( Fig 6C ) . This leads to tissue damage and mortality in young alcoholic Mtb-infected mice . To determine the clinical relevance of our mouse studies , we compared IFN-α levels in the culture supernatants of γ-irradiated Mtb-cultured PBMCs from young and old alcoholic and non-alcoholic healthy LTBI+ individuals . We found that PBMCs from young alcoholic LTBI+ individuals produced significantly higher amounts of IFN-α after culture with γ-irradiated Mtb than those from young non-alcoholic , old alcoholic and old non-alcoholic healthy LTBI+ individuals ( Fig 7 ) . Our findings suggest that young alcoholic LTBI+ individuals have a higher risk of developing active TB infection . In conclusion , our studies demonstrate that alcohol increases the mortality of young but not old mice infected with Mtb . The increased mortality of alcohol-fed Mtb-infected young mice is due to IFN-α production by Ly6G+ cells . Further characterization of the exact phenotype of CD11b+ Ly6G+ cells and the delineation of the mechanisms through which alcohol enhances IFN-α production by Ly6G+ cells during Mtb infection will facilitate the development of therapies for alcoholic individuals with latent and active Mtb . Our findings may also be applicable to other intracellular pathogen infections .
All animal studies were performed with specific pathogen-free , 6- to 8- week -old and 17- to 22-month-old male and female C57BL/6 mice ( Jackson Laboratory and National Cancer Institute ) . The Institutional Animal Care and Use Committee of the University of Texas Health Science Center at Tyler approved the studies . The animal procedures involving the care and use of mice were conducted in accordance with the guidelines of the NIH/OLAW ( Office of Laboratory Animal Welfare ) . Blood was obtained from 17 non-alcoholic and 20 alcoholic healthy LTBI+ individuals who were 18–75 years of age . PBMCs were isolated from freshly collected blood samples . All subjects were HIV seronegative . The alcoholic LTBI+ individuals had a history of drinking at least 10–12 drinks per week . All human studies were approved by the Institutional Review Board of the Bhagwan Mahavir Medical Research Centre , and informed written consent was obtained from all participants . All human subjects involved in our study were adults . All animal studies were approved by the Institutional Animal Care and Use Committee of the University of Texas Health Science Center at Tyler ( Protocol #554 ) . All animal procedures involving the care and use of mice were undertaken in accordance with the guidelines of the NIH/OLAW ( Office of Laboratory Animal Welfare ) . All mice were maintained on a standard rodent chow diet ( LabDiet , catalog number 5053 , St . Louis , MO , 4 . 07 kcal/gm ) until the beginning of the experiment , when they were randomized into control or alcohol-containing liquid diet groups . The mice were fed alcohol using the Lieber-DeCarli liquid diet formulation ( Dyets Inc . , catalog number 710260; Bethlehem , Pa . 4 . 5 kcal/gm ) , which supplies 36% of the caloric intake as ethanol , or were fed an isocaloric liquid control diet ( LCD ) ( Dyets Inc . , catalog number 710027; Bethlehem , Pa , 4 . 5 kcal/gm ) as previously described [57] . The animals were fed the respective liquid diets for 5 of 7 days and the chow diet for 2 of 7 days . Animals in the liquid ethanol diet ( LED ) group were given water containing 20% ( wt/vol ) ethanol on the two chow diet days . The weights of the mice were recorded weekly . Mice were fed the alcohol and control diets , and after three months , they were infected with Mtb H37Rv using an aerosol exposure chamber as described previously [58] . Briefly , Mtb H37Rv was grown to the mid-log phase in liquid medium and then frozen in aliquots at -70°C . Bacterial counts were determined by plating on 7H10 agar supplemented with oleic albumin dextrose catalase ( OADC ) . For infection , the bacterial stocks were diluted in 10 ml of normal saline ( to 0 . 5 ×106 CFU [colony forming units]/ml , 1 ×106 CFU/ml , 2 ×106 CFU/ml , and 4 × 106 CFU/ml ) and placed in a nebulizer within an aerosol exposure chamber custom made by the University of Wisconsin . In the preliminary studies , groups of three mice were exposed to the aerosol at each concentration for 15 min . After 24 h , the mice were euthanized , and homogenized lung samples were plated on 7H10 agar plates supplemented with OADC . CFUs were counted after 14–22 days of incubation at 37°C . The aerosol concentration that resulted in ~50–100 bacteria in the lungs was used for the subsequent studies . For some experiments , mice were treated with anti-IFNAR-1 antibodies . One month after control and alcohol diet feeding , the mice were challenged with aerosolized Mtb . After 3 months , the mice received 0 . 3 mg of anti-IFNAR-1 ( BioXcell , Clone: MAR1-5A3 , Catalog number: BP0241 ) or isotype-matched control Ab ( rat IgG1 , Clone: MOPC-21 , Catalog number: BE0083 ) intravenously every 4 days for up to 2 months . Lungs were harvested from the alcohol and control diet-fed mice at the indicated time points after Mtb challenge and were placed into 60-mm dishes containing 2 ml of Hank's balanced salt solution ( HBSS ) . The tissues were minced with scissors into pieces no larger than 2–3 mm , and the fluid was discharged onto a 70-μm filter ( BD Biosciences , San Jose , CA ) that had been pre-wetted with 1 ml of PBS containing 0 . 5% bovine serum albumin ( BSA , Sigma-Aldrich ) suspended over a 50-ml conical tube . The syringe plunger was then used to gently disrupt the lung tissue before washing the filter with 2 ml of cold PBS/0 . 5% BSA . The total number of viable cells in the lungs was determined with the trypan blue exclusion method . For flow cytometry experiments , we gated based on the total lung CD45+ cells ( leukocytes ) and measured various cell populations . For flow cytometry , we used FITC anti-CD3 , PE anti-CD8 , APC anti-CD4 , APC anti-NK1 . 1 , APC anti-CD11b , FITC anti-Ly6G , PE anti-IFN-α , APC CD11-C , and FITC anti-F4/80 antibodies ( all from BioLegend ) . The antibodies used for the in vivo neutralization experiments were purchased from BioXcell ( anti-mouse IFNAR-1 , Clone: MAR1-5A3 , Catalog number: BP0241 , and mouse IgG1 isotype control , Clone: MOPC-21 , Catalog number: BE0083 ) . Anti-Ly6G ( Sigma-Aldrich; MABF474 ) , anti-F4/80 ( Abcam; ab6640 ) , anti-IFN-α-FITC conjugated ( R&D Systems; 22100–3 ) , anti-cleaved caspase-3 ( Cell Signaling Technology; 9661S ) , anti-CD163 ( Santa Cruz Biotechnology , INC; sc-58965 ) , anti-Ly6C ( Santa Cruz Biotechnology , INC; sc-52650 ) , anti-CD115 ( Santa Cruz Biotechnology , INC; sc-46662 ) , anti-CD200 ( Santa Cruz Biotechnology , INC; sc-53100 ) , anti-CD11c ( Abcam; ab33483 ) , anti-CD68 ( Abcam; ab53444 ) , anti-RIP-1/3 ( Santa Cruz Biotechnology , INC; sc-133102/sc-374639 ) and secondary antibodies ( goat anti-rat IgG ( H+L ) -Alexa 647 , goat anti-rabbit IgG ( H+L ) , Alexa Fluor 488 , and goat anti-mouse IgG ( H+L ) , Alexa Fluor 594 ) were obtained from Life Technologies , and fluoroshield mounting medium with DAPI ( Abcam , ab104139 ) was used for the confocal microscopy analyses . For surface staining , 106 cells were resuspended in 100 μl of staining buffer ( PBS containing 2% heat-inactivated FBS ) and Abs . The cells were then incubated at 4° C for 30 min , washed twice and fixed in 1% paraformaldehyde before acquisition using a FACS Calibur flow cytometer ( BD Biosciences ) . In some experiments , intracellular staining for IFN-α was performed . Controls for each experiment included cells that were unstained , cells to which PE-conjugated rat IgG had been added and cells that were single stained , either for a surface marker or for intracellular molecules . For IFN-α analysis , we gated based on CD11c , F4/80 , CD11b or Ly6G-positive cells and determined the percentages or the number of IFN-α expressing cells . In the lung homogenates , the following 27 cytokines and chemokines were measured using a multiplex ELISA kit ( M60009RDPD , Bio-Rad ) . The cytokines and chemokines analyzed were IL-1b , IL-1ra , IL-2 , IL-4 , IL-5 , IL-6 , IL-7 , IL-8 , IL-9 , IL-10 , IL-12 ( p70 ) , IL-13 , IL-15 , IL-17 , basic FGF , eotaxin , G-CSF , GM-CSF , IFN-γ , IP-10 , MCP-1 ( MCAF ) , MIP-1a , MIP-1b , PDGF-BB , RANTES , TNF-α and VEGF . RNA was isolated from lungs using TRIzol ( Invitrogen ) according to the manufacturer's instructions . Complementary DNA ( cDNA ) was generated from 0 . 5 mg of RNA and random hexamer primers using a Maxima First Strand cDNA Synthesis Kit for RT-qPCR ( BIO-RAD ) according to the manufacturer's instructions , and real-time PCR was performed . Gene expression for RIP-1 and RIP-3 was determined using Sybr green master mix ( Qiagen ) , gene-specific primers ( Sigma-Aldrich ) and an ABI Prism 7600 . All gene expression levels were normalized to β-actin internal controls , and the fold changes were calculated using the 2-ΔΔCT method . IFN-α levels were measured using ELISA kits ( Abcam , USA , catalog number: ab213479 ) according to the manufacturer’s instructions . Serum was collected without anti-coagulant by cardiac puncture from control and alcohol diet-fed mice . Serum alcohol levels were determined by using an ethanol assay kit as per the manufacturer’s guidelines ( Abcam , USA , catalog number: ab65343 ) . At the specified time points , mice were euthanized , and the harvested lungs were placed in 10% neutral buffered formalin ( Statlab , McKinney , TX , USA ) for 48 hours to inactivate the infectious agent . Paraffin-embedded blocks were cut into 5 μm-thick sections . For morphometric lesion analyses , the lung sections were stained with hematoxylin and eosin ( H&E ) and examined in a blinded manner to assess the necrotic lesions as previously described by Sibila et al . [59] . Briefly , each lung lobe was quantified for the lesion area and percentage of the lung lesions by using digital software ( NIH ImageJ; developed at the U . S . National Institutes of Health and available on the Internet at https://imagej . nih . gov/ij/ ) . Two investigators , DT and SC , independently assessed the immunohistochemical readouts using morphometric analyses . Confocal microscopy was performed to colocalize IFN-α-producing Ly6G and RIP1/3-expressing F4/80 cells in the lung sections . The lung tissues were stored in 10% neutral buffered formalin; then , the samples were paraffin embedded and cut into 5 μM thick sections that were deparaffinized and rehydrated . The tissue sections were subjected to heat-induced antigen retrieval in 10 mM sodium citrate buffer ( pH 6 . 0 ) . Then , the lung tissue sections were incubated in 0 . 025% Triton X-100 in PBST for 10 min and washed 3 × 5 min using PBS . Nonspecific binding was blocked with 5% goat serum in PBST for 1 hour , and the slides were washed 2 × 5 min with PBS . The slides were then incubated at 4°C overnight in PBST with the appropriate dilutions of the following primary antibodies: anti-Ly6G ( 1:200 ) , anti-F4/80 ( 1:50 ) , anti-IFN-α-FITC-conjugated ( 1:50 ) , anti-cleaved caspase-3 ( 1:400 ) , anti-CD68 ( 1:100 ) , anti-CD115 ( 1:50 ) , anti-CD200 ( 1:50 ) , anti-CD163 ( 1:50 ) , anti-CD11c ( 1:100 ) , anti-Ly6C ( 1:50 ) and anti-RIP-1/RIP-3 ( 1:50 ) ; subsequently , the slides were washed thoroughly 3 × 5 min with PBS . Then , the tissue sections were stained with the respective secondary antibodies at 1:1000 dilutions ( v/v ) , washed again with PBS for 3 × 5 min , and mounted with fluoroshield mounting medium with DAPI . The slides were then examined and analyzed under a laser-scanning confocal microscope ( Zeiss LSM 510 Meta ) . An IgG isotype secondary control was used for all the confocal microscopy studies , and Zen 2009 software ( Carl Zeiss ) was used for image acquisition; then , the images were processed/quantified uniformly for each experiment using ImageJ NIH software . Representative images from three different independent experiments are shown . Data analyses were performed using GraphPad Prism ( GraphPad Software , Inc . , La Jolla , CA ) . The results are expressed as the mean ± SE . For normally distributed data , comparisons between groups were performed using a paired or unpaired t-test and ANOVA as appropriate . Mouse survival was compared using the Kaplan- Meier log-rank test . | Chronic alcohol consumption modulates the host immune defense mechanism ( s ) and makes the host susceptible to various fungal , viral and bacterial infections , including Mycobacterium tuberculosis ( Mtb ) . However , limited information is available about the mechanisms involved in alcohol-mediated host susceptibility to Mtb and other intracellular bacterial infections . In the current study , we fed control and alcohol diets to young and old mice and determined the mortality rates and the immune mechanisms involved in host susceptibility to Mtb infection . We found that alcohol increases the mortality of young mice but not old mice infected with Mtb . The increased mortality in alcohol-fed Mtb-infected young mice was due to IFN-α production by CD11b+Ly6G+ cells . We also found that PBMCs from young alcoholic individuals with latent tuberculosis infection ( LTBI ) produced significantly higher amounts of IFN-α than those from young non-alcoholic , old alcoholic and old non-alcoholic LTBI+ individuals . Our findings suggest that young alcoholic LTBI+ individuals have a higher risk of developing active TB infection . Our studies provide the first evidence that chronic alcohol consumption induces IFN-α production in young Mtb-infected mice and increases their mortality rates . Further characterization of CD11b+Ly6G+ cells and delineation of the mechanisms through which alcohol enhances IFN-α production in Ly6G+ cells during Mtb infection will facilitate the development of therapies for alcoholic individuals with latent and active Mtb . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"death",
"rates",
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"chemical",
"compounds",
"immunology",
"social",
"sciences",
"diet",
"light",
"microscopy",
"organic",
"compounds",
"nutrition",
"microscopy",
"confocal",
"microscopy",
"po... | 2018 | Alcohol enhances type 1 interferon-α production and mortality in young mice infected with Mycobacterium tuberculosis |
In a living cell , the antiparallel double-stranded helix of DNA is a dynamically changing structure . The structure relates to interactions between and within the DNA strands , and the array of other macromolecules that constitutes functional chromatin . It is only through its changing conformations that DNA can organize and structure a large number of cellular functions . In particular , DNA must locally uncoil , or melt , and become single-stranded for DNA replication , repair , recombination , and transcription to occur . It has previously been shown that this melting occurs cooperatively , whereby several base pairs act in concert to generate melting bubbles , and in this way constitute a domain that behaves as a unit with respect to local DNA single-strandedness . We have applied a melting map calculation to the complete human genome , which provides information about the propensities of forming local bubbles determined from the whole sequence , and present a first report on its basic features , the extent of cooperativity , and correlations to various physical and biological features of the human genome . Globally , the melting map covaries very strongly with GC content . Most importantly , however , cooperativity of DNA denaturation causes this correlation to be weaker at resolutions fewer than 500 bps . This is also the resolution level at which most structural and biological processes occur , signifying the importance of the informational content inherent in the genomic melting map . The human DNA melting map may be further explored at http://meltmap . uio . no .
Currently , a community-wide effort is being pursued to understand how the genome itself , in concert with its epigenetic modifications and its organization in the cell nucleus , functions to provide each cell with the functionality and gene regulation that it requires at various levels of organization , such as cell state and tissue specificity . Two important and distinct approaches towards this goal may be found: ( 1 ) data-driven statistical learning methods that can provide results in a relatively short time , with little knowledge of the biological mechanisms involved; ( 2 ) knowledge-driven physical modeling that can provide a mechanistic understanding of the system in terms of its molecular constituents . However , by nature the latter is a relatively slow and difficult process . In most of the fifty years following the discovery of DNA structure [1] , the base sequence of DNA has been the primary focus , as epitomized by the completion of the human genome [2 , 3] . Now , with the sequence in hand , efforts have intensified to relate local sequence motifs or physical characteristics to involvement in particular DNA-dependent processes . Prediction algorithms for a number of DNA features based on sequence information have been developed , such as prediction of genes , alternative splicing , and transcription factor binding [4–6] , leading to an increasing number of annotations being available for the most widely studied organisms [7] . Studies of chromatin DNA in its nuclear environment have found that chromosomal compartmentalization occurs in the nucleus , and that apparently most of the genes locate in the inner part of the nucleus [8 , 9] . The importance of chromatin modification for gene expression has been recognized [10] . Still , many of the organizing principles of the structural elements of DNA within the eukaryotic nucleus remain poorly understood . The interest in the physical organization of DNA may be considered as a reflection of the fact that many complex biological processes require an integrative approach , given the relative shortage of laboratory-based observations . Sequence-oriented annotations of diverse types , physical modeling , and experimental observation of chromatin organization are now providing rich , yet disparate , sources of information on scales ranging from a few base pairs to the whole genome . A systematic integration and statistical exploration of the combined data is believed to provide a more complete picture of the functionality of DNA in its natural context . In parallel with the determination of the DNA sequence , efforts have been made to model the molecular behavior of DNA . A central issue has been prediction of how the DNA denatures , or melts , to dynamically create local single-stranded regions , to which a number of single-strand binding elements can attach , and thus influence such functions as gene transcription and a number of other DNA-dependent functions . Algorithms for this aspect of DNA behavior have been built on the foundations of polymer theory and statistical mechanics , in which the pioneering work of Poland and Scheraga [11] established a method to predict the melting profiles of DNA . When these algorithms were applied to the exonic sequences of specific genes , it became possible to develop technology to separate mixed mutant and wild-type sequences , and to observe the quantitative sequence-specific distribution of point mutations in human tissues and large human populations [12–14] . DNA melting algorithms aim at quantitative predictions of in vitro experiments in dilute DNA solutions at specific temperatures , ionic strengths , and denaturing solutes [15] . Few attempts have been made to bring quantitative modeling inside the cell nucleus . A precise model of chromatin DNA would be very complex , accounting for a number of physical structures . Among these would be the generation and response to superhelical stress [16] , the histone-based hierarchical folding , the maintenance of topological order , the nucleus wall–inducing confinement , attachment , and molecular crowding , the protein-driven replication and transcription processes , as well as other nonequilibrium effects to be discovered . Due to the lack of such chromatin models , the aim of the present type of investigation is neither a quantitative prediction in vivo nor in vitro . Rather , as pioneered by L . S . Lerman , who applied a DNA melting algorithm to the human beta-globin gene [14] , and by G . J . King [17] , who applied melting maps in a study of yeast chromosome III , the thought is that the existing melting algorithms , while describing in vitro experiments , may also reflect qualitatively some of the chromatin DNA properties in vivo . Therefore , the primary interests here are the qualitative features , such as propensities for single-strandedness , rather than the quantitative detail . To produce qualitative information that is physically realistic , however , a quantitatively accurate model and algorithm is required . An important requirement for any melting algorithm to be useful on long genomic sequences is that its computation time grows linearly with sequence length . For decades , the only available linear algorithm was the Poland–Fixman–Freire ( PFF ) algorithm [18 , 19] . It calculates the melting properties according to the classical Poland–Scheraga model [11] , originating in the 1960s , which considers a base pair to have simply two distinct states , helix and coil . The key element of the Poland–Scheraga model are the loop entropies [11] , whose scaling behaviors have been derived from various random walk polymer models that may take into account excluded volume effects , while other effects , such as chain stiffness in smaller loops ( <30 bp ) , are less well-understood [15] . The statistical weight of interior loops given by the loop entropy factor cannot be expressed as a product over base pairs , which intrinsically makes an exact calculation grow quadratically with sequence length , as in the Poland algorithm [19] . However , an approximation of the loop entropy factor was incorporated into the Poland algorithm in 1977 by M . Fixman and J . J . Freire [18] , providing the linear but approximative PFF algorithm . Various implementations of the PFF algorithm have been available in the scientific community , among which Lerman's MELT87 implementation was the first widely available code . Only in recent years have other linear algorithms become available [20 , 21] . In one such algorithm [21] , some of us introduced a Forward–Backward method ( analogously to the Poland algorithm ) for the recursive calculation of partition functions in the Poland–Scheraga model . Another recent algorithm [20] calculates the melting properties according to the Dauxois–Peyrard–Bishop model [22 , 23] . This model does not explicitly have a loop entropy factor that could slow down the computation . Instead , the energetics relies on each base pair having a continuum of possible states , mimicking a gradually varying geometry of the hydrogen-bonded bases . Calculations involve an integration over these continuous variables that can be algorithmically complex , but a recent approximative discretization method [24] has provided a linear algorithm that may be applicable to whole-genome computation of melting temperature [20] . Any algorithm that is quadratic or slower could be applied in linear time , by using the basic windowing technique of dividing up the sequence into pieces to be calculated individually and merging the results afterward . Such an Alexandrian solution is usually problematic due to the associated errors . For the Poland–Scheraga model , these errors could be kept small by first locating the pieces according to successive helical regions [25] . However , neither the PFF nor the more recent linear melting algorithms rely on a windowing technique for their speed . For those algorithms , windowing cannot provide much further reduction in computation time , but it may reduce the required RAM considerably . For more complex melting algorithms , on the other hand , windowing may be the only choice for analyzing long genomic sequences . One such example is C . J . Benham's SIDD model of superhelical DNA melting [26–28] . This model has distinct helix and coil states as does the Poland–Scheraga model , but rather than focusing on loop entropies and excluded volume , it models the torsional stresses imposed with a fixed linking number . Unwinding transitions can be induced by both increased temperature and negative superhelicity . SIDD calculations with a windowing technique have identified the bubbles or SIDD sites in yeast and various microbial genomes at physiologically reasonable values of the temperature and linking number [29] . The present work , however , applies the original PFF algorithm , which has the speed required for a human genomic calculation . Earlier studies similar to the present one have investigated genomic profiles of the local GC contents [30] , partly because they provide an indirect reflection of DNA's biophysical properties . It would , however , be misleading to consider the melting temperature as simply a function of the local di-nucleotide distribution . This misconception is propagated by web tools such as the WEB-THERMODYN [31] , or the EMBOSS' DAN [32] , in which melting temperature predictions appropriate for oligonucleotides are applied to a sliding window in longer sequences . The profiles thus obtained are essentially just the compositional profiles in disguise , and should provide similar results in genomic analyses . The problem with GC contents and sliding window approaches is that they fail to exhibit the cooperative physics . Base pair melting temperatures are determined by the organization of the sequence into cooperative domains of tens to hundreds of base pairs with well-defined boundaries . Each domain has its own melting temperature , resulting in a characteristic appearance of a melting map . The location of the domain boundaries partly depends on the long-range effects of the loop entropies , the extent of which , however , has not been fully investigated . The calculation of a melting map is nontrivial , in the sense that the whole sequence must be considered to account for these long-range effects . Although cooperativity features are absent in GC profiles , some correlation is expected between the melting map and the local GC contents . For example , a genome has a mosaic structure organized on many levels , with GC contents varying between isochores , exons , and introns , and in local motifs with abrupt changes in composition . By applying segmentation algorithms to genomic GC profiles [33 , 34] , a set of boundaries between regions of different GC contents can be found . The GC variations may in some cases “force” the thermodynamic domain boundaries to coincide with the GC-based boundaries , but the a priori expectation is that the two sets of boundary locations are different due to cooperativity . The first step in exploring cooperativity on a genomic level presented here opens a route to unravel novel mechanistic implications of the DNA sequence . Combining a knowledge-driven modeling approach and data-driven statistical explorations , we now report the complete calculation of the human genomic melting map , and we report a first explorative examination of the potential information content of the melting map . We discuss the large-scale computational challenges , such as the algorithmic complexity , the high-precision floating point formats , a Fixman–Freire approximation for very large sequence lengths , and the hardware requirements . We find that the cooperativity of DNA dominates at sequence resolution below 500 base pairs , and most importantly , that neighboring melting domains influence each other such that GC content is no longer a sufficient predictor of single-strandedness . This has important implications for understanding the interactions in chromatin .
The melting profiles were calculated using the Poland algorithm [19] with the Fixman–Freire approximation [18] . We employed a Fortran implementation that has its origin in the MELT87 program written by L . S . Lerman [14] , but we have made important modifications of the code to enable large-scale genomic computations . First , we use data types with ultra-high precision ( exponent terms to 4 , 500 , with up to 800 significant digits ) , to prevent the well-known problem of arithmetic overflow in dynamic programming . Second , we extend the accuracy of the Fixman–Freire approximation to very large loop sizes . In the Fixman–Freire scheme , the loop entropy factor Ω ( x ) = σ ( 2x + d ) −α ( which is a function of x , the number of melted base pairs in a loop plus one ) is approximated by Ω′ ( x ) = const × σ · f ( 2x + d ) , where the power function has been replaced by some multi-exponential function The parameters I , An , and Bn ( n = 1 , . . , I ) can be determined by fitting f ( x ) to x−α . The MELT87 code contained a hard-coded set of I = 10 exponentials . Although it is known that with a fixed number I , the Fixman–Freire approximation breaks down for long enough sequences , the consequences for the melting calculations have been largely ignored in the literature . Figure 1 shows a plot of the exact loop entropy factor Ω ( x ) , together with an I = 10 and an I = 21 approximation . The I = 10 approximation is only accurate for loop sizes up to the order of 104 , whereafter it decreases exponentially . The I = 21 approximation is accurate for loop sizes up to the order of 108 , whereafter it also decreases exponentially . When we first applied an I = 10 approximation in the calculation for human chromosomes , we observed “ceiling” artefacts imposing upper limits on the melting temperatures . An interpretation of this observation is that very large loops should be included in the partition function at high temperatures , despite their low statistical weights . To take large loops properly into account , we derived a Fixman–Freire approximation for arbitrary sequence length . Here , the parameters I , An , and Bn are given algebraically by the following expressions: I ≥ 1 + ln ( 2N ) where N is the sequence length , Bn = en−I , and Previously , such parameters have been obtained using more complex algorithms for solving the curve-fitting problem . The straightforward mathematical expressions provided here may be advantageous in terms of simpler programming and numerical reproducibility . We have set up a Web tool for calculating these parameters for given user input , including graphs that show the accuracy of the approximation ( see http://meltmap . uio . no/tools/loopentropy . html ) . For all the human chromosomes , we used the set of I = 21 exponentials . The extra computation time spent ( about twice that for an I = 10 set ) removes the artefacts that would otherwise have required extra validation efforts . The Poland algorithm also uses parameters that determine the free energy contributions of the helical segments . Several sets of experimentally determined parameters are available , but a comparative study by SantaLucia [35] has shown that a consensus among them exists . The choice of parameters should not be very important for the present type of study . Here we use Gotoh and Tagashira's ten nearest neighbor parameters [36] . An advantage of this set is that it was specifically designed for the Poland algorithm , with their modification that free energies are assigned to nearest-neighbor doublets rather than base pairs , and the parameters were determined by fitting calculated melting curves to experimental curves ( at salt concentration 19 . 5 mM ) . For the loop entropy factor , we used σ = 3 . 5 · 10−5 , α = 1 . 75 , and d = 0 . We do not distinguish methylated cytosine from unmethylated cytosine , and we use the parameters for unmethylated cytosine in both cases . Unknown bases in the sequence ( denoted by N ) are assigned their own parameters obtained by averaging over the four bases A , C , G , and T . The output of the Poland algorithm is a probability profile showing the probability of each position to be in the helical state calculated for a given temperature . For each chromosome , we calculate all probability profiles in the range 45° C to 110 °C at every 0 . 1 °C temperature increment . From this set of probability profiles , we derive the melting temperatures Tm ( x ) at which the probability at position x equals 50% . The resulting Tm-profiles or melting maps summarize the main features of melting along the sequences . The melting maps are stored in a format rounded to two digits after the decimal point . The complete calculation of probability profiles for all human chromosomes takes approximately 22 CPU days on an HP Superdome ( 64 × Itanium 2 processors , 1 . 5 Ghz , 6 MB cache ) . The calculation requires at least 13 GB RAM to process the longest chromosome ( ~240 Mbps ) with seven arrays extended precision . In some of the downstream analyses , we used “zoomed-out” melting maps , i . e . , averaging melting temperatures in nonoverlapping windows of a certain size ( varied from 10 bp to 1 Mbp ) . As the melting maps represent cooperative melting events of many base pairs , many features are still present at a lower resolution . UCSC Golden Path Human Genome Sequence Release hg17 ( May 2004 ) [37] , containing the Build35 assembled by the International Human Genome Project sequencing centers , were downloaded and used in our calculation of DNA melting profiles . We also calculated another set of melting profiles for 24 randomized chromosome sequences . When generating the randomized chromosomes , we ensured that the total length , and the number of A , T , G , C , and N of each chromosome ( N represents unknown bases , which are mostly located in euchromatic gaps ) , as well as the start and end positions of each consecutive N stretch , correspond to their human chromosome counterparts . Only the base compositions , not the di- or tri-nucleotide compositions , were specified . The randomized chromosomes are not completely featureless , however , since they contain the same stretches of N's as in the human chromosomes . The melting map algorithm was executed with the randomized chromosome set as input , and all the downstream statistical analyses were performed similarly . A characteristic of a melting domain is that each base pair has the same melting temperature . The flat plateaus of a melting map may give an indication of the location of domains . Two alternative segmentation methods were developed to identify flat segments of the melting map . First , we identified flat and nonflat segments of a given constant size ( e . g . , 100 bp or 1 Kbp ) by ranking the standard deviations ( SDs ) of the melting temperatures within nonoverlapping windows . Those segments having high SDs were designated as nonflat , and the ones with low SDs were designated as flat . Second , we also defined three types of segments , called up , down , and flat , based on the stepwise change in melting temperature , ΔTm = Tm ( x + 1 ) − Tm ( x ) , between neighboring positions within a segment ( in the 5′ to 3′ direction ) . An up segment was defined as a consecutive series of stepwise increases in Tm by at least ΔTm ≥ 0 . 13°C . Vice versa , a down segment was defined as a consecutive series of stepwise decreases by ΔTm ≤ −0 . 13°C . The flat segments consisted of small stepwise changes of |ΔTm | ≤ 0 . 01 °C . Neighboring positions with intermediate-sized stepwise change ( 0 . 01 °C < |ΔTm | < 0 . 13 °C ) were not assigned to any segment . In Chromosome 21 , 1 . 07% of all neighboring positions were part of up segments , 1 . 07% were part of down segments , 88 . 82% were part of flat segments , and the rest was not categorized . For each identified segment [xstart , xend] , the length L = xend − xstart + 1 , the average melting temperature , and the end-to-end step height ( Tm ( xend ) − Tm ( xstart ) ) , were determined . Segments of L < 9 were discarded for downstream analyses . The Pearson correlation coefficient was used to quantify the direction and magnitude of coordinated relations between various pairs of continuous variables , for example , the melting temperature versus the GC contents , and the recombination rates , respectively . The local GC content was calculated for every nonoverlapping window of different lengths ( varied from 10 bps to 1 Mbps ) as the ratio of G + C over the total number of A + T + G + C within each window . Note that the Ns do not contribute in the above definition , i . e . , we made no assumptions of the base pair composition with respect to unknown bases . When investigating the correlation between melting profiles and the recombination rates , we used the DeCODE data source , as publicly available [38 , 39] . This set of data has a resolution of 1 Mbp . For exploratory analysis of annotation correlations , the EpiGRAPH analysis service [40 , 41] was utilized . This service provided statistical association analysis for more than 1 , 000 human genomic annotations , including DNA sequence properties and patterns , repeat frequency and distribution , CpG island frequency and distribution , predicted DNA structure properties , predicted transcription binding sites , evolutionary conservation , and SNPs . These annotations were tested against discretized pairwise classes of input melting regions . As these calculations were computationally demanding , only sets of up to 100 cases per class were performed for each analysis . The enormous size of the human genome , with the shortest chromosome being more than 45 Mbps , requires an approach that can , beyond local details , reveal possible global patterns in an analysis of the melting map . Wavelet analysis , having evolved from Fourier transformations , has become an increasingly popular and useful tool for analyzing signals that contain nonstationary power at many different frequencies . It has been found in previous studies [42 , 43] using wavelet analysis of the GC contents of human chromosome sequences that regular nonlinear oscillatory behavior occurs . By association to the same underlying organizational principles , we therefore similarly examined if we could identify wavelengths by spectral analysis of the calculated DNA melting map . We applied the Morlet wavelet [44] to identify possible dominant periodic components in the melting map , using MATLAB Wavelet Toolbox ( The MathWorks , http://www . mathworks . com ) . Although the wavelet decomposition algorithm has been under continuous improvement since its inception , wavelet analysis is still computationally demanding . We performed continuous wavelet transformation , using a 1-Kbp window-averaged melting profile over a wide range of scales from 20 Kbps to 5 Mbps , at steps of 20 Kbps . Subsequently , the scale-averaged wavelet power spectra were computed for examining the underlying rhythm of fluctuations in power over various scales . We also randomly chose some melting map stretches of 2~3 Mbp in length from various chromosomes , and performed the wavelet analysis within each nonoverlapping 10-Kbp or 20-Kbp segments . This analysis was done at base-pair level , that is , using scales from 2 bp to 1 , 024 bp at steps of 2 bp to capture oscillations at a high resolution .
The fundamental feature of a melting map is the occurrence of thermodynamically stable and unstable regions , having relatively high and low melting temperatures , respectively . For the human chromosomes , we obtained statistics of melting map features using nonoverlapping averaging window sizes from 10 bp to 1 Mbp . Table 1 shows the basic statistics of melting temperatures averaged over nonoverlapping 1-Kbp segments . The highest 1-Kbp window averaged melting temperatures varied between the chromosomes in the range 86 . 35 °C to 88 . 40 °C , where the latter occurred within the 60 , 074th 1-Kbp segment of Chromosome 20 , containing a number of GC-rich repeat motifs such as ( CGG ) n . The lowest melting temperature per chromosome varied from 48 . 85 °C to 51 . 82 °C , where Chromosome 16 displayed the low 48 . 85 °C at the 10 , 501th 1-Kbp segment , which fully overlapped with a repeat of ( TA ) n type . The 100-bp segments displayed similar statistics ( see http://meltmap . uio . no/results/misc/meltmap_chr_global_stat_100bp . pdf ) . The broad ranges of melting temperatures in the human melting map were in contrast with those of the randomized chromosomes , which had melting temperatures in a narrow range of 70 ± 6 °C . A similar picture arose for the GC contents of 1-Kbp windows . In the human genome , the bulk of GC contents ( in 1-Kbp nonoverlapping window ) ranges roughly from 20% to 80% , while in the randomized chromosomes , the ranges are about six times as narrow . We compared the correlations between GC content and the melting temperature using different window sizes , ranging from 10 bps to 1 Mbps . The human genomic melting map showed a very strong correlation with the local GC content within windows of 1 Kbp and larger; above 0 . 99 for all chromosomes ( see Figure 2 ) . As also shown in Figure 2 , the correlation was found to be relatively low at small window sizes , increasing roughly log-linearly until reaching more than 0 . 98 at a window size of 500 bps . Thus , a likely interpretation is that the main features differentiating the GC content and melting temperature lies in the sub-500 bp range . Figure 2 also shows that , for all window sizes , the Tm and GC correlation was smaller for the randomized chromosomes than for the human chromosomes . Correlating a genome-wide recombination rate [38] with the melting map , we observed ten chromosomes where the melting temperature and recombination rates ( at a 1-Mbp resolution level , in a gender-mixed population ) correlated positively with a significant p-value of below 4 . 16E-4 ( using Bonferroni correction ) , as shown in Table 2 . Chromosomes 4 ( r = 0 . 602 ) , 5 ( r = 0 . 599 ) , and 13 ( r = 0 . 560 ) displayed the highest correlations . For three chromosomes ( 15 , 20 , and 21 ) , there was a negative correlation . Between the SD of the melting temperature over averaged 1-Mbp segments and the recombination rate , the correlation was weaker . Similarly , the Pearson correlations of the SNP frequency distribution [45] with the mean and SD of the melting temperatures were calculated for all chromosomes using 1 Mbp nonoverlapping windows . No strong correlations between these features were found to be statistically significant ( see Table 2 ) . SNP frequencies in flat segments on Chromosome 21 were investigated but did not reveal strong correlation with melting temperature ( unpublished data ) . For a more detailed analysis of the stable and unstable regions , we defined the melting temperatures above 90 °C and below 50 °C as extreme temperatures ( see http://meltmap . uio . no/results/extreme_tms . html ) . We defined stable ( i . e . , high-Tm ) and unstable ( i . e . , low-Tm ) regions as consecutive stretches of extreme high/low temperatures . Twenty high-Tm regions ( average melting temperature 90 . 25 °C , average region length 347 bps ) and 20 low-Tm regions ( average melting temperature 49 . 5 °C , average region length 460 bp ) were randomly chosen from various chromosomes . EpiGRAPH analysis ( see Methods ) between these stable and unstable regions ( see http://meltmap . uio . no/results/EpiGraph/061011_125400_423443447_Attributes . html ) indicated that unstable regions were associated with AT richness , low levels of evolutionary conservation , and high SNP frequency , and also exhibited frequent overlap with tandem repeats . The stable regions correlated with not only physical parameters of DNA , such as high solvent accessibility ( as illustrated in Figure 3A ) , high DNA rise and roll , but also informational content , such as gene overrepresentation . By comparing the melting temperature distributions of exonic and nonexonic regions of each chromosome , we further observed that for most of the chromosomes , there seemed to be an offset of about 2 °C toward higher temperature for exons , compared with nonexons ( see http://meltmap . uio . no/results/exon_vs_nonexon . html ) . The perhaps most interesting aspect of a melting map may in fact lie within the melting domain segmentation . In a macroscopic view , the human genomic melting map broadly follows the local GC content . However , with increased resolution , the cooperative melting characteristics appear as expected . Relatively flat plateaus of nearly equal melting temperatures are widespread , interspersed with step-like areas of both minor and large changes . We applied our segmentation definitions to examine the details of the melting map and its cooperativity . We thus performed an EpiGRAPH comparison of the annotation differences between two classes of randomly selected cases among the up/down versus flat segments of lengths 20~22 bp from Chromosome 21 , all having average Tm = 68 °C . The up/down class consisted of 50 segments with end-to-end step heights above ±6 °C . The flat class consisted of 50 flat segments with end-to-end step heights below 0 . 1 °C ( see http://meltmap . uio . no/results/EpiGraph/061119_190322_497232509_Attributes . html ) . A statistically significant association between Alu-type SINE repeat structures and the up/down class was found , compared with the flat class ( see Figure 3B ) . There were also an increased number of transcription start sites and genes in the up/down segments compared with the flat segments . The analysis was repeated with similar findings using Chromosome 22 . These observations applied also to the segmented data of nonoverlapping 100-bp windows based on SDs of the melting temperatures ( see http://meltmap . uio . no/results/EpiGraph/061109_194104_300525259_Attributes . html ) . When comparing segments with increasing melting temperatures ( up segments ) with decreasing temperatures ( down segments ) ( in a 5′-3′ orientation ) , no significant differences were found . In an attempt to capture possible functional differences between flat segments of different sizes ( below 100 bps ) , we performed EpiGRAPH analyses using 25 cases from each of two classes , which contained flat segments at ~68 °C in the length of ~60 bps and ~20 bps , respectively , from Chromosome 21 . Both options of choosing data with and without repeat masking were examined . It was found that the correlation of the two classes with the kurtosis and SD of the physical parameters ( twist , rise , and roll ) , as well as those for C and G frequency , had Bonferroni-adjusted p-values being equal or smaller than 10−6 , but no significant correlation to functional annotations were found ( unpublished data ) . To further investigate the relationship between sequence statistics and melting features , we examined the GC contents of all 50 bp flat segments of Chromosome 21 . Intriguingly , we observed not one , but three , distinct bands in a scatter plot ( see Figure 4 ) , indicating that regions having identical sequence composition ( i . e . , equal GC content ) , may have qualitatively different structural stabilities ( i . e . , melting temperatures ) . We extended the analysis also to flat segments of lengths 20 , 100 , 150 , and 200 bp . The separation between the three bands increases toward approximately 5 °C with decreasing segment length and , on the other hand , it disappears at approximately 200 bp . This phenomenon can also be observed for other human chromosomes , as well as for the randomized chromosomes . We found a simple rule that relates the three bands to neighboring regions and cooperativity . We grouped all the 50 bp flat segments into three categories ( I , II and III ) based on the difference between the average melting temperatures of the segments and their neighbor regions ( also 50-bp long ) at both sides . The segments in category I had higher melting temperatures on both sides than on themselves , and those in category III had lower melting temperatures on both sides than on themselves . The other segments were clustered into category II . When Figure 4 was color-coded according to this category definition , we found that the three visually distinct bands overlapped very well with the three mathematically defined categories . This provides an understanding of the impact on DNA thermal stability from neighboring regions in terms of cooperativity [15] , and this could not have been revealed by sequence-based statistics alone . Wavelet analysis on the low resolution melting map ( i . e . , using averaged Tm of nonoverlapping 1-Kbp windows across the genome ) was performed in order to uncover possible oscillating patterns of melting temperatures along the whole sequence . Although different chromosomes did not reveal an identical pattern of periodicities , ~200 Kbp , ~400 Kbp , and ~1 Mbp were the common wavelengths observed in multiple chromosomes . This was also conducted for the GC content ( using 1-Kbp nonoverlapping windows ) , and it tended to fall into the same general picture as that of the melting map . In contrast , no visible peaks could be found in the spectral analysis of the randomized chromosomes ( see http://meltmap . uio . no/results/wavelet_global . html ) . As it was possible that some periodic patterns with scales fewer than 1 , 000 bps also exist locally , we performed high-resolution ( scales from 2 bps to 1 , 024 bps ) spectral analysis on several randomly chosen ENCODE regions ( see http://meltmap . uio . no/results/wavelet_local . html ) . In these analyses , a frequency of approximately 150 bps , roughly corresponding to the nucleosome length , was frequently seen . We also observed a general oscillation pattern in the range of 500~600 bps in most of the regions examined . A sample plot of local wavelet analysis is shown in Figure 5 .
We here present the complete melting map of the human genome . While we find a high correlation ( 0 . 99 ) between the local GC content and the melting map averaged in 1-Kbp windows for all human chromosomes , Figure 2 demonstrates that when calculations are performed on windows of fewer than 500 base pairs , the correlation is much smaller . This suggests that the characteristics of cooperativity being present in the melting map ( as steps and flat plateaus ) are not observable from GC content , as expected . For instance , at the window size of 100 bps , only 88 . 8% of the variation in the melting map could be attributed to the changes of GC content , while the rest should be attributed to the cooperativity . This is a qualitative difference that apparently correlation based on larger than 500 bp windows fails to recognize . The randomized chromosomes derived from the selected randomization procedure were intended to serve as a baseline for comparison and for verification of our analyses . As these were required only to have the same overall frequencies of As , Cs , Gs , and Ts as the corresponding human chromosomes , the di-nucleotide composition was not preserved in the randomization procedure . Other randomization procedures would be worthwhile exploring for biological questions . The randomized chromosomes displayed a much more uniform GC content than the human chromosomes . As a consequence , the mosaic structure of the human chromosomes , presumably reflecting biological function , was not preserved . The correlations for the human chromosomes between DNA annotations and melting map features were not found for the randomized ones , well in line with our expectations . We correlated the melting map with various physical and functional features of the genome at several levels of resolution in a first exploration of the information contained . The recombination rate [30 , 39] and the SNP frequency [46 , 47] have previously been reported to be positively correlated to the GC content . In a recent study on compositional symmetry of DNA and recombination rate , a negative correlation was found , indicating that asymmetry favors recombination [48] . As asymmetry will reflect regions of changing melting temperature , this was expected . In an examination of the SNP frequencies in flat melting segments , we only observed relatively minor variations in the extreme ends of the temperature range . For the remainder of the temperature range of flat segments , there appears to be a constant frequency of SNPs . Thus , more detailed studies are required to clarify whether DNA bubble openings may be important for SNP-inducing processes . Also , the chromosomally variable correlation of recombination and the melting map remains to be elucidated . Recently , an isochore map was published [49] , that relies on observing the SD of the GC content in 100-Kbp windows , and defining the borders of isochores as abrupt shifts . It is not unlikely that the domain nature of the DNA melting map could be helpful , for defining isochores by melting temperatures , or for defining borders of GC isochores , although these topics remain to be explored . In a zoomed-in view , the domain structure of the melting map is distinct from the GC content . To explore possible correlations between these domains and the plethora of other existing annotations for the human genome , it was necessary to extract the corresponding segments from the melting map in an automated way . No clear and rigorous definition of a melting domain exists , and it is not known how accurately domains can be determined from a melting map . Thus , we tried several approaches to locate segments at various resolution levels . Given a constant size of a window , we defined a flat window as having minimal SD of melting temperature , as opposed to nonflat windows that have high SD . To identify melting segments of various lengths , we chose an approach based on incremental step of consecutive melting temperatures . Both approaches were shown to be useful for this first exploratory effort . Future studies , however , should address the segmentation aspect more elaborately . Better segmentation algorithms can be applied , for example , by modifying existing algorithms used for analyzing the GC contents [33 , 34] . A better knowledge of how the melting cooperativity manifests itself in a melting map may also contribute to the development of segmentation algorithms . In an approach of Yeramian et al . [50] , the thermodynamic boundaries are located from a set of probability profiles at a range of temperatures , instead of using the melting map , but an automated method has not been provided . Another strategy is to calculate the bubble boundary locations directly , instead of first calculating the probability profiles and the melting map , followed by a segmentation to extract the information . For example , an algorithm for calculating the possible locations of melted and helical regions ( stitch profiles ) has been developed [51 , 52] , but as its computation time scales quadratically , a genome-wide analysis must await the development of a linear version of that algorithm . However , using the simple segmentation algorithm described , we were able to identify , and to some extent quantify , the effect of cooperativity between neighboring regions . We clearly demonstrated that the influence of neighbor regions is visible for the segments at lengths fewer than 150 base pairs , and increases in importance down to 20-bp segments . Also , it is observed that the cooperative effect depends on the length of the central segment , melting temperature differences between neighbor regions , and the lengths of neighboring regions . Cooperativity is generally not considered for most computer prediction algorithms of biological functions . As has been shown by Benham et al . for promoter locations in prokaryotes [53] , it seems reasonable to expect improved predictions taking this aspect into consideration . The extreme instances of high and low melting temperatures were statistically compared with existing annotations . The low-Tm regions naturally coincided with AT-rich regions , but were also found to coincide with regions of poorer conservation and relatively frequent SNP and repeat occurrence . The high-Tm regions correspondingly were found to be associated with GC-rich sequences , but more interestingly also to various physical parameters of DNA , such as high solvent accessibility and rise , and a higher association to genes . These findings are generally expected from the underlying sequence composition . It is for instance well-known that the gene frequency is higher in GC-rich regions , as are the bendability and B to Z conformation transitions . It is thought that these parameters relate to a propensity for gene transcription [54] . As the cooperative effect was found to be more pronounced for shorter melting segments , we focused on those of length fewer than 100 bps . Among these , a selection of flat and nonflat segments significantly identified Alu type repeats as a major contributor of nonflat segments . The general structure for these retrotransposed sequences of a few hundred base pairs consists of a GC-rich transcription start site , a variable middle part , and an AT-rich tail part [55] . In fact , we show that the Alu-type repeats represent a considerable fraction of the areas in the genome having steep melting temperature changes . An overrepresentation of transcription start sites and increased gene frequency was also found in nonflat segments , as compared with flat segments . Recently , it has been observed that Alu-type sequence may have significant effects on gene expression , either through their influence on alternative splicing , through adenosine-inosine editing , or through protein translation influences [56] . Alu sequences have also been reported as having an increased fraction of SNPs [57] , both in the GC-rich body , but also in the GC-poor tail of the Alu sequences , possibly related to increased recombination rates , thus underlining the possibility that AT replication slippage may be significant in the generation of SNPs in the AT-rich tail of the Alu sequence , as suggested previously [57] . We found that exonic sequences had a shift toward higher melting temperature , compared with non-exons across the whole spectrum of melting temperatures , as well as a larger tail at the high melting temperature side . Others have previously found correlations between the occurrence of exons and relatively stable , high-Tm regions [58 , 59] . However , this seems to be a species-dependent feature , possibly due to the occurrence of long intron regions in higher eukaryotes . In a study of genes where the introns had been removed , a correlation between thermodynamic boundaries and exon boundaries was suggested [60] , and explained as a propensity of intronic sequences to be inserted at the boundary between open and closed DNA . This correlation also conforms well with the correlation between the GC content and exons , as previously described [61] . Previously , others have performed spectral analysis of selected regions of the human genome GC content and intra-strand asymmetry [42 , 62] , in order to uncover wavelengths or oscillations along the DNA , and reported two significant periods ( 110 Kbp and 400 Kbp ) , which roughly correspond to sizes relevant for DNA loops and vertebrate replicons . We could identify these features in the genomic melting map as well . The computational cost for a genome-wide spectral analysis at base pair level resolution is presently too high . However , in a local wavelet analysis of several ENCODE selected genomic regions , we could identify a wavelength of approximately 150 bps . Some supportive studies for this observation also can be found [63–66] , for instance showing that there is a general correlation between nucleosome complex local DNA double helix curvature and DNA chain bending as a function of the sequence composition . This paper is primarily concerned with establishing basic relationships between the human genomic melting map and available biological annotations . We are aware that the melting model employed in this work was developed for in vitro predictions . DNA in vivo , being much more densely packed together with histones and other macromolecules , seems too encumbered to form melting bubbles freely . Nevertheless , single-stranded regions are widespread in vivo , driven by molecular motors , rather than by temperature . Both replication and transcription rely on local opening of the DNA to take place , and also these processes should be scrutinized for possible dependencies on the melting map . For instance , DNA mutation rates may be related to the probabilities and lifetimes of bubbles exposing the bases . In a study using the Dauxois–Peyrard-Bishop model on selected transcriptional promoters , it was indicated that the algorithm could identify the transcription start site [67] , although the validity of this finding is disputed [68] . Studies using the SIDD model have shown that sites susceptible to opening correlate with replication origins and transcriptionally active regions [69–71] . Such topics should be addressed at the genomic level . The observation that neighboring segments , or potential bubbles , influence each other suggests that any prediction algorithm of sequence features related to single-stranded states would benefit greatly from including cooperativity as represented by the melting map . Today , such algorithms , for example , for transcription factor binding sites prediction ( see recent review by Bajic et al . [72] ) , are often hampered by large numbers of false positive predictions . We expect that prediction algorithms relying on sequence motifs alone could be improved . The central hypothesis for this melting map exploration is that the predictions of in vitro melting may reflect also the in vivo behavior . It is reasonable to believe that the sequence-dependent bubble openings are functionally important in vivo . However , the correlations with biological annotations do not necessarily indicate causalities , and their possible physical or biological origins cannot be deduced from this preliminary work alone . Some correlations may stem from sequence composition due to evolutionary changes , reflected in the melting map , but with no relation to bubble openings and functional role . Further studies of the melting map and its association to annotations and DNA structures are clearly warranted and at present made possible on a genomic scale . One such interesting topic would be the evolutionary aspects of the melting curve across related species with available sequence information . It is our opinion that the development of further refined melting models , that include more aspects of the in vivo physical constraints and flexibility of chromatin DNA as actually experienced in the different settings in the nucleus , could benefit from the knowledge gained in studies like the present one . This would significantly influence the understanding of the mechanisms behind a number of central structural and functional aspects of cells . The following materials ( including online tools , downloadable data , and results ) can be accessed via the Web page http://meltmap . uio . no/: 1 ) brief introduction of the project ( http://meltmap . uio . no/intro . html ) , 2 ) source code of the melting map calculation ( http://meltmap . uio . no/code . html ) , 3 ) online tool for calculating parameters of loop entropy factor approximation ( http://meltmap . uio . no/tools/loopentropy . html ) , 4 ) human genomic melting map data files ( based on UCSC release hg17 ) ( http://meltmap . uio . no/rawdata . html ) , 5 ) browser of human genomic melting map ( http://meltmap . uio . no/browser ) , 6 ) human genomic melting map being included in the Ensembl genome browser as a DAS source ( http://meltmap . uio . no/tools/ensembl . html ) , and supplementary analytical results ( http://meltmap . uio . no/analysis . html ) . | In a living cell , DNA both is an information carrier and carries out important structural tasks , such as organizing its replication and distributing the chromosomes to the daughter cells . DNA is frequently depicted as an antiparallel double-stranded helix , but DNA may rather be viewed as having a dynamically changing structure . This is because in performing most of these tasks , it is necessary for the DNA helix to become single-stranded locally , or unwound , thereby creating “bubbles” in the double strand , much as what happens when a twisted rope with two strands is untwisted . In the cell , this happens by the aid of the enzymatic machinery , but it may also be observed in experiments when a gradual increase in temperature produces bubbles . Our calculations in producing a melting map are based on temperature changes , but may be viewed as a map of bubble formation tendencies along the genome . In DNA , an opening bubble does not open one base at a time , but rather as a cooperative event , in that several base pairs act in concert to form a bubble , and we use an algorithm that takes this aspect into consideration . We then explore the correlations between the melting map and many known features of the human genome . We also demonstrate the extent of cooperativity , and find that the melting map carries information otherwise not available . Once the melting map is calculated , a number of more detailed studies of relationships to DNA structure and function are made possible , as well as improvements of algorithms for modelling DNA with associated proteins as they occur in the natural cellular environment . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"molecular",
"biology",
"genetics",
"and",
"genomics",
"homo",
"(human)",
"computational",
"biology"
] | 2007 | The Human Genomic Melting Map |
Neural population equations such as neural mass or field models are widely used to study brain activity on a large scale . However , the relation of these models to the properties of single neurons is unclear . Here we derive an equation for several interacting populations at the mesoscopic scale starting from a microscopic model of randomly connected generalized integrate-and-fire neuron models . Each population consists of 50–2000 neurons of the same type but different populations account for different neuron types . The stochastic population equations that we find reveal how spike-history effects in single-neuron dynamics such as refractoriness and adaptation interact with finite-size fluctuations on the population level . Efficient integration of the stochastic mesoscopic equations reproduces the statistical behavior of the population activities obtained from microscopic simulations of a full spiking neural network model . The theory describes nonlinear emergent dynamics such as finite-size-induced stochastic transitions in multistable networks and synchronization in balanced networks of excitatory and inhibitory neurons . The mesoscopic equations are employed to rapidly integrate a model of a cortical microcircuit consisting of eight neuron types , which allows us to predict spontaneous population activities as well as evoked responses to thalamic input . Our theory establishes a general framework for modeling finite-size neural population dynamics based on single cell and synapse parameters and offers an efficient approach to analyzing cortical circuits and computations .
When neuroscientists report electrophysiological , genetic , or anatomical data from a cortical neuron , they typically refer to the cell type , say , a layer 2/3 fast-spiking interneuron , a parvalbumin-positive neuron in layer 5 , or a Martinotti cell in layer 4 , together with the area , say primary visual cortex or somatosensory cortex [1–4] . Whatever the specific taxonomy used , the fact that a taxonomy is plausible at all indicates that neurons can be viewed as being organized into populations of cells with similar properties . In simulation studies of cortical networks with spiking neurons , the number of different cell types , or neuronal populations , per cortical column ranges from eight [5] to about 200 [6] with 31’000 to 80’000 simulated neurons for one cortical column , but larger simulations of several columns adding up to a million neurons ( and 22 cells types ) have also been performed [7] . In the following , we will refer to a model where each neuron in each population is simulated explicitly by a spiking neuron model as a microscopic model . On a much coarser level , neural mass models [8–10] , also called field models [11–13] , population activity equations [14] , rate models [15] , or Wilson-Cowan models [16] represent the activity of a cortical column at location x by one or at most a few variables , such as the mean activity of excitatory and inhibitory neurons located in the region around x . Computational frameworks related to neural mass models have been used to interpret data from fMRI [17 , 18] and EEG [9] . Since neural mass models give a compact summary of coarse neural activity , they can be efficiently simulated and fit to experimental data [17 , 18] . However , neural mass models have several disadvantages . While the stationary state of neural mass activity can be matched to the single-neuron gain function and hence to detailed neuron models [11 , 14 , 19–22] , the dynamics of neural mass models in response to a rapid change in the input does not correctly reproduce a microscopic simulation of a population of neurons [14 , 22 , 23] . Second , fluctuations of activity variables in neural mass models are either absent or described by an ad hoc noise model . Moreover , the links of neural mass models to local field potentials are difficult to establish [24] . Because a systematic link to microscopic models at the level of spiking neurons is missing , existing neural mass models must be considered as heuristic phenomenological , albeit successful , descriptions of neural activity . In this paper we address the question of whether equations for the activity of populations , similar in spirit but not necessarily identical to Wilson-Cowan equations [16] , can be systematically derived from the interactions of spiking neurons at the level of microscopic models . At the microscopic level , we start from generalized integrate-and-fire ( GIF ) models [14 , 25 , 26] because , first , the parameters of such GIF models can be directly , and efficiently , extracted from experiments [27] and , second , GIF models can predict neuronal adaptation under step-current input [28] as well as neuronal firing times under in-vivo-like input [26] . In our modeling framework , the GIF neurons are organized into different interacting populations . The populations may correspond to different cell types within a cortical column with known statistical connectivity patterns [3 , 6] . Because of the split into different cell types , the number of neurons per population ( e . g . , fast-spiking inhibitory interneurons in layer 2/3 ) is finite and in the range of 50–2000 [3] . We call a model at the level of interacting cortical populations of finite size a mesoscopic model . The mathematical derivation of the mesoscopic model equations from the microscopic model ( i . e . network of generalized integrate-and-fire neurons ) is the main topic of this paper . The small number of neurons per population is expected to lead to characteristic fluctuations of the activity which should match those of the microscopic model . The overall aims of our approach are two-fold . As a first aim , this study would like to develop a theoretical framework for cortical information processing . The main advantage of a systematic link between neuronal parameters and mesoscopic activity is that we can quantitatively predict the effect of changes of neuronal parameters in ( or of input to ) one population on the activation pattern of this as well as other populations . In particular , we expect that a valid mesoscopic model of interacting cortical populations will become useful to predict the outcome of experiments such as optogenetic stimulation of a subgroup of neurons [29–31] . A better understanding of the activity patterns within a cortical column may in turn , after suitable postprocessing , provide a novel basis for models of EEG , fMRI , or LFP [9 , 17 , 18 , 24] . While we cannot address all these points in this paper , we present an example of nontrivial activity patterns in a network model with stochastic switching between different activity states potentially linked to perceptual bistability [32–34] or resting state activity [18] . As a second aim , this study would like to contribute to multiscale simulation approaches [35] in the neurosciences by providing a new tool for efficient and consistent coarse-grained simulation at the mesoscopic scale . Understanding the computations performed by the nervous system is likely to require models on different levels of spatial scales , ranging from pharmacological interactions at the subcellular and cellular levels to cognitive processes at the level of large-scale models of the brain . Ideally , a modeling framework should be efficient and consistent across scales in the following sense . Suppose , for example , that we are interested in neuronal membrane potentials in one specific group of neurons which receives input from many other groups of neurons . In a microscopic model , all neurons would be simulated at the same level; in a multi-scale simulation approach , only the group of neurons where we study the membrane potentials is simulated at the microscopic level , whereas the input from other groups is replaced by the activity of the mesoscopic model . A multiscale approach is consistent , if the replacement of parts of the microscopic simulation by a mesoscopic simulation does not lead to any change in the observed pattern of membrane potentials in the target population . The approach is efficient if the change of simulation scale leads to a significant speed-up of simulation . While we do not intend to present a systematic comparison of computational performance , we provide an example and measure the speed-up factor between mesoscopic and microscopic simulation for the case of a cortical column consisting of eight populations [5] . Despite of its importance , a quantitative link between mesoscopic population models and microscopic neuronal parameters is still largely lacking . This is mainly due to two obstacles: First , in a cortical column the number of neurons of the same type is small ( 50–2000 [3] ) and hence far from the N → ∞ limit of classic “macroscopic” theories in which fluctuations vanish [14 , 36–38] . Systematic treatments of finite-size networks using methods from statistical physics ( system size expansion [39] , path integral methods [40 , 41] , neural Langevin equations [42–45] ) have so far been limited to simplified Markov models that lack , however , a clear connection to single neuron physiology . Second , spikes generated by a neuron are generally correlated in time due to refractoriness [46] , adaptation and other spike history dependencies [28 , 47–51] . Therefore spike trains are often not well described by an ( inhomogeneous ) Poisson process , especially during periods of high firing rates [46] . As a consequence , the mesoscopic population activity ( i . e . the sum of spike trains ) is generally not simply captured by a Poisson model either [52–54] , even in the absence of synaptic couplings [55] . These non-Poissonian finite-size fluctuations on the mesoscopic level in turn imply temporally correlated synaptic input to other neurons ( colored noise ) that can drastically influence the population activity [53 , 54 , 56] but which is hard to tackle analytically [57] . Therefore , most theoretical approaches rely on a white-noise or Poisson assumption to describe the synaptic input [58–62] , thereby neglecting temporal correlations caused by spike-history dependencies in single neuron activity . Here , we will exploit earlier approaches to treating refractoriness [23] and spike frequency adaptation [55 , 63] and combine these with a novel treatment of finite-size fluctuations . Our approach is novel for several reasons . First , we use generalized integrate-and-fire models that accurately describe neuronal data [25 , 26] as our microscopic reference . Second , going beyond earlier studies [58–60 , 64] , we derive stochastic population equations that account for both strong neuronal refractoriness and finite population size in a consistent manner . Third , our theory has a non-perturbative character as it neither assumes the self-coupling ( refractoriness and adaptation ) to be weak [65] nor does it hinge on an expansion around the N → ∞ solution for large but finite N [55 , 66 , 67] . Thus , it is also valid for relatively small populations and non-Gaussian fluctuations . And forth , in contrast to linear response theories [55 , 68–72] , our mesoscopic equations work far away from stationary states and reproduce large fluctuations in multistable networks . In the Results section we present our mesoscopic population equations , suggest an efficient numerical implementation , and illustrate the main dynamical effects via a selected number of examples . To validate the mesoscopic theory we numerically integrate the stochastic differential equations for the mesoscopic population activities and compare their statistics to those of the population activities derived from a microscopic simulation . A detailed account of the derivation is presented in the Methods section . In the discussion section we point out limitations and possible applications of our mesoscopic theory .
A population α of size Nα is represented by its population activity A N α ( t ) ( Greek superscripts label the populations , Fig 1C ) defined as A N α ( t ) = 1 N α ∑ i = 1 N α s i α ( t ) . ( 1 ) Here , s i α ( t ) = ∑ k δ ( t - t i , k α ) with the Dirac δ-function denotes the spike train of an individual neuron i in population α with spike times t i , k α . Empirically , the population activity is measured with a finite temporal resolution Δt . In this case , we define the coarse-grained population activity as A N α ( t ) = Δ n α ( t ) N α Δ t , ( 2 ) where Δnα ( t ) is the number of neurons in population α that have fired in a time bin of size Δt starting at time t . The two definitions converge in the limit Δt → 0 . An example of population activities derived from spiking activity in a cortical circuit model under a step current stimulation is shown in Fig 1B . To bridge the scales between neurons and populations , the corresponding mean-field model should ideally result in the same population activities as obtained from the full microscopic model ( Fig 1D ) . Because of the stochastic nature of the population activities , however , the qualifier “same” has to be interpreted in a statistical sense . The random fluctuations apparent in Fig 1B and 1D are a consequence of the finite number of neurons because microscopic stochasticity is not averaged out in the finite sum in Eq ( 1 ) . This observation is important because estimated neuron numbers reported in experiments on local cortical circuits are relatively small [3 , 73] . Therefore , a quantitatively valid population model needs to account for finite-size fluctuations . As mentioned above , we will refer to the population-level with finite size populations ( N ∼ 50 to 2000 per population ) as the mesoscopic level . In summary , we face the following question: is it possible to derive a closed set of evolution equations for the mesoscopic variables A N α ( t ) that follow the same statistics as the original microscopic model ? To address this question , we describe neurons by generalized integrate-and-fire ( GIF ) neuron models ( Fig 1A ( inset ) and Methods , Sec . “Generalized integrate-and-fire model” ) with escape noise [14] . In particular , neuron i of population α is modeled by a leaky integrate-and-fire model with dynamic threshold [49 , 80] . The variables of this model are the membrane potential u i α ( t ) and the dynamic threshold ϑ i α ( t ) = u th + ∫ - ∞ t θ α ( t - t ′ ) s i α ( t ′ ) d t ′ ( Fig 1A , inset ) , where uth is a baseline threshold and θα ( t ) is a spike-triggered adaptation kernel or filter function that accounts for adaptation [26 , 47 , 81–84] and other spike-history effects [14 , 84] via a convolution with the neurons spike train s i α ( t ) . In other words , the dynamic threshold depends on earlier spikes t i , k α of neuron i: ϑ i α ( t ) ≡ ϑ α ( t , t i , k α < t ) . Spikes are elicited stochastically depending on the present state of the neuron ( Fig 1A , inset ) . Specifically , the probability that neuron i fires a spike in the next time step [t , t + Δt] is given by λi ( t ) Δt , where λ i α ( t ) is the conditional intensity of neuron i ( also called conditional rate or hazard rate ) : λ i α ( t ) = f α u i α ( t ) - ϑ α ( t , t i , k α < t ) ( 3 ) with an exponential function f α ( x ) = c α exp ( x / Δ u α ) . Analysis of experimental data has shown that the “softness” parameter Δ u α of the threshold is in the range of 1 to 5 mV [85] . The parameter cα can be interpreted as the instantaneous rate at threshold . Besides the effect of a spike on the threshold as mediated by the filter function θα ( t ) , a spike also triggers a change of the membrane potential . In the GIF model ( Methods , Sec . “Generalized integrate-and-fire model” ) , the membrane potential u i α ( t ) is reset after spiking to a reset potential ur , to which u i α ( t ) is clamped for an absolute refractory period tref . Absolute refractoriness is followed by a period of relative refractoriness , where the conditional intensity Eq ( 3 ) is reduced . This period is determined by the relaxation of the membrane potential from the reset potential to the unperturbed or “free” potential , denoted h ( t ) , which corresponds to the membrane potential dynamics in the absence of resets . The GIF model accurately predicts spikes of cortical neurons under noisy current stimulation mimicking in-vivo like input [25 , 26] and its parameters can be efficiently extracted from single neuron recordings [26 , 27] . Variants of this model have also been suggested that explicitly incorporate biophysical properties such as fast sodium inactivation [86 , 87] , conductance-based currents [88] and synaptically-filtered background noise [89] . For a first analysis of the finite-size effects , we consider the special case of a fully-connected network of Poisson neurons with absolute refractory period [14] . In this case , the conditional intensity can be represented as λ A ( t | t ^ ) = f ( h ( t ) ) Θ ( t - t ^ - t ref ) , where tref is the absolute refractory period , Θ ( ⋅ ) is the Heaviside step function and h ( t ) is the free membrane potential , which obeys the passive membrane dynamics τ m d h d t = - h + μ ( t ) + τ m J ( ϵ * A N ) ( t ) , ( 16 ) where τm is the membrane time constant , μ ( t ) = urest + RI ( t ) ( where urest is the resting potential and R is the membrane resistance ) accounts for all currents I ( t ) that are independent of the population activities , J is the synaptic strength and ϵ ( t ) is a synaptic filter kernel ( see Methods , Eq ( 27 ) for details ) . For the mathematical analysis , we assume that the activity AN ( t ) and input μ ( t ) have started at t = −∞ so that we do not need to worry about initial conditions . In a simulation , we could for example start at t = 0 with initial conditions AN ( t ) = δ ( t ) for t ≤ 0 and h ( 0 ) = 0 . For the conditional intensity given above , the effective rate Λ ( t ) , Eq ( 11 ) , is given by Λ ( t ) = f ( h ( t ) ) because the variance v ( t , t ^ ) is zero during the absolute refractory period t - t ref ≤ t ^ < t . As a result , the mesoscopic population Eq ( 13 ) reduces to the simple form A ¯ ( t ) = f ( h ( t ) ) 1 - ∫ t - t ref t A N ( t ^ ) d t ^ . ( 17 ) This mesoscopic equation is exact and could have been constructed directly in this simple case . For N → ∞ , where AN ( t ) becomes identical to A ¯ ( t ) , this equation has been derived by Wilson and Cowan [16] , see also [14 , 23 , 92] . The intuitive interpretation of Eq ( 17 ) is that the activity at time t consists of two factors , the “free” rate λfree ( t ) = f ( h ( t ) ) that would be expected in the absence of refractoriness and the fraction of actually available ( “free” ) neurons that are not in the refractory period . For finite-size populations , these two factors explicitly reveal two distinct finite-size effects: firstly , the free rate is driven by the fluctuating population activity AN ( t ) via Eq ( 16 ) and hence the free rate exhibits finite-size fluctuations . This effect originates from the transmission of the fluctuations through the recurrent synaptic connectivity . Secondly , the fluctuations of the population activity impacts the refractory state of the population , i . e . the fraction of free neurons , as revealed by the second factor in Eq ( 17 ) . In particular , a large positive fluctuations of AN in the recent past reduces the fraction of free neurons , which causes a negative fluctuation of the number N A ¯ ( t ) Δ t of expected firings in the next time step . Therefore , refractoriness generates negative correlations of the fluctuations 〈ΔA ( t ) ΔA ( t′ ) 〉 for small |t − t′| . We note that such a decrease of the expected rate would not have been possible if the correction term in Eq ( 13 ) was absent . However , incorporating the effect of recent fluctuations ( i . e . fluctuations in the number of refractory neurons ) on the number of free neurons by adding the correction term , and thereby balancing the total number of neurons , recovers the correct Eq ( 17 ) . The same arguments can be repeated in the general setting of Eq ( 13 ) . Firstly , the conditional intensity λ A ( t | t ^ ) depends on the past fluctuations of the population activity because of network feedback . Secondly , the fluctuations lead to an imbalance in the number of neurons across different states of relative refractoriness ( i . e . fluctuations do not add up to zero ) which gives rise to the “correction term” , i . e . the second term on the r . h . s . of Eq ( 13 ) . We wondered how well the statistics of the population activities obtained from the integration of the mesoscopic equations compare with the corresponding activities generated by a microscopic simulation . As we deal with a finite-size system , not only to the first-order statistics ( mean activity ) but also higher-order statistics needs to be considered . Because there are several approximations involved ( e . g . full connectivity , quasi-renewal approximation and effective rate of fluctuations in the refractory density ) , we do not expect a perfect match . To compare first- and second-order statistics , we will mainly use the power spectrum of the population activities in the stationary state ( see Methods , Sec . “Power spectrum” ) .
Our theory provides a general framework to replace spiking neural networks that are organized into homogeneous populations by a network of interacting mesoscopic populations . For example , the excitatory and inhibitory neurons of a layer of a cortical column [5] may be represented by one population each , as in Fig 1 . Weak heterogeneity in the neuronal parameters are allowed in our theory because the mesoscopic equations describe the population-averaged behavior . Further subdivisions of the populations are possible , however , such as a subdivision of the inhibitory neurons into fast-spiking and non fast-spiking types [26] . Populations that show initially a large degree of heterogeneity can be further subdivided into smaller populations . In this case , a correct description of finite-size fluctuations , as provided by our theory , will be particularly important . However , as with any mean-field theory , we expect that our theory breaks down if neural activity and information processing is driven by a few “outlier” neurons such that a mean-field description becomes meaningless . Further limitations may result from the mean-field and quasi-renewal approximation , Eq ( 4 ) . Formally , the mean-field approximation of the synaptic input requires dense connectivity and the heterogeneity in synaptic efficacies and in synapse numbers to be weak . Moreover , the quasi-renewal approximation assumes slow threshold dynamics . However , as we have demonstrated here , our mesoscopic population equations can provide in concrete applications excellent predictions even for sparse connectivity ( Figs 5D–5G , 8 and 9 ) and may qualitatively reproduce the mesoscopic statistics in the presence of fast threshold dynamics ( Fig 4D and 4E ) . Using our mesoscopic population equations it is possible to make specific predictions about the response properties of local cortical circuits . For instance , recent progress in genetic methods now enables experimentalists to selectively label and record from genetically identified cell types , such as intratelencephalic ( IT ) , pyramidal tract ( PT ) and corticothalamic ( CT ) neurons among the excitatory neurons , and vasoactive intestinal peptide ( VIP ) , somatostatin ( Sst ) and parvalbumin ( Pvalb ) expressing neurons among the interneurons [4] . These cell types have received much attention recently as it has been proposed that they may form a basic functional module of cortex , the canonical circuit [4 , 110] . The genetic classification of cells defines subpopulations of the cortical network . A model of the canonical circuits of the cortex in terms of interacting mesoscopic populations can be particularly useful if used to describe experiments that use optogenetic stimulation of genetically-defined populations by light , which in our framework can be represented as a transient external input current . To build a mesoscopic population model based on our theory demands some assumptions about microscopic parameters such as ( i ) typical neuron parameters for each subpopulation , ( ii ) structural parameters as characterized by average synaptic efficacies and time scales of connections between and within populations , and ( iii ) estimates of neuron numbers per subpopulation . Parameters for a typical neuron of each population could be extracted by the efficient fitting procedures presented in [26 , 27] . Structural parameters and neuron numbers have been estimated , for instance , for barrel columns in rodents somato-sensory cortex [3 , 73] and other studies ( see e . g . , [5] ) . Our population equations could then be used to make predictions about circuit responses to light stimuli , e . g . by imaging the activity of a genetically-defined subpopulation in one column in response to optogenetic stimulation of another cell class in another column . As a first step in this direction , we have demonstrated here that our population equations correctly predict the mesoscopic activities ( means and fluctuations ) of a simulation of a detailed , microscopic network model of a cortical microcircuit [5] under thalamic stimulation of layer 4 and 6 neurons . Using a population density method , mean activities of this model have also been predicted in a recent study to analyze its computational properties [107] with a special focus on predictive coding . Our population density approach goes beyond that study by also predicting finite-size fluctuations of the activities and their effects on the mesoscopic network dynamics such as finite-size induced stochastic oscillations . Predicting activities in real experiments is , however , complicated by the fact that the parameters of a microscopic network model are typically underconstrained given the lack or uncertainty of measured or estimated parameters [111] . Here , our population equations provide an efficient means to constrain unknown microscopic parameters by requiring consistence with mesoscopic experimental data . While the canonical circuit represents a model of interacting populations on the mesoscopic level , recent interest in macroscopic models of entire brain areas or even of whole brains has risen [6 , 7] . Population dynamics can be used in this context as a means to reduce large parts of the macroscopic neuronal network to a system of interacting populations that is numerically manageable , and requires less detailed knowledge of synaptic connectivity ( mean synaptic coupling of populations as opposed to individual synapses ) . However , even this information about mesoscopic network structure might not be available given that it corresponds to an M × M matrix of mean synaptic efficacies , where the number M of populations , or respectively cell types , might be large . In this case , our population equations can be utilized to efficiently constrain unknown structural parameters , such as synaptic weights , such that the resulting mesoscopic activities are consistent with experimental data . This leads in turn to experimentally testable predictions for synaptic connectivities . Such an approach [111] has been recently applied to a network model of primate visual cortex demonstrating the usefulness of mean-field theories for predicting structural properties of large-scale cortical networks . An interesting complementary route for further studies is a multiscale model , in which a large-scale model is simulated in terms of reduced , mesoscopic populations but with one or several areas in focus that are simulated in full microscopic detail . As knowledge of anatomy and computational capacity increases , more and more mesoscopic populations can be replaced by a microscopic simulation , while at any time in this process the full system is represented in the model . We therefore expect our population dynamics model to be a useful tool to continuously integrate experimental data into multiscale models of whole mammalian brains . Simplified whole brain models of interacting neuronal areas have recently been proposed [112 , 113] . Furthermore , large-scale neuro imaging data are routinely modeled by phenomenological population models such as neural mass , Wilson-Cowan , or neural field models [9 , 22] . Our new population dynamics theory could be used in such approaches as an accurate representation of the fluctuations of neural activity in the reduced areas . For example , in macroscopic data such as resting state fMRI , EEG or MEG , the endogenously generated fluctuations of brain activity are of major interest [113] . A fortiori the same applies to mesoscopic data such as local field potentials ( LFP ) or voltage-sensitive dye ( VSD ) , in which finite-size fluctuation are expected to be large . Our theory paves the way for relating macroscale fluctuations to the underlying networks of spiking neurons and their activity , and so to the neuronal circuits that underlie the computations of the brain . Another interesting application of our population model is to predict the activity of neural networks grown in cultures . This model system is much more accessible and controllable ( e . g . , by optogenetic stimulations ) than cortical networks in-vivo but may still provide valuable insights into the complex network activity of excitatory and inhibitory neurons as proposed in a recent study [114] . In particular , in that study the authors propose a critical role for short-term plasticity [115] . Although we have here used static synapses , an extension of our mesoscopic mean-field theory to synaptic short-term plasticity is feasible . Furthermore , finite-size fluctuations appear to be particularly important in cell cultures as suggested by a previous theoretical study [62] . Our mesoscopic population theory thus represents a framework to predict spontaneous as well as evoked activity in neuronal cell cultures . From a theoretical point of view , our study represents a generalization of deterministic , macroscopic population equations for an infinite number of spiking neurons with refractoriness [14 , 23 , 63] to stochastic , mesoscopic population equations for a finite number of neurons . The resulting dynamics can be directly used to generate single stochastic realizations of mesoscopic activities , in analogy to a Langevin dynamics . Our work is thus conceptually different from earlier studies of finite-size effects [66–68] , who also considered finite networks of spiking neurons and refractoriness but derived deterministic evolution equations for moment and cross-correlation functions and hence characterized the ensemble dynamics . Furthermore , in contrast to these studies , our theory is not based on a perturbation expansion around the N → ∞ limit , and thus captures large and non-Gaussian fluctuations in strongly nonlinear population dynamics such as bistable networks . Outside the low-rate Poisson firing regime , spiking neurons exhibit history dependencies in their spike trains , the most prominent of which is neuronal refractoriness , i . e . the strongly reduced firing probability depending on the time since the last spike . On the population level this means that a positive ( negative ) fluctuation of the population rate affects the underlying refractory state of the population because more ( less ) neurons than expected become refractory . This altered refractory state in turn tends to decrease ( increase ) the mean and variance of the population activity shortly after the fluctuation . More generally , fluctuations of the population activity influence the population density of state variables , which in turn influences fluctuations . In this study , we have worked out how to incorporate this interplay between fluctuations of the population activity and fluctuations of the refractory density into a mesoscopic population dynamics . The key insight to achieve this was ( i ) to exploit the normalization condition for the density of microscopic states ( in our case , the density of last spike times ) and ( ii ) to associate density fluctuations with a time-dependent but state-independent average rate that emphasizes the microscopic rates of those states that exhibit the largest finite-size fluctuations ( in our case , the weighted average rate with respect to the variance v ( t , t ^ ) ) . Our work is thus in marked contrast to previous stochastic rate models for finite-size systems in the form of stochastic Wilson-Cowan equations [62 , 104] , or stochastic neural field equations [39 , 116] . In these models , finite-size fluctuations of the rate may feed back through the recurrent connections but the strong negative self-feedback due to refractoriness is neglected . This is the case even if the stationary or dynamic transfer function employed in the rate dynamics corresponds to a spiking neuron model [62 , 72] . Furthermore , fluctuations of the population rate have often been implemented ad hoc by a phenomenological white-noise source , which was added to the macroscopic ( i . e . deterministic ) rate dynamics [32 , 102 , 112] . The intensity of the noise is a free parameter in these cases . Our mesoscopic equations are also driven by a noise source , but two differences to these studies are noteworthy: First , it is derived from a microscopic model and does not contain any free parameter; and second , the noise is white given the predicted mean activity but since the activity predicted in one time step depends on fluctuations in all earlier time steps , the effective noise leads to a colored noise spectrum—even if coupling is removed ( see Fig 3 ) . This observation is consistent with previous studies [55 , 58 , 60 , 69] , in which the power spectrum of the fluctuations about a steady-state has been calculated analytically . On the population level , refractoriness can be taken into account by population density equations such as the Fokker-Planck equation for the membrane potential density ( see e . g . [36 , 58 , 59] , or [107 , 117 , 118] for related master equations ) , or the population integral equation for the refractory density [14 , 23 , 88 , 89] . These studies were mainly concerned with macroscopic populations , which formally correspond to the limit N → ∞ . For the refractory density formalism , we have shown here how to extend the population integral equation to the case of finite population size . To this end , we corrected for the missing normalization of the mesoscopic density ( e . g . Q ( t , t ^ ) = S ( t | t ^ ) A N ( t ^ ) in Eq ( 13 ) or q ( t , t ^ ) in Eq ( 80 ) ) , and thereby accounted for the interplay between fluctuations and refractoriness . Finite-size fluctuations of the population rate have also been used in the Fokker-Planck formalism [58–60] but the immediate effect of these fluctuations on the membrane potential density at threshold , and hence the refractoriness , has been neglected: in fact , a positive ( negative ) fluctuation of the population rate increases ( decreases ) the number of neurons at the reset potential while the number of neurons close to the threshold has to decrease ( increase ) such that the microscopic density remains normalized . The finite-N membrane potential density used by Mattia and Del Giudice [60] does not account for this normalization effect . Whereas the numerical integration of their equation may still give a satisfying solution in the low-rate , Poissonian-firing regime , where refractory effects can be neglected , it becomes unstable at higher rates unless the density is renormalized manually at every time step [119] . How to correct for the missing normalization in the Fokker-Planck approach is still an unsolved theoretical question . In this respect , using analogies to and insights from our approach might be promising . The quasi-renewal approximation [55 , 63] allowed us to develop a finite-size theory for an effectively one-dimensional population density equation even in the presence of adaptation . Here , the only microscopic state variable is the last spike time t ^ , or equivalently the age of the neuron τ = t - t ^ . Longer lasting spike history effects such as adaptation are captured by the dependence of the conditional intensity on the population activity AN , which as a mesoscopic mean-field variable does not need to be treated as a state variable . Furthermore , Chizhov and Graham have shown that the one-dimensional population density method in terms of the age τ can also capture multiple gating variables in conductance-based neuron models with adaptation [88] . Such one-dimensional descriptions have great advantages compared to population density equations that include adaptation by additional state variables and which thus require a multi-dimensional state-space [37 , 120–122]: Firstly , the numerical solution of the density equations grows exponentially with the number of dimensions and becomes quickly infeasible if multiple adaptation variables are needed as e . g . in the case of multi-timescale adaptation [28] or if an adiabatic approximation of slow variables [64 , 120 , 123] is not possible . Secondly , it is unclear how to treat finite-size fluctuations in the multi-dimensional case . Our theory is based on an effective fully-connected network , in which neurons are coupled by the actual realization of the stochastic population activity ( the “mean field” ) , both in the microscopic and mesoscopic model . Thus , in the limit of a fully-connected network , the problem of self-consistently matching the input and output statistics , which arises in mean-field theories , is automatically satisfied to any order by our finite-size theory . This is in marked contrast to the opposite limit of a sparsely-connected network [59] . In that case , the mean-field variables correspond to the statistics of the spike trains ( e . g . rate and auto-correlation function ) rather than to the actual realization of the population activity . These statistics must be matched self-consistently for input and output , which is a hard theoretical problem [56 , 57 , 124 , 125] . Between these two limit cases , where the network is randomly connected with some finite connection probability 0 < p < 1 , our examples ( Figs 5 , 8 and 9 ) indicate that the approximation by an effective fully-connected network can still yield reasonable results even for relatively sparse networks with p = 5% . We emphasize that in our microscopic network model we used a fixed in-degree in order to avoid additional variability due to the quenched randomness in the number of synapses . This allowed us to focus on dynamic finite-size noise in homogeneous populations and its interactions with refractoriness . In contrast , the heterogeneity caused by the quenched randomness is a further finite-size effect [58] that needs to be examined in a future study . As an integral equation , the mesoscopic population model is formally infinitely dimensional and represents a non-Markovian dynamics for the population activity AN . Such complexity is expected given that the derived population equations are general and not limited to a specific dynamical regime . Loosely speaking , the equations must be rich enough , and hence sufficiently complex , in order to reproduce the rich repertoire of dynamical regimes that fully connected networks of spiking neurons are able to exhibit ( e . g . limit cycles , multi-stability , cluster states [23] ) . For a mathematical analysis , however , it is often desirable to have a low-dimensional representation of the population dynamics in terms of a few differential equations , at least for a certain parameter range . Apart from the dynamics in the neighborhood of an equilibrium point ( see e . g . [126] ) or in the limit of slow synapses [127] , such “firing rate models” are difficult to link to the microscopic model already in the deterministic ( macroscopic ) case ( for notable exceptions see [128 , 129] ) , let alone the stochastic , finite-size case . Here , our mesoscopic population rate equations can serve as a suitable starting point for deriving low-dimensional dynamics that links microscopic models to mesoscopic rate equations with realistic finite-size noise . There are several ways to extend our mesoscopic population model towards more biological realism . We already mentioned the possibility to include short-term synaptic plasticity in our mean-field framework . Furthermore , the hazard function could be generalized to capture Gaussian current noise as arising from background spiking activity [5 , 57–60 , 98] . Approximate mappings of white and colored current noise to an effective hazard function in the escape noise formalism are available [88 , 89] and might be combined with our mesoscopic population model . Yet another extension concerns the synaptic input model . Here we only looked at current input but , as shown by Chizhov and Graham [88] , it is straightforward to extend population theories of the renewal type to the case of conductance inputs . In the simplest case , the synaptic current of neuron i embedded in population α and driven by populations β can be modelled by a linear voltage-dependence: I syn , i α ( t ) = - ∑ β = 1 M u i α ( t ) - E α β ∑ j ∈ Γ i β g α β * s j β ( t ) ( 18 ) ( cf . corresponding expression Eq ( 22 ) in Methods for current-based synapses ) . Here , Eαβ is the reversal potential of a synapse from population β , and gαβ ( t ) is the conductance response ( in nS ) elicited by a spike of a presynaptic neuron in population β . The same mean-field arguments as for the current-based model carry over to the case of conductance-based synapses . For example , the membrane potential u A α ( t , t ^ ) of a current-based leaky integrate-and-fire neuron with a last spike time at time t ^ follows the equation τ m α ∂ u A α ∂ t = - u A α + μ α ( t ) + τ m α ∑ β = 1 M p α β N β w α β ( ϵ α β * A N β ) ( t ) , ( 19 ) where at t = t ^ and during an absolute refractory period u A ( t , t ^ ) = u r is at the reset potential ( see Methods , Eq ( 30 ) for details ) . In the case of conductance-based input , Eq ( 18 ) , we only need to replace Eq ( 19 ) by τ m α ∂ u A α ∂ t = - u A α + μ α ( t ) - R α ∑ β = 1 M p α β N β u A α - E α β ( g α β * A N β ) ( t ) . ( 20 ) where Rα is the membrane resistance . How to model nonlinear voltage-dependence of synaptic currents such as N-methyl-D-aspartate ( NMDA ) currents within a mean-field approximation is less obvious but approximations also exist for this case [20] . It will be an interesting question for the future how well these approaches work with the finite-N theory developed in the present study . Alternatively , effective current models [130 , 131] with activity-dependent , effective time constant τm ( t ) and effective resting potential urest ( t ) could be another solution to treat conductance inputs . Here , we have used a discrete set of populations . In large-scale models of the brain , one often regards the spatial continuum limit , resulting in so-called stochastic neural field equations [116] . These equations represent a compact description of neural activity and do not depend on a specific discretization of space . Just as discrete firing rate models , these field equations must be considered phenomenological because the link to neuronal parameters is not clear ( note however that such equations have been derived from non-spiking , two-state neuron models for N < ∞ [39] , and from spiking models for N → ∞ [132 , 133] ) . By taking the spatial continuum limit , our mesoscopic population equations can be formulated as a stochastic neural field equation that is directly derived from a finite-size , spiking neural network . It would be interesting to employ this continuous extension of our mesoscopic equations to study the effect of spike-history effects on the stochastic behavior of bumps and waves in neural fields . A first simple comparison of the computational performance in Results , “Mesoscopic dynamics of cortical microcolumn” , demonstrated already that the mesoscopic population dynamics outperformed the microscopic simulation by a speed-up factor of around 120 . In this example , the numerical integration of the population dynamics has not been particularly optimized with respect to time step Δt and history length T . A systematic comparison under the condition of some given accuracy , has the potential for an even larger speed-up because the population equations can be integrated with a larger time step than the spiking neural network . In addition to that , we have also compared the mesoscopic model to the full microscopic simulation of the refractory density ( cf . Results , “Finite-size mean-field theory” ) and found a moderate enhancement in performance for sufficiently large networks ( N ≳ 100 ) . These computational aspects will be investigated in a separate study .
An important variable that characterizes the internal state of a neuron is the time of its last spike , or , equivalently , the time elapsed since the last spike ( “age” of the neuron ) [88] . The time since the last spike is a good predictor of the refractory state of a neuron at time t . Our approach is to use a population density description for this refractory state [23 , 68 , 88 , 89] , in which the coupling of neurons as well as the adaptation of single neurons are mediated by the mesoscopic population activities A N α ( t ) defined by Eq ( 1 ) . To this end , we replace the conditional firing rate λ i α ( t ) of a neuron i in population α by an effective rate λ A α ( t | t ^ i α ) that only depends on its last spike time t ^ i α and the history of the population activity { A N α ( t ′ ) } t ′ < t [63] . Here and in the following , the subscript A indicates the dependence on the history of A N α ( t ) . We note that the expected total activity A ¯ α ( t ) of population α at time t is the average of all the conditional firing rates summed over all neurons in this population: A ¯ α ( t ) = ( 1 / N α ) ∑ i λ i α ( t ) . The effective rate λ A α ( t | t ^ i α ) is determined such that it approximates the conditional intensity on average: 1 N α ∑ i = 1 N α λ i α ( t ) ≈ 1 N α ∑ i = 1 N α λ A α ( t | t ^ i α ) . ( 28 ) To find such an approximation , we proceed in two steps [55]: first , the membrane potential u i α ( t ) is approximated by a function u A α ( t , t ^ i α ) using a mean-field approximation of the synaptic input . For fully connected populations , this first approximation turns into an exact statement . Second , the dynamic threshold ϑ i α ( t ) is approximated by a function ϑ A α ( t , t ^ i α ) using the quasi-renewal approximation [63] . For renewal neurons , the second approximation becomes exact . Once we have found an expression for the mean-field approximation Eq ( 28 ) , we are in a position to use a population density description with respect to the last spike times t ^ i α . In the following two paragraphs we explain the above two steps in detail . Using the mean-field approximation Eq ( 32 ) , we have reduced the model to a population of time-dependent renewal processes [23 , 68] , where the conditional intensity of neuron i is λ A α ( t | t ^ i α ) . Neurons are effectively coupled through the dependence of λ A α ( t | t ^ i α ) upon the membrane potential u A α ( t | t ^ i ) , which in turn depends on the activities A N β of all populations β that are connected to population α . This is the only place where population labels different from α appear . For the sake of notational simplicity , we will omit the population label α and the subscript A in this section , keeping in mind that all quantities refer to population α and that the coupling with other populations is implicitly contained in u A α ( t | t ^ i ) . In continuous time , we consider the rescaled variables A N ( t l ) = Δ n ( t l ) N Δ t , A ¯ ( t l ) = Δ n ¯ ( t l ) N Δ t . ( 60 ) Here , A ¯ ( t ) can be interpreted as the expected population activity given the past activity AN ( t′ ) , t′ < t . For Δt small but positive , the spike count Δn ( t ) is an independent Poisson number with mean Δ n ¯ ( t ) = N A ¯ ( t ) Δ t . Thus , on a coarse-grained time scale , the continuum limit of the population activity may be written in the following suggestive way A N ( t ) = d n ( t ) N d t , d n ( t ) ∼ Pois ( N A ¯ ( t ) d t ) , ( 61 ) where dt denotes an infinitesimal ( but temporally coarse-grained ) time step and dn ( t ) is an independent Poisson-distributed random number with mean N A ¯ ( t ) d t . In the limit dt → 0 , the spike count in an infinitesimal time step is a Bernoulli random number , where dn ( t ) = 1 with probability N A ¯ ( t ) d t and n ( t ) = 0 with probability 1 - N A ¯ ( t ) d t . Therefore , in this limit the population activity AN ( t ) converges to a sequence of Dirac δ-functions occurring at random times tpop , k with rate N A ¯ ( t ) . Thus , AN ( t ) can be written more formally as a population spike train or “shot-noise” A N ( t ) = 1 N ∑ k δ ( t - t pop , k ) , ( 62 ) where ( t pop , k ) k ∈ Z is a point process with conditional intensity λ pop ( t | H t ) = N A ¯ ( t ) . Here , the condition H t denotes the history of the population activity {AN ( t′ ) }t′<t , or equivalently , the history of spike times {tpop , k}tpop , k<t , up to ( but not including ) time t . To obtain the dynamics for A ¯ ( t ) , we also introduce the rescaled variables S ( t l | t k ) = ⟨ m ^ l , k ⟩ Δ n ( t k ) , v ( t l , t k ) = ⟨ Δ m ^ l , k 2 ⟩ N Δ t . ( 63 ) The function S ( t | t ^ ) can be interpreted as the survival probability of neurons that have fired their last spike at time t ^ . Furthermore , for small Δt the firing probability is given by Pλ ( tl|tk ) = λ ( tl|tk ) Δt + O ( Δt2 ) . Thus , the continuum limit of Eq ( 59 ) reads A¯ ( t ) =limΔt→0{ ∑k=−∞tΔt−1λ ( t| kΔt ) S ( t |kΔt ) Δn ( kΔt ) N +∑k=−∞tΔt−1λ ( t|kΔt ) v ( t , kΔt ) Δt∑k=−∞tΔt−1v ( t , kΔt ) Δt ( 1−∑k=−∞tΔt−1S ( t|kΔt ) Δn ( kΔt ) N ) } . ( 64 ) The sums in this equation can be regarded as the definition of stochastic integrals , which allows us to rewrite Eq ( 64 ) as A ¯ ( t ) = ∫ - ∞ t λ ( t | t ^ ) S ( t | t ^ ) A N ( t ^ ) d t ^ + Λ ( t ) 1 - ∫ - ∞ t S ( t | t ^ ) A N ( t ^ ) d t ^ . ( 65 ) Here , Λ ( t ) = ∫ - ∞ t λ ( t | t ^ ) v ( t , t ^ ) d t ^ ∫ - ∞ t v ( t , t ^ ) d t ^ ( 66 ) is an effective rate corresponding to the effective firing probability PΛ ( t ) . Note that according to Eq ( 64 ) , the stochastic integrals in Eq ( 65 ) extend only over last spike times t ^ < t not including time t ^ = t . Taking the continuum limit of Eq ( 48 ) we find that the survival probability satisfies the differential equation ∂ S ( t | t ^ ) ∂ t = - λ ( t | t ^ ) S ( t | t ^ ) , S ( t ^ | t ^ ) = 1 . ( 67 ) This equation has the simple solution S ( t | t ^ ) = exp - ∫ t ^ t λ ( t ′ | t ^ ) d t ′ . ( 68 ) Similarly , we find from Eq ( 51 ) that the rescaled variance obeys the differential equation ∂ v ∂ t = - 2 λ ( t | t ^ ) v + λ ( t | t ^ ) S ( t | t ^ ) A N ( t ^ ) , v ( t ^ | t ^ ) = 0 . ( 69 ) The set of coupled Eqs ( 62 ) – ( 69 ) defines the mesoscopic population dynamics . We emphasize that not only AN ( t ) depends on A ¯ ( t ) ( cf . Eq ( 62 ) ) but that there is also a feedback of AN ( t ) onto the dynamics of A ¯ ( t ) , cf . Eq ( 65 ) . In fact , A ¯ ( t ) can be regarded as a deterministic functional of the past activities up to but not including time t . In particular , AN ( t ) is not an inhomogeneous Poisson spike train because the specific realization of the spike history of AN ( t ) determines the conditional intensity function for the point process ( tpop , k ) via Eq ( 65 ) . Furthermore , we note that , in the case of synaptic coupling or adaptation , also the variables S and v depend on the history of the population activity through the dependence of λ ( t | t ^ ) on the membrane potential u ( t , t ^ ) and the threshold ϑ ( t , t ^ ) ( cf . Eqs ( 30 ) and ( 33 ) ) . For large N , the population activity can be approximated by a Gaussian process . To this end , we note that in the discrete time description the spike counts Δn ( tl ) are conditionally independent random numbers with mean and variance N A ¯ ( t l ) Δ t . Therefore , in the large-N limit , the variable Δ W ( t l ) = Δ n ( t l ) - N A ¯ ( t l ) Δ t N A ¯ ( t l ) ( 70 ) is normally distributed with mean zero and variance Δt , and hence corresponds to the increment of a Wiener process . Using Eq ( 60 ) for the population activity and taking the continuum limit Δt → 0 , we obtain A N ( t ) = A ¯ ( t ) + A ¯ ( t ) N ξ ( t ) , ( 71 ) where ξ ( t ) = limΔt→0 ΔW ( t ) /Δt is a Gaussian white noise with correlation function 〈ξ ( t ) ξ ( t′ ) 〉 = δ ( t − t′ ) . This Gaussian approximation has the advantage that the multiplicative character of the noise in Eq ( 71 ) becomes explicit because ξ ( t ) is independent of the state of the system . It also explicitly reveals that the finite-size fluctuations scale like 1 / N . We stress again that AN ( t ) is not a white-noise process with time-dependent mean , as Eq ( 71 ) might suggest at first glance , but it is a sum of two mutually correlated processes , ( i ) a white-noise term proportional to ξ ( t ) that reflects the fact that the population activity is a δ-spike train and ( ii ) a colored “noise” A ¯ ( t ) that arises from the filtering of ξ ( t ) through the dynamics in Eq ( 65 ) . As a result , the auto-correlation function of AN ( t ) contains a δ-peak and a continuous part , consistent with previous theoretical findings [55] . In particular , at short lags the auto-correlation function may be negative as a result of refractoriness: in this case , ξ and A ¯ are anti-correlated in line with the intuitive picture discussed in the Results section , Fig 2 , that a positive fluctuation ξ ( t ) is associated with the creation of a “hole” in the distribution of last spike times leading to a reduced activity after time t . In the frequency domain , refractoriness corresponds to a trough in the power spectrum at low frequencies [94] as visible , for instance , in Fig 3 . These considerations clearly highlight the non-white character of the finite-size fluctuations in our theory . It is straightforward to generalize the population equations to several populations by adding a population label α = 1 , … , M . For the sake of completeness , we explicitly state the full set of equations . The activity of population α is given by A N α ( t ) = 1 N α ∑ k δ ( t - t pop , k α ) , ( 72 ) where ( t pop , k α ) k ∈ Z is a point process with conditional intensity λ pop α ( t | H t ) = N α A ¯ α ( t ) . Here , the expected activity A ¯ ( t ) depends explicitly on the history H t = { A N β ( t ′ ) } t ′ < t , β = 1 , … , M by the following set of equations A ¯ α ( t ) = ∫ - ∞ t λ α ( t | t ^ ) S α ( t | t ^ ) A N α ( t ^ ) d t ^ + Λ α ( t ) 1 - ∫ - ∞ t S α ( t | t ^ ) A N α ( t ^ ) d t ^ ( 73 ) λ α ( t | t ^ ) = c α exp u α ( t , t ^ ) - ϑ α ( t , t ^ ) Δ u α , Λ α ( t ) = ∫ - ∞ t λ α ( t | t ^ ) v α ( t , t ^ ) d t ^ ∫ - ∞ t v α ( t , t ^ ) d t ^ , ( 74 ) ∂ S α ∂ t= - λ α ( t | t ^ ) S α , S α ( t ^ | t ^ ) = 1 , ( 75 ) ∂ v α ∂ t= - 2 λ α ( t | t ^ ) v α + λ α ( t | t ^ ) S α ( t | t ^ ) A N α ( t ^ ) , v α ( t ^ , t ^ ) = 0 , ( 76 ) ∂ u α ∂ t= - u α - μ α ( t ) τ m α + ∑ β = 1 M w α β p α β N β ( ϵ α β * A N β ) ( t ) , u α ( t ^ , t ^ ) = u r ( 77 ) ϑ α ( t , t ^ ) = u th + θ α ( t - t ^ ) + ∫ - ∞ t ^ θ ˜ α ( t - t ′ ) A N α ( t ′ ) d t ′ . ( 78 ) For each population , the system of Eqs ( 73 ) – ( 78 ) contains a family of ordinary differential equations for the variables S , u and v parametrized by the continuous parameter t ^ with - ∞ < t ^ < t , and five integrals over this parameter . In the next section , we show that the family of ordinary differential equations is equivalent to three first-order partial differential equations . Furthermore , in Sec . “Population equations for a finite history” , we reduce the infinite integrals to integrals over a finite range , which will be useful for the numerical implementation of the population equations . There is an equivalent formulation of the population equation in terms of first-order partial differential equations for the density of ages τ = t - t ^ [23 , 68 , 88 , 89] . The representation in terms of age τ as a state variable is useful because it parallels the Fokker-Planck formalism for the membrane potential density [14 , 36 , 58–60] or related density equations [117 , 118] , in which the state variable is the membrane potential of a neuron . To keep the notation simple we consider in the following population α but drop the index α wherever confusion is not possible . Thus , we write e . g . S for Sα and AN for A N α but we keep the index β as well as double indices αβ occurring in Eq ( 77 ) . The density of ages at time t is defined as q ( t , τ ) = S ( t|t − τ ) AN ( t − τ ) . We recall that because of finite-size fluctuations , q is not a normalized probability density . Furthermore , we regard the functions λ , u and v as functions of t and τ . With these definitions the population equation , Eq ( 65 ) , can be rewritten as A ¯ ( t ) = ∫ 0 ∞ λ ( t , τ ) q ( t , τ ) d τ + Λ ( t ) 1 - ∫ 0 ∞ q ( t , τ ) d τ . ( 79 ) The stochastic activity AN ( t ) then follows from Eqs ( 14 ) or ( 15 ) . Eq ( 79 ) yields the expected population rate at time t for a given density of ages . In the Fokker-Planck formalism , this would correspond to the calculation of the rate from the membrane potential density as the probability flux across the threshold . Noting that ∂ t S ( t | t ^ ) A N ( t ^ ) = ( ∂ t + ∂ τ ) q ( t , τ ) , we find from Eq ( 67 ) the following first-order partial differential equation for the density of ages q ( t , τ ) : ( ∂ t + ∂ τ ) q = - λ ( t , τ ) q , q ( t , 0 ) = A N ( t ) . ( 80 ) Similarly , u and v obey from Eqs ( 77 ) and ( 76 ) , respectively , ( ∂ t + ∂ τ ) u= - u - μ τ m + ∑ β = 1 M w α β p α β N β ( ϵ α β * A N β ) ( t ) , ( 81 ) ( ∂ t + ∂ τ ) v= - λ ( t , τ ) [ 2 v - q ] ( 82 ) with boundary conditions u ( t , 0 ) = ur and v ( t , 0 ) = 0 . These functions , together with the threshold dynamics ϑ ( t , τ ) = u th + θ ( τ ) + ∫ τ ∞ θ ˜ ( τ ′ ) A N ( t - τ ′ ) d τ ′ , ( 83 ) determine λ ( t , τ ) and Λ ( t ) via Eq ( 74 ) , i . e . λ ( t , τ ) = c exp u ( t , τ ) - ϑ ( t , τ ) Δ u , Λ ( t ) = ∫ 0 ∞ λ ( t , τ ) v ( t , τ ) d τ ∫ 0 ∞ v ( t , τ ) d τ . ( 84 ) The Eqs ( 75 ) – ( 77 ) of the previous section can be regarded as the characteristic equations corresponding to the partial differential Eqs ( 80 ) – ( 82 ) ( “method of characteristics” ) . To simulate the population activity forward in time , the integrals in Eq ( 65 ) over the past need to be evaluated , starting at t ^ = - ∞ . For biological systems , however , it is sufficient to limit the integrals to a finite history of length T . This history corresponds to the range t - T ≤ t ^ < t , where we have to explicitly account for the dependence of the conditional firing rate λ ( t | t ^ ) on the last spike time t ^ . We will call neurons with last spike time in this range “refractory” because they still experience some degree of ( relative ) refractoriness caused by the last spike . The remaining part of the integral corresponding to the range - ∞ < t ^ < t - T receives a separate , compact evaluation . We will refer to neurons with their last spike time in this range as “free” because their conditional intensity is free of the influence of the last spike . How should we choose the length of the explicit history T ? First of all , this length can be different for different populations and is mainly determined by the time scale of refractoriness , i . e . the time it needs to forget the individual effect of a single spike in the past . Furthermore , it depends on the properties of the spike-triggered kernel , i . e . the dynamic threshold that is responsible for adaptation . More precisely , we determine the length of the history by the following conditions: first , the conditional intensity is insensitive to the precise timing of the last spike at t ^ < t - T if T ≫ max [ t ref , τ rel ] . ( 85 ) Here , tref is the absolute refractory period and τrel is the time scale of the relative refractory period . For the GIF model τrel = τm . Second , we demand that T is chosen such that for t > T , the quasi-renewal kernel θ ˜ ( t ) = Δ u [ 1 - e - θ ( t ) / Δ u ] can be well approximated by the original spike-triggered kernel θ ( t ) . Taylor expansion of the exponential yields the condition θ ( t ) ≪ Δ u , ∀ t > T . ( 86 ) The length of the history T needs to be chosen such that both conditions , Eqs ( 85 ) and ( 86 ) are fulfilled . It is important to note that condition Eq ( 86 ) does not require the time window T to be larger than the largest time scale of the spike-triggered kernel . For instance , consider the kernel θ ( t ) = J θ τ θ e - t / τ θ , where Jθ and τθ are adaptation strength and time scale , respectively . In particular , the adaptation strength Jθ sets the reduction in firing rate compared to a non-adapting neuron in the limit of strong drive irrespective of the time scale τθ ( see e . g . [80 , 82] ) . Condition Eq ( 86 ) can be fulfilled for a given T if either τθ is small enough such that the exponential e - t / τ θ is small , or , for a fixed adaptation strength Jθ , by increasing the adaptation time scale τθ such that Jθ/τθ ≪ Δu . We characterize the fluctuations of the stationary population activity by the power spectrum defined as C ˜ ( f ) = lim T → ∞ ⟨ | A ˜ ( f ; T ) | 2 ⟩ T , ( 132 ) where A ˜ ( f ; T ) = ∫ 0 T A N ( t ) e 2 π i f t d t ( 133 ) is the Fourier transform of the population activity on a time window of length T . For a population of renewal neurons the power spectrum is known analytically . It is given by [134] C ˜ ( f ) = r N 1 - | P ˜ ISI ( f ) | 2 | 1 - P ˜ ISI ( f ) | 2 , ( 134 ) where P ˜ ISI ( f ) is the Fourier transform of the interspike interval density P ISI ( t ) = λ ( t | 0 ) exp - ∫ 0 t λ ( s | 0 ) d s ( 135 ) and r is the stationary firing rate given by r = ∫ 0 ∞ exp - ∫ 0 t λ ( s | 0 ) d s d t - 1 . ( 136 ) Note that the power of the fluctuations in Eq ( 134 ) scales like 1/N , vanishing in the macroscopic limit N → ∞ . For the LIF model with escape noise , the hazard rate λ ( t|0 ) is given by λ ( t | 0 ) = f u ( t , 0 ) - u th , u ( t , 0 ) = μ + ( u r - μ ) exp - t - t ref τ m ( 137 ) for t > tref and λ ( t|0 ) = 0 for t ≤ tref . To model the cortical column of [5] in our framework , we used the parameters of the original publication and modified the model in two ways: First , the background Poisson input was replaced by a constant drive and an increased escape noise such that the populations exhibited roughly the same stationary firing rates . Specifically , we set ur = 0 mV , uth = 15 mV andΔu = 5 mV; and , using the mesoscopic dynamics , fitted the resting potentials of the GIF model ( here denoted by μ ^ α ) without adaptation Jθ = 0 to obtain firing rates r ^ that roughly match the target firing rates . Second , we introduced adaptation on excitatory cells with strength Jθ and time scale τθ , and re-adjusted the resting potential as follows u rest = μ ^ + J θ r ^ . This yielded correct stationary firing rates in the presence of adaptation . The resulting parameters of the modified model are summarized in Table 2 . | Understanding the brain requires mathematical models on different spatial scales . On the “microscopic” level of nerve cells , neural spike trains can be well predicted by phenomenological spiking neuron models . On a coarse scale , neural activity can be modeled by phenomenological equations that summarize the total activity of many thousands of neurons . Such population models are widely used to model neuroimaging data such as EEG , MEG or fMRI data . However , it is largely unknown how large-scale models are connected to an underlying microscale model . Linking the scales is vital for a correct description of rapid changes and fluctuations of the population activity , and is crucial for multiscale brain models . The challenge is to treat realistic spiking dynamics as well as fluctuations arising from the finite number of neurons . We obtained such a link by deriving stochastic population equations on the mesoscopic scale of 100–1000 neurons from an underlying microscopic model . These equations can be efficiently integrated and reproduce results of a microscopic simulation while achieving a high speed-up factor . We expect that our novel population theory on the mesoscopic scale will be instrumental for understanding experimental data on information processing in the brain , and ultimately link microscopic and macroscopic activity patterns . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"action",
"potentials",
"medicine",
"and",
"health",
"sciences",
"neural",
"networks",
"population",
"dynamics",
"membrane",
"potential",
"condensed",
"matter",
"physics",
"computational",
"biology",
"electrophysiology",
"neuroscience",
"computational",
"neuroscience",
"pop... | 2017 | Towards a theory of cortical columns: From spiking neurons to interacting neural populations of finite size |
Dengue is the most important mosquito-borne viral disease affecting humans . The only prevention measure currently available is the control of its vectors , primarily Aedes aegypti . Recent advances in genetic engineering have opened the possibility for a new range of control strategies based on genetically modified mosquitoes . Assessing the potential efficacy of genetic ( and conventional ) strategies requires the availability of modeling tools that accurately describe the dynamics and genetics of Ae . aegypti populations . We describe in this paper a new modeling tool of Ae . aegypti population dynamics and genetics named Skeeter Buster . This model operates at the scale of individual water-filled containers for immature stages and individual properties ( houses ) for adults . The biology of cohorts of mosquitoes is modeled based on the algorithms used in the non-spatial Container Inhabiting Mosquitoes Simulation Model ( CIMSiM ) . Additional features incorporated into Skeeter Buster include stochasticity , spatial structure and detailed population genetics . We observe that the stochastic modeling of individual containers in Skeeter Buster is associated with a strongly reduced temporal variation in stage-specific population densities . We show that heterogeneity in container composition of individual properties has a major impact on spatial heterogeneity in population density between properties . We detail how adult dispersal reduces this spatial heterogeneity . Finally , we present the predicted genetic structure of the population by calculating FST values and isolation by distance patterns , and examine the effects of adult dispersal and container movement between properties . We demonstrate that the incorporated stochasticity and level of spatial detail have major impacts on the simulated population dynamics , which could potentially impact predictions in terms of control measures . The capacity to describe population genetics confers the ability to model the outcome of genetic control methods . Skeeter Buster is therefore an important tool to model Ae . aegypti populations and the outcome of vector control measures .
Mosquito-borne dengue virus serotypes cause approximately 50 million cases of dengue fever per year , 500 , 000 cases of dengue hemorrhagic fever ( DHF ) or dengue shock syndrome ( DSS ) , and result in approximately 12 , 500 fatalities annually [1] , [2] . Since the 1950s , the incidence of DHF/DSS has increased over 500-fold [2] , due to increases in human population , uncontrolled urbanization and international travel [3] . The major vector for dengue is the mosquito Aedes aegypti which thrives in households with open , water-filled containers in which larvae develop . Lack of reliable piped water service and garbage disposal systems in endemic subtropical and tropical countries provide mosquito vectors with ample development sites [4] . Presently , there is no commercially available clinical cure for dengue and no vaccine has successfully completed clinical trials [5] , leaving vector control as the only viable option for dengue prevention . Several practices are used to control dengue vector populations , including reduction or elimination of larval development sites and insecticides targeting immatures or adults . In the case of Ae . aegypti , the Container Inhabiting Mosquito Simulation Model ( CIMSiM ) [6] , [7] is the most detailed tool available for understanding population dynamics and the expected effects of different intervention strategies on adult female densities . CIMSiM is a weather-driven , dynamic life table simulation model of Ae . aegypti populations that incorporates a high level of detail about the life history of this species . Results from CIMSiM are used as the entomological input of a companion model , DENSiM [8] , that models dengue transmission dynamics based on the mosquito population dynamics simulated by CIMSiM . CIMSiM and DENSiM have proven useful in characterizing local Aedes aegypti population dynamics [9] and predicting general impacts of control measures on dengue prevalence and incidence [10] , [11] . Despite its considerable detail , three things that CIMSiM does not take into account are spatial heterogeneity in habitat availability , potential impacts of stochastic effects – both of which could significantly affect population dynamics – and the genetics of the simulated population . A stochastic spatial model of Ae . aegypti population dynamics has also been developed separately [12] , [13] that does not include any genetic component . A lack of a genetic modelling is not critical when dealing with most conventional methods of vector control unless evolution of insecticide resistance is of concern . Recent advances in molecular biology and genetic engineering , however , have presented the possibility of employing a number of control methods based on genetically engineered mosquitoes [14] . Genetic strategies fall into two broad categories: population suppression and population replacement . Population suppression methods , such as the Release of Insects carrying a Dominant Lethal ( RIDL ) , which is a form of the sterile insect technique [15] , [16] , aim to reduce the density of vectors by releasing genetically engineered male mosquitoes that mate with native females and cause mortality of offspring before they emerge as adults . Population replacement strategies aim to replace the resident , competent vector population with mosquitoes that are genetically engineered to not transmit a pathogen [17]–[19] . For both approaches a model that can predict the outcome of releasing an engineered strain in a given location and across different ecological and epidemiological circumstances is critically needed to provide guidance for which particular approach ( or combination thereof ) would be the most effective and to anticipate any undesirable outcome . To address this need , we developed a modeling tool , Skeeter Buster , that can predict how Ae . aegypti population dynamics and population genetics might be affected by stochasticity and spatial variation in Ae . aegypti habitat . Skeeter Buster builds on the biologically rich components of CIMSiM , while adding stochasticity , explicit spatial structure and genetics . The construction of this model is the first step of our project that aims at evaluating conventional and genetic vector management tools and their potential success in controlling dengue incidence in human populations . To that end , Skeeter Buster will ultimately be associated with an epidemiological model . This tandem modeling tool will be comparable to the CIMSiM/DENSiM association , and will allow a direct assessment of the effects of vector control measures ( including genetic approaches ) on dengue prevalence and incidence . The latest version of Skeeter Buster with a user-friendly interface is available for Windows platforms at http://www . skeeterbuster . net , and the source code is available on request from the authors . In this first paper , we explain the characteristics and specificities of Skeeter Buster , and present results from simulations that compare the population dynamics predictions of Skeeter Buster to those of CIMSiM . Examples presented are not intended to explore the vast parameter space that is associated with this model . All parameters are , however , adjustable in the user-friendly version of the modeling tool . Subsequent articles will describe a detailed sensitivity and uncertainty analysis of this model , as well as validation against a data set of Ae . aegypti population dynamics in Iquitos , Peru . These detailed analyses will indicate whether there are specific details in the model that are not important for predicting the dynamics of Ae . aegypti , and could be dropped from the model . These analyses may also point to specific parameters in the model that have major effects on the mosquito dynamics , and therefore require better empirical estimates .
Because Skeeter Buster was built using many algorithms from CIMSiM , the two models share a number of identical characteristics . We schematically describe the relationships between the two models , with identical components represented in grayscale and specific additions in Skeeter Buster in color ( Figure 1 ) . CIMSiM is a deterministic , weather-driven model that follows cohorts on a daily time scale for each immature stage ( eggs , larvae , pupae ) as well as female adults . Because they are not considered to impact the population dynamics , adult males are not modeled in CIMSiM . Environmental parameters include daily weather data ( temperature , precipitation , and relative humidity ) and an external input of food into containers . Based on these variables , CIMSiM calculates the number of individuals within all cohorts present in the model at a given time , their cumulative physiological development , weight , fecundity and gonotrophic status , as well as the transitions between life stages . With the exception of food input , all calculations in CIMSiM are applied to cohorts in a strictly deterministic fashion . CIMSiM calculates water temperature and water level in all containers based on the available local weather data . The amount of food available in each container is also calculated daily . In addition to food depletion by consumption , three factors affect the amount of food: the external food input , a daily decay factor and the conversion of dead immatures to nutritional resources . The daily survival probability for each life stage includes a temperature-dependent component . Cumulative physiological development of each life stage is also based on temperature using an enzyme kinetics approach [20] assuming that a single enzyme determines the development rate of the insect ( see equations in Text S2 . 2 ) . Completion of physiological development at a given stage is attained when cumulative development reaches a threshold value ( specific for each life stage ) . Hatch of embryonated eggs is determined by water level and water temperature in the container . Larval weight is modeled in parallel with the amount of food in each container according to the equations in [21] ( see in Text S2 . 4 ) . Pupation requires larvae to complete physiological development as well as reach a sufficient weight . Fecundity of female adults is based on their weight , and females distribute their eggs among available containers based on the size of these containers . These general characteristics of CIMSiM are all incorporated into Skeeter Buster , but with three major differences . First , Skeeter Buster is a stochastic model . For a given event ( e . g . survival ) applied to a specific cohort , a probability is defined for the cohort , and that same probability is applied independently to all individuals within the cohort . The number of individuals to which the event occurs is obtained by drawing a number from a binomial distribution defined by that probability and the total number of individuals in the cohort . Second , Skeeter Buster models several distinct locations ( hereafter called “properties” ) . In the simplest setup , properties are arranged on a rectangular grid , and sets of distinct water-holding containers are assigned to individual properties ( indoor or outdoor location of each container is specified ) . Immature cohorts are associated with a specific container within a property , and emerging adults are associated with a specific property . Finally , because Skeeter Buster also models the genetics of the population , cohorts are further distinguished by genotype . Skeeter Buster also includes a number of components lacking in CIMSiM ( see Figure 1 ) . First , because of the genetic component of Skeeter Buster , male adults are now included in the model . Consequently , an important new component is the modeling of mating in the population . Mating is restricted to individuals present at the same property . Adults can disperse from one property to another , and containers can also be transported between properties , with the assumption that egg cohorts are carried along in the container . In the following sections , we describe the Skeeter Buster model in more detail . We first describe the dynamics within a single property and within individual containers , and then describe the spatial structure of the model and mosquito movement among properties . We provide a complete description of the processes involved in Skeeter Buster . Some of these processes are similar to those in CIMSiM and are described in [6] . Therefore , we only describe those processes briefly in the main text , and refer the reader to supporting material ( Text S2 , Text S4 , Dataset S1 ) for more details about the equations and parameters that are identical to their equivalent in CIMSiM . While CIMSiM models a single representative area with a default size of 1 ha , Skeeter Buster models multiple properties independently . Each property hosts a specific set of containers both inside and outside of buildings , and the immature cohorts in these containers as well as the adults emerging from those are specifically assigned to that particular property . Properties are laid out on a rectangular grid , each cell of the grid representing a single property . The grid is not associated with explicit geographic distances , and the property is the only fundamental unit of distance . Although in this paper we consider properties to be at the scale of meters ( individual houses in a dense urban setting ) , one property in the model can be considered to be larger units such as a block of properties or a village ( and parameters can be adjusted accordingly ) if needed for specific questions . Properties located on the edges of the grid have fewer immediate neighbors than those in the interior of the grid . To deal with these edge locations we employ one of three boundary assumptions . First , periodic boundaries assume that opposite borders of the grid are connected to each other to form a toric topology . Second , solid boundaries prevent mosquitoes from migrating across the border , and force them to stay in the border property . Third , with random boundaries , mosquitoes migrating across a border are reintroduced at a random location on the edges of the grid . Properties can be identified by their coordinates ( xi , yi ) on the grid . Distance between properties ( xi , yi ) and ( xj , yj ) is defined as |xj−xi|+|yj−yi| ( with appropriate adjustments depending on the boundary conditions ) . In this paper , we only report results from model runs that use solid boundary conditions . Adults can disperse between properties . Skeeter Buster allows for both short and long range dispersal . Short range dispersal allows adult male and female mosquitoes to move to nearest neighbor properties . We assume that this movement occurs with probability 0 . 3 for each mosquito on each day . We estimated this probability by simulating an empirical mark-release-recapture study in Thailand [45] , and measuring the necessary daily dispersal probability to match the distribution of captured marked mosquitoes found in that study ( Figure 5 ) . In the model , for each dispersing adult , one of four directions is randomly chosen , and the adult is moved to the nearest property in that direction ( von Neumann neighborhood ) . Adult mosquitoes can disperse to properties at a further distance in the grid by long range dispersal . There is no clear consensus in the literature about the extent of long range dispersal of adult Ae . aegypti ( e . g . how often this happens , or how far adults migrate ) [45]–[47] . In Skeeter Buster , each adult can disperse long distances with a daily probability; we assume a default value of 0 . 02 . A maximum distance is also defined for long range dispersal events ( default value of 20 properties , corresponding to ∼200 meters in a dense urban setting ) . Within this range , an actual distance is chosen at random ( uniformly between 1 property and the maximum distance ) , and the destination property is chosen randomly among properties situated at this particular distance . We assume that the dispersal probabilities for both short and long range dispersal are independent of age , sex [45] , parous state , mating status , size or developmental percentage . Finally , we also allow the possibility for displacement of containers from one property to another . With some daily probability , any particular container is removed from its original property and allocated to another randomly chosen property in the grid . To account for the movement of immature cohorts associated with container displacement , all egg cohorts present in a moving container remain unaltered by this process . Larval and pupal cohorts , however , are discarded . In this paper , unless otherwise specified , the daily movement probability is assumed to equal zero . To develop the Skeeter Buster simulation program , we chose to rewrite a clone of CIMSiM in C++ as a first basis , because it provides several clear advantages for model development . From a practical point of view , CIMSiM was originally written in Visual Basic , a coding language that is tied to the PC platform and that has undergone alterations that hinder recompilation of the code on recent machines . We instead chose to use standards-compliant C++ to provide maximum flexibility , e . g . in allowing the code to be ported to and run on other computer systems , and to prevent future obsolescence of the code . Another , more important , motivation for our strategy was to provide some means to verify our simulation code , ensuring that all procedures would work in Skeeter Buster according to the algorithms presented in the original published model [6] . The complexity of the CIMSiM ( or Skeeter Buster ) simulation code offers many opportunities for the occurrence of coding errors; these could be difficult to identify without an independent rewrite of the code . Rewriting CIMSiM allowed us to reveal and correct some inconsistencies between the original model and presented algorithms , as well as apparent malfunctions in the original release of CIMSiM ( see Text S1 , Figure S1 and Figure S2 ) . For all the above reasons , we felt that the rewriting process of CIMSiM was a necessary step prior to working with confidence when expanding the initial model to build Skeeter Buster . We rewrote CIMSiM in C++ ( hereafter refered to as C++ CIMSiM ) by exactly following the algorithms described in [6] . We tested C++ CIMSiM by systematically comparing its output to the output of the original CIMSiM program with identical parameters . Whenever the output was different , we contrasted the source code of the C++ CIMSiM to the algorithms published in [6] as well as to the source code of the original CIMSiM . We corrected several coding errors in C++ CIMSiM ( see Text S1 ) . We observed several differences between the operation of the original CIMSiM program and the algorithms described in [6] . In order to verify C++ CIMSiM , we had to deliberately include these differences and coding errors from the original source code into the C++ CIMSiM source code during this testing phase . We attributed rare remaining differences in the outputs to malfunctions of the original CIMSiM executable . We were able to mimic such malfunctions by deliberately altering specific cohorts of larvae on specific occasions in C++ CIMSiM ( see Text S1 ) . Finally , we were able to match the output of the original CIMSiM executable and the output of C++ CIMSiM ( Figure 6 ) . Because C++ CIMSiM is not affected by the malfunctions in the original CIMSiM executable and is more flexible in terms of desired output , we used C++ CIMSiM in our comparisons to Skeeter Buster . Skeeter Buster was developed by expanding and modifying this C++ code according to the model specificities described above ( see Text S4 for a detailed list of modifications ) . A user-friendly graphical interface was developed for PC/Windows systems , and allows the user to vary parameter values . This part of the code is more specific to the particular system , but a similar interface could be developed for other systems ( or could be developed in a portable framework such as Java ) .
CIMSiM and Skeeter Buster handle multiple containers of the same type in different ways . While CIMSiM models a single representative container , and multiplies the results according to the density of such containers per hectare , Skeeter Buster models each container individually . In order to compare these two approaches , we first set both CIMSiM and Skeeter Buster to have the equivalent of 100 containers of each of the above three types in an area of one hectare , with completely random mating of the mosquitoes within this area . For Skeeter Buster , this was equivalent to modeling a single “property” with a 1 ha yard in which 100 containers of each type are placed . We compare the outcome of this simulation to that of CIMSiM set up with the same three types of containers , each with a density of 100/ha . Both approaches model a similar 1-ha area . The primary difference is that Skeeter Buster models the dynamics in each of the 300 containers individually , whereas CIMSiM simulates the dynamics in groups of only 3 representative containers . We compared the stage-specific densities of eggs , larvae , pupae and adult females within the 1-ha area from single runs of both Skeeter Buster and CIMSiM ( Figure 7 ) . For all developmental stages of Ae . aegypti , a common characteristic of the output from Skeeter Buster is that the temporal variation in density is reduced compared to CIMSiM . Although it may appear paradoxical to observe less variation in a stochastic model , this result can be explained by two major differences between these two models . First , because of the stochasticity incorporated in Skeeter Buster , the demographic dynamics in each container are independent and not synchronized , which reduces the variability when the total density across all 300 containers is considered . Second , in Skeeter Buster , individuals within a given larval cohort do not necessarily all pupate on the same day , and pupation can be spread across several days . The same effect applies for larvae maturation and pupae maturation . As a result , the ‘cohort effect’ is quickly lost in the simulation , reducing the temporal variation in densities . The average stage-specific densities , taken over the entire year , in Skeeter Buster are similar to those obtained from CIMSiM . Minor differences in average densities can be explained by the different daily mortality rates used in Skeeter Buster , or by minor changes in the oviposition procedures ( see Text S4 ) . These changes also affect the periodicity of these time series , with the interval between peaks of female adult densities appearing to be slightly shorter in Skeeter Buster ( see Figure 7D , and Text S3 , Figure S5 and Figure S6 for a more detailed analysis of time series periodicity ) . We incorporate spatial structure in Skeeter Buster by considering simulations using the same 300 containers ( 100 of each type ) as before , but now distributed among 100 individual properties . Properties are laid out on a 10×10 grid , and migration between individual properties can occur ( see Methods ) . To explore the impact of habitat heterogeneity , we consider two container distributions . First , a homogeneous container distribution in which each property has exactly 3 containers , 1 of each type; in other words , all properties have an identical container distribution . Second , a heterogeneous container distribution , in which all 300 containers are randomly assigned to one of the 100 properties . In this case , the overall number of containers remains the same as in the homogenous case , but individual properties can have different types and numbers of containers . We present a snapshot of the spatial variation in the density of the population , as the number of pupae per property , at the end of a 1-yr simulation with the homogeneous container distribution , on Figure 8 . Because of the effects of both stochasticity in local dynamics and dispersal , there is clear spatial heterogeneity among population densities between individual properties , even when their container composition is the same . We compare the time series of female adult density in the whole population for both types of habitat heterogeneity described above , as well as for the non-spatial case described in the previous section ( Figure 9 ) . Both average densities and temporal variances are comparable in all three cases , and therefore do not appear to be affected by habitat heterogeneity . Habitat heterogeneity however has a strong effect on the level of spatial variation ( between properties ) in the population . We quantify this variation by measuring the coefficient of variation in the number of pupae among individual properties at a given time ( denoted as CVp ) . We measured CVp ( Figure 10 ) in the two above-defined setups ( homogeneous or heterogeneous ) , and under three different assumptions about adult dispersal between properties : ( 1 ) both short range and long range dispersal are allowed , with daily probabilities of 0 . 3 and 0 . 02 , respectively; ( 2 ) only short range dispersal is allowed , or: ( 3 ) no dispersal at all . The results of analysis of variance for CVp are also summarized ( Table 1 ) . These results show a clear effect of the spatial distribution of containers on CVp . As expected , the values of CVp are significantly higher when the container distribution is heterogeneous . Dispersal also has a significant effect . For both container distributions , CVp is significantly higher when no dispersal occurs . On the other hand , the values of CVp when short and long range dispersal occur do not differ from the case when only short range dispersal is allowed , suggesting that long range dispersal does not affect spatial variance among properties within the specified level of heterogeneity . Similarly , there is a significant effect of the interaction between container distribution and dispersal pattern . The effects of dispersal on CVp are more pronounced when the container setup is heterogeneous . Finally , we describe how the genetic structure of the population is affected by spatial factors such as the distribution of containers ( homogeneous or heterogeneous ) and adult dispersal . We follow the dynamics of a single locus with two alleles that do not differentially impact fitness ( i . e . two neutral alleles ) . Both alleles are initially introduced into the population in egg cohorts homozygous for one of the two alleles , each at a frequency of 0 . 5 . Simulations are set up with 400 properties ( 20×20 grid ) , with the same three container types as above , and run for 5 years . We arbitrarily define 25 subpopulations that consist of non-overlapping 4×4 squares within the 20×20 grid . ( Here , we use 400 properties instead of 100 to allow us to partition our grid into a larger number of subpopulations , facilitating the spatial analysis that follows . ) Short range dispersal is set to its default value ( 0 . 3 daily dispersal probability ) , and we examine the effects of varied amounts of long range adult and container movement on the genetic structure of the population . We calculated the global FST values based on this neutral locus at the end of the simulations ( Figure 11 ) . FST values , representing the level of genetic differentiation within the overall population ( between subpopulations ) , are higher in the case of a heterogeneous distribution of containers , but decrease quickly when the daily probability of long range dispersal increases . We also calculate pairwise FST values between all 25 subpopulations . We can test the existence of isolation by distance in our simulated population by examining the correlation between the genetic distance between two subpopulations ( given by the pairwise FST value ) and their geographic distance . More specifically , following the method described in [52] , we regress the values of FST/ ( 1−FST ) for pairs of subpopulations against the logarithms of their geographic distances . Isolation by distance is characterized by a significant correlation between these two distances . A stronger isolation by distance is associated with a higher slope of the regression line . We measured the values of this slope for different assumptions concerning habitat heterogeneity and adult movement ( Figure 12 ) . For both types of container distribution , long range dispersal , even at daily probabilities as low as 0 . 02 , prevents the occurrence of isolation by distance at the scale of the simulation considered here ( 20×20 properties ) . Finally , we also examine the impact of container displacement ( and the associated movement of egg cohorts ) between properties . We measured the impact of this movement on final FST values for a neutral allele , assuming that there is no long range dispersal ( Figure 13 ) . Only the plastic buckets ( 1-gallon and 5-gallon ) are moved since larger containers are not typically moved among households . It appears that moving containers across the city can have an impact on the population structure even when these events are rare , although increasing this probability does not seem to impact FST values as much as adult dispersal .
The results from Skeeter Buster presented in this paper using simplified container and property setups highlight the impact of spatial structure and heterogeneity on the population dynamics of Ae . aegypti . First , the simulated population dynamics differ markedly between CIMSiM and Skeeter Buster when a large number of identical containers within one property are considered in Skeeter Buster . Because each of these containers is simulated individually in the stochastic Skeeter Buster , the overall population dynamics is an average over a large number of containers whose individual dynamics are typically not synchronized . Additionally , containers in different properties are associated with a different local population . Identical containers in Skeeter Buster can therefore exhibit very different dynamics from one another . As a consequence , the variability in densities of Ae . aegypti at the level of the population is greatly reduced ( see Fig . 7 ) . Beyond the effect of simulating individual containers , the explicit simulation of individual properties in Skeeter Buster does not seem to affect the overall population dynamics , at least in the settings investigated here ( see Fig . 9 ) . However , this inclusion of multiple properties allows a quantitative description of spatial heterogeneity among properties in terms of Ae . aegypti densities and age composition that could not be modeled by CIMSiM . We show here that the level of heterogeneity among properties in Ae . aegypti population density can be high even when a homogeneous container distribution is considered . Future studies based on Skeeter Buster will reveal if and how much this heterogeneity is predicted to affect both dengue transmission dynamics and the impact of vector control strategies . Because there is evidence that heterogeneity among properties in densities of female adults could be important for both [49] , [53] , we conclude that it is an important feature to include in our modeling tool . Among the possible strategies for decreasing dengue incidence , approaches using genetic tools to control the mosquito population appear to be promising , but their applicability in field situations is still under evaluation . Skeeter Buster was designed to aid this evaluation , and simulate the efficiency and practicality of these approaches in order to guide the development of genetic control programs . We therefore incorporated explicit genetics in the model , and describe here the basic population genetic structure predicted by this model . While long range dispersal does not seem to affect the spatial variance in densities , Figs . 11 and 12 show that long range dispersal can significantly affect the genetic structure of the population . Even relatively rare long range dispersal events ( daily probability lower than 2% ) are associated with lower FST values in the population and dramatically reduce the observed instances of isolation by distance among subdivisions of our modeled population . The transfer of containers between properties in the grid can also impact the genetic structure , although its impact does not seem to be as important as that of adult long range dispersal ( Figure 13 ) . The existence of strong genetic spatial structure in the population is important to the potential fate of an allele introduced into specific locations within a population . Strong genetic structure could impede or slow the spread of a novel allele to distant parts of the population . For this reason , the ability of Skeeter Buster to model this genetic structure is an important addition for predicting the outcome of genetic control strategies in Ae . aegypti populations . The spatial scale examined in Skeeter Buster is at the level of individual properties , that is , in the case of Iquitos , distances of an order of magnitude of hundreds of meters . Field studies of genetic structure at this level are rare . FST values reported from small-scale clusters ( kilometers ) in within-city studies [54]–[57] are variable but consistent with the highest values observed in the simulations presented in this paper . This would suggest a limited amount of adult dispersal between these geographically close sites , without excluding the possibility of gene flow maintained by displacement of immatures through human activities and transportation . More generally , these results emphasize the need to characterize the dispersal patterns of Ae . aegypti in natural populations . While adults are generally considered to migrate only short distances ( modeled by our short range dispersal ) [45] , dispersal to longer distances has been observed [46] , but how often such long range dispersal events occur is unknown . Overall , the results presented here are consistent with our assertion that Skeeter Buster provides a realistic description of Ae . aegypti population dynamics and will be a valuable tool in the development of city-wide genetic strategies for prevention of dengue and control of its major mosquito vector . Ultimately , this entomological simulation will be a component of a framework from which dengue transmission can be modeled , and control measures can be evaluated . However , two important requirements have to be fulfilled before these further steps can be carried out . First , the outcome of the model must be validated with population data from an actual field site: this will rely on a more elaborate property setup and container distribution than the examples presented here . Skeeter Buster allows for detail at the individual container level , and therefore enables a specific Ae . aegypti population in a particular location to be modeled . But , to achieve such a location-specific level of accuracy , Skeeter Buster requires intensive field work to obtain a description of the container distribution and relative productivity in this particular location . In a subsequent paper , we will illustrate this location-specific simulation capacity with a case study of the city of Iquitos , Peru . Second , since this model relies on a very high number of procedures and parameters , all of which are associated with some level of uncertainty , it is crucial to carry a broad-scale uncertainty and sensitivity analysis of this model . This analysis will ensure that the results of the simulations are robust enough within the range of the existing uncertainties on parameter values , or , if not , the analysis will highlight the traits predicted to account for the highest percentage of uncertainty in predicted population dynamics and genetics , providing guidelines for the most needed additional field or laboratory studies . | Dengue is a viral disease that affects approximately 50 million people annually , and is estimated to result in 12 , 500 fatalities . Dengue viruses are vectored by mosquitoes , predominantly by the species Aedes aegypti . Because there is currently no vaccine or specific treatment , the only available strategy to reduce dengue transmission is to control the populations of these mosquitoes . This can be achieved by traditional approaches such as insecticides , or by recently developed genetic methods that propose the release of mosquitoes genetically engineered to be unable to transmit dengue viruses . The expected outcome of different control strategies can be compared by simulating the population dynamics and genetics of mosquitoes at a given location . Development of optimal control strategies can then be guided by the modeling approach . To that end , we introduce a new modeling tool called Skeeter Buster . This model describes the dynamics and the genetics of Ae . aegypti populations at a very fine scale , simulating the contents of individual houses , and even the individual water-holding containers in which mosquito larvae reside . Skeeter Buster can be used to compare the predicted outcomes of multiple control strategies , traditional or genetic , making it an important tool in the fight against dengue . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"ecology/population",
"ecology",
"infectious",
"diseases/epidemiology",
"and",
"control",
"of",
"infectious",
"diseases",
"genetics",
"and",
"genomics/population",
"genetics"
] | 2009 | Skeeter Buster: A Stochastic, Spatially Explicit Modeling Tool for Studying Aedes aegypti Population Replacement and Population Suppression Strategies |
The innate-immune restriction factor MxA inhibits influenza replication by targeting the viral nucleoprotein ( NP ) . Human influenza virus is more resistant than avian influenza virus to inhibition by human MxA , and prior work has compared human and avian viral strains to identify amino-acid differences in NP that affect sensitivity to MxA . However , this strategy is limited to identifying sites in NP where mutations that affect MxA sensitivity have fixed during the small number of documented zoonotic transmissions of influenza to humans . Here we use an unbiased deep mutational scanning approach to quantify how all single amino-acid mutations to NP affect MxA sensitivity in the context of replication-competent virus . We both identify new sites in NP where mutations affect MxA resistance and re-identify mutations known to have increased MxA resistance during historical adaptations of influenza to humans . Most of the sites where mutations have the greatest effect are almost completely conserved across all influenza A viruses , and the amino acids at these sites confer relatively high resistance to MxA . These sites cluster in regions of NP that appear to be important for its recognition by MxA . Overall , our work systematically identifies the sites in influenza nucleoprotein where mutations affect sensitivity to MxA . We also demonstrate a powerful new strategy for identifying regions of viral proteins that affect inhibition by host factors .
Influenza proteins must evade immunity while maintaining their ability to function and interact with host cell factors [1] . The effects of immune pressure on influenza evolution have been most studied in the context of adaptive immunity , with numerous studies showing how antibodies and T-cells drive the fixation of immune-escape mutations in viral proteins [2–5] . However , the innate immune system also exerts selection on influenza virus via the interferon-stimulated expression of restriction factors , some of which target viral proteins and inhibit their function . The first anti-influenza restriction factor to be discovered , the murine protein Mx1 , was initially described over 50 years ago [6–8] . It is now known that Mx1 and its human ortholog MxA inhibit influenza virus by interacting with the viral nucleoprotein ( NP ) [9–14] . However , the exact mechanistic details of the inhibitory interaction between MxA and NP remain incompletely understood . Influenza virus counteracts the inhibitory effects of MxA through two distinct strategies: it generally blocks the interferon-response that drives expression of MxA [15] , and it fixes specific amino-acid mutations in NP that reduce its sensitivity to MxA [16 , 17] . The importance of the second of these two strategies has been elegantly demonstrated by studies comparing the MxA sensitivity of different viral strains . NPs from avian and swine influenza viruses are more sensitive to human MxA than NPs from human influenza [16 , 17] . NPs from non-human influenza virus have been introduced into circulating human influenza strains twice over the last century: once in 2009 from swine influenza [18] , and once in or before 1918 probably from avian influenza [19–21] . By functionally characterizing the effects of mutations at sites that differ between these human influenza NPs and their predecessors from non-human viral strains , Mänz et al [17] identified a small set of sites in NP where mutations affect MxA resistance . Riegger et al [22] subsequently identified another site in NP where a mutation has increased the MxA resistance of an avian H7N9 virus that has undergone numerous non-sustained zoonotic transmissions to humans . Characterizing sites that differ between non-human and human influenza strains is a powerful strategy to identify mutations that have historically contributed to the adaptation of NP to avoid recognition by MxA . However , it is an incomplete approach for mapping the full set of sites in NP that affect sensitivity to MxA . Evolution is stochastic [23–25] , meaning that any given adaptation event will sample only some of the possible mutations that confer MxA resistance . In addition , adaptation of non-human influenza to humans only favors mutations at NP sites that initially encode an amino acid that confers relatively high sensitivity to MxA . Sites at which avian and swine influenza viruses already possess MxA-resistant amino acids will not be identified by cross-species comparison . Therefore , more systematic approaches are needed to fully characterize the sites in NP that affect MxA resistance . Systematic measurement of how all amino-acid mutations affect a protein phenotype has recently become possible with the advent of deep mutational scanning [26 , 27] . This massively parallel experimental technique involves generating a library of mutants , imposing a functional selection , and using deep sequencing to determine the frequency of each mutation before and after selection . Deep mutational scanning has already been used to examine the functional effects of most mutations to several influenza proteins [28–35] . Here we use deep mutational scanning to quantify how every amino-acid mutation to the NP of a human influenza strain affects sensitivity to MxA . This unbiased approach enables us to identify mutations that both increase and decrease MxA sensitivity . Therefore , in addition to re-identifying some sites where mutations have previously been shown to adapt influenza to MxA , we are able to identify new sites that affect MxA sensitivity . We individually confirm the effects of mutations at 12 of the sites identified by our high-throughput experiments , thereby validating the accuracy of the deep mutational scanning approach . At most of the sites where mutations have the largest effect on MxA sensitivity , almost all known influenza A strains already possess an amino acid that confers high resistance . Overall , our work finds new sites affecting MxA resistance that could not have been identified by comparing viral strains across species , and introduces a framework for comprehensively profiling the effect of all mutations to viral proteins on recognition by restriction factors .
Our goal is to understand which sites in influenza NP determine its sensitivity to MxA . We can do this by experimentally quantifying how MxA sensitivity is affected by all amino-acid mutations to NP that support viral replication . Systematic measurements of this type can be made using the deep mutational scanning approach outlined in Fig 1 . This approach involves creating influenza viruses that carry a diverse set of NP mutations , growing these viruses in cells that do or do not express human MxA , and then using deep sequencing to identify mutations that are enriched or depleted in one condition versus the other . Mutations that are enriched in MxA-expressing cells relative to control cells increase MxA resistance , whereas mutations that are relatively depleted in MxA-expressing cells increase MxA sensitivity . We chose to perform our deep mutational scan on a NP from a human-adapted influenza strain , A/Aichi/2/1968 ( H3N2 ) . We reasoned that use of a human-adapted NP should allow us to better detect mutations that decrease MxA resistance , as well as identify any resistance-enhancing mutations that have not already fixed in human influenza virus . The use of a human-adapted NP makes our approach complementary to previous studies that have focused on mutations that increase the MxA resistance of non-human strains of influenza [17 , 22] . We have previously described duplicate libraries of influenza viruses that carry nearly all amino-acid mutations to the Aichi/1968 NP that are compatible with viral growth [32]; these libraries formed the starting point for the work performed here . Briefly , these virus libraries were generated by creating plasmid pools encoding all random codon mutants of the Aichi/1968 NP gene , using these plasmid pools to generate pools of mutant viruses , and then passaging the viruses in cell culture at a low multiplicity of infection . We mutagenized all 498 codons in the NP gene except for the N-terminal methionine . Each residue can be mutated to 19 non-wildtype amino acids , so our plasmid libraries sampled from 19 × 497 = 9 , 443 amino-acid mutations . Only mutations that support the growth of replication-competent viruses remain in the libraries after the passaging in cell culture . During viral infection of normal human cells , MxA expression is induced by activation of the interferon response , which varies from cell to cell [36 , 37] . But our controlled experiment ( Fig 1 ) requires cells that never or always express a functional human MxA . For our MxA-deficient cells , we chose MDCK-SIAT1 cells , a variant of the Madin Darby canine kidney cell line . We chose these cells for two reasons . First , the canine MxA ortholog lacks anti-influenza activity against all influenza strains tested to date [38 , 39] so the likelihood that this gene will exert selection on our virus library is small . Second , MDCK-SIAT1 cells support robust growth of influenza , and are therefore well-suited to maintaining the diversity of our virus library . To create MxA-expressing cells , we engineered a MDCK-SIAT1 cell line that constitutively expresses human MxA . We also created a control cell line that constitutively expresses the T103A mutant of MxA , which is inactive against influenza [40] . For both cell lines , we verified MxA protein expression ( Fig 2A ) . We also verified that constitutive expression of wildtype but not T103A MxA profoundly inhibits viral replication ( Fig 2B ) . This inhibition demonstrates that even human-adapted influenza NP is sensitive to sufficiently high levels of MxA . We then infected our virus libraries into all three cell lines as indicated in Fig 1A . In order to maintain the diversity of the libraries , we infected each cell line with 5 × 106 TCID50 of virus . We performed the infections at a multiplicity of infection ( MOI ) of 0 . 1 to reduce viral co-infection and subsequent genetic complementation . We then isolated viral RNA from infected cells after 48 hours , and deep sequenced NP to determine the frequency of each mutation in each selective condition . We used overlapping paired-end Illumina reads to reduce the sequencing error rate ( S1 Fig shows that this strategy reduced the net rates of errors associated with sequencing , PCR , and reverse transcription to below the frequency of the actual mutations of interest ) . We performed this experiment independently for each of our two NP virus libraries , meaning that all high-throughput measurements were made in true biological duplicate . Our expectation was that analyzing the deep sequencing data would enable us to identify mutations that affect MxA resistance as shown in Fig 1B . We estimated the effect of each mutation on MxA resistance by computing the logarithm of its frequency in the MxA-expressing cells relative to the non-expressing cells . We refer to this quantity as the mutation differential selection . We estimated the total differential selection at each site by summing the absolute values of the differential selection on each mutation at the site . Fig 1B graphically illustrates our measures of mutation and site differential selection . Fig 3A shows that our two independent replicates of deep mutational scanning yielded reproducible estimates of the differential selection at each site . The estimates of the differential selection on individual mutations were also significantly correlated between replicates , although they were noisier than the per-site ones ( S2 Fig ) . One way to test if the sensitivity of our experiments exceeded their noise is to compare the magnitudes of the differential selections observed in the actual selection with MxA-expressing cells versus the control selection with cells expressing inactive MxA ( Fig 1A ) . Fig 3B shows the distribution of differential selection values across all sites as estimated in the MxA and control selections for each replicate . For each replicate , there was a long tail of sites with strong differential selection in the MxA selection that exceeded any value observed in the control selection with inactive MxA . S2 Fig shows that similar results are obtained when examining differential selection at the level of mutations rather than sites . Therefore , at a subset of sites , MxA exerts selection that substantially exceeds the background noise of the experiments . We next tested whether our results were consistent with existing knowledge about how mutations to NP affect MxA resistance . For this test and the remainder of this paper , we use the average of the measurements from the two replicates ( S3 Fig shows these average measurements for all sites . ) We examined the three NP mutations previously shown to be mostly responsible for the increased MxA resistance of the human 1918 pandemic H1N1 strain relative to its avian influenza ancestors ( according to [17] , these are R100V , L283P , and F313Y ) . The Aichi/1968 NP in our study is a descendant of this 1918 NP and retains all three MxA-resistance mutations . We therefore expected that reverting these mutations would increase MxA sensitivity , barring effects due to changes in NP sequence context between 1918 and 1968 . Consistent with this expectation , the differential selection for reverting each mutation to its consensus identity in avian NP was negative ( S2 Fig ) . Two of three mutations also occurred at sites that had differential selection that greatly exceeded the median in both the MxA and control selections ( Fig 3C ) . These results show that despite a half-century of sequence divergence , the sites of the mutations that conferred MxA resistance on the 1918 virus remain important determinants of this phenotype in the NP used in our study . We next sought to identify the NP sites that most affected MxA resistance . There were 29 sites where the differential selection from MxA exceeded the background noise in the control selection . Fig 4A shows these 29 sites ranked by their differential selection values . Site 283 , which is one of the sites most responsible for the MxA resistance of the 1918 pandemic virus [17] , ranks second in our data ( Fig 4A ) . But most sites predicted by our deep mutational scan to have the largest effects on MxA resistance have not previously been described as impacting this phenotype . Interestingly , at all these sites , the greatest differential selection is from mutations away from the wildtype amino acid that increase MxA sensitivity . Strikingly , 26 of the 29 sites where mutations most influence MxA resistance have the same consensus amino acid in avian and human influenza NP sequences ( Fig 4A ) . Therefore , although these sites appear to be important determinants of the restriction of NP by MxA , they have not undergone extensive substitution during the adaptation of influenza virus to humans—presumably because they already possess an amino acid that confers resistance . The broad conservation of these sites also explains why they have not previously been identified by studies examining mutations that fixed during recent influenza evolution in nature . The sites that most affected MxA resistance are on the surface of the monomeric structure of NP ( Fig 4B ) . Twelve of these sites clustered at the base of the NP body domain , which is also the location of the three mutations that contributed to the MxA resistance of the 1918 virus [17] . The remaining sites mostly clustered either in a flexible basic loop known to affect RNA binding ( residues 73 to 91 ) or in the N-terminus of NP , which is disordered [41–44] . Overall this structural mapping reinforces the central importance of the base of the NP body domain to MxA sensitivity , but shows that other NP surface regions may also contribute . The two NP sites with the greatest differential selection were 51 and 283 ( Fig 4A ) . The mutation L283P has previously been characterized as increasing the MxA resistance of the 1918 virus relative to its avian ancestor [17] , and our high-throughput data concur that mutating this site to the avian identity ( or indeed to almost any amino acid other than P ) substantially increases MxA sensitivity ( Fig 4A , S4 Fig ) . But while site 283 has clearly undergone important change during influenza evolution , site 51 is almost completely conserved as D across all NPs from human , swine , equine , and avian influenza A strains ( S1 Table ) . Our high-throughput data suggest that mutating site 51 to most other amino acids should greatly increase MxA sensitivity . Structurally , site 51 is located near site 283 ( Fig 4B ) , and is adjacent to sites 52 and 53 , where mutations have been shown to affect the MxA resistance of the 2009 pandemic H1N1 [17] and H7N9 [22] strains , respectively . Interestingly , a mutation at site 51 that increased MxA sensitivity arose as a secondary change in an avian influenza virus that was engineered for increased MxA resistance [45] . To validate the finding of our high-throughput experiments that site 51 is a major determinant of MxA sensitivity , we engineered variants of the Aichi/1968 NP carrying a variety of mutations at this site . We selected five amino-acid mutations that our high-throughput data suggest should reduce MxA resistance by varying degrees ( S5 Fig ) . As a control , we also designed a synonymous mutation at site 51 ( D51Dsyn ) that was not expected to affect MxA sensitivity . To test these mutations , we measured the effect of each mutation on polymerase activity in the presence and absence of MxA . Active influenza polymerase can be reconstituted in cells , and this polymerase activity is sensitive to inhibition of NP by MxA [9] . We expected that polymerase activity would be more inhibited for mutant NPs that had increased MxA sensitivity . In the absence of MxA , the D51Dsyn mutation had similar polymerase activity to the wildtype NP while all five amino-acid mutations modestly decreased polymerase activity ( Fig 5A ) . We compared these activities in the absence of MxA to those measured in cells expressing MxA , and quantified the effect of each mutation on MxA resistance by dividing its activity in the presence of MxA by its activity in the absence of MxA . The wildtype NP and the D51Dsyn mutant were slightly inhibited by MxA , with activity decreasing to ∼80% of its original value ( Fig 5B ) . As predicted by our high-throughput deep mutational scanning , all five amino-acid mutants at site 51 were more strongly inhibited , with activity decreasing to between 24% and 54% of its original value ( Fig 5B ) . This result indicates that multiple different mutations away from D at site 51 substantially increase MxA sensitivity as measured by polymerase activity . To confirm that the decreased MxA resistance of polymerase activity correlated with the effect of MxA on viral replication , we carried out competition experiments between wildtype and mutant viruses . Such competition experiments provide a sensitive and internally controlled way to measure the relative fitness of two viral variants . We used reverse genetics to generate influenza viruses carrying wildtype NP , the D51Dsyn mutation , or the D51N mutation ( S7 Fig ) . We mixed each mutant virus with wildtype virus at a 1:1 ratio of infectious particles , and infected MxA-expressing or non-expressing cells at a low MOI . At 10 and 54 hours post-infection , we isolated viral RNA and determined the frequency of each variant by deep sequencing . As expected , the wildtype D51 variant greatly increased in frequency relative to the MxA-sensitive D51N mutant in MxA-expressing cells , whereas the two variants remained at similar frequencies in cells not expressing MxA ( Fig 6 ) . Also as expected , the wildtype D51 variant and its synonymous variant remained at roughly equal frequencies in the control competitions ( Fig 6 ) . This competition experiment verifies that an amino-acid mutation away from the wildtype identity at site 51 strongly increases MxA sensitivity as measured by viral growth . To expand the validation of the deep mutational scanning beyond site 51 , we chose 11 additional mutations for testing in the viral competition assay . We chose these mutations from the sets of sites ( Fig 4A ) and mutations ( S4 Fig ) under the strongest differential selection in our deep mutational scanning based on the following criteria: they were distributed across different regions of NP , they had consistent differential selection in both replicates of the deep mutational scan , and the deep mutational scanning data suggested that they supported good viral growth . The deep mutational scanning data predict that eight of these mutations ( E294R , N309R , L466G , Q4Y , M105G , Q12S , T23H , and S50C ) should increase MxA sensitivity , while three ( I41T , Q399R , and R102A ) should increase MxA resistance . We generated duplicate stocks of each mutant virus using reverse genetics . All mutant viruses grew to similar titers as wildtype except for the R102A mutant , which was significantly attenuated ( S7 Fig ) . We then tested each mutation in duplicate in the competition assay described in the previous section to determine its effect on MxA sensitivity . S8 Fig shows the full data from each duplicate competition assay . To summarize these data , for each mutation we computed the ratio of its frequency relative to wildtype in MxA-expressing cells versus the control non-expressing cells at both 10 and 54 hours . If a mutation increases MxA sensitivity then this enrichment ratio will be less than one , while if a mutation increases MxA resistance then this enrichment ratio will be greater than one . All mutations clearly had the predicted effect on MxA sensitivity at the 54-hour timepoint , and at least weakly had the predicted effect at the earlier 10-hour timepoint when selection has had less time to act ( Fig 7 ) . As expected , D51N and the eight additional putative sensitizing mutations were depleted in the MxA-expressing cells relative to the control cells , with D51N having the strongest effect ( Fig 7 ) . The three putative resistance mutations were all enriched in the MxA-expressing cells relative to the control cells , validating that they do indeed increase resistance ( Fig 7 ) . The largest increase in resistance was conferred by the R102A mutation . This resistance mutation also substantially attenuates viral growth ( S7 Fig ) , perhaps explaining why it is not fixed in the human influenza NP . However , there does not seem to be any general trend for mutations to have similar effects on viral growth and MxA resistance , as we identify attenuating mutations that increase both MxA resistance ( e . g . , R102A ) and sensitivity ( e . g . , D51N ) , while also identifying mutations with both effects on MxA sensitivity that do not greatly affect viral growth ( e . g . , E294R and Q399R ) . As expected , the control synonymous D51Dsyn mutation had no effect on MxA sensitivity ( Fig 7 ) . These results demonstrate that our deep mutational scanning approach accurately identifies mutations that increase both sensitivity and resistance to MxA . We also tested all of these mutations and a few others in a polymerase activity assay ( S9 Fig ) . This assay is easier than traditional viral growth assays , and so has often been used in the literature to test for mutational effects on MxA resistance . Notably , less than half of the mutations that were validated to affect viral sensitivity to MxA in Fig 7 also had significant effects in the polymerase activity assay . A similar discrepancy between the polymerase activity and viral growth assays has been observed in other studies , and is believed to result from the fact that the polymerase activity assay tests only a limited part of the influenza viral life cycle [46 , 47] . Therefore , a clear strength of our deep mutational scanning approach is it makes it feasible to test large numbers of mutations for their effects on viral growth .
We have used deep mutational scanning to experimentally estimate how MxA sensitivity is affected by every NP amino-acid mutation that supports viral growth . Our approach screens all mutations compatible with viral replication , and can identify changes that increase or decrease MxA resistance . In contrast , previous approaches have focused on mutations that fixed during viral adaptation to MxA in nature or in the lab [17 , 22] . These approaches have different strengths . Examining mutations that fix during viral adaptation elucidates the evolutionary pathways to MxA resistance . Our approach systematically maps how all sites in NP contribute to MxA resistance , without regard to whether mutating these sites in the starting virus can increase MxA resistance . The two approaches will yield similar results if all sites start with amino acids that confer sensitivity to MxA . But the approaches will yield different results if most sites in the starting virus already have amino acids that confer resistance to MxA . The most striking finding of our work is that the latter situation predominates: most sites with the largest effect on MxA resistance already possess amino acids that confer resistance . Furthermore , at most of these sites , the resistant amino acid is conserved across human and avian influenza strains . As a consequence , most sites that we identified could not have been found by looking for mutations that adapt influenza virus to MxA in nature or the lab for the simple reason that these sites are already fixed to a resistant amino acid . Why are so many sites in NP already fixed at MxA-resistant identities ? One speculative possibility is that homologs of MxA in other hosts have selected for some level of generalized MxA resistance in all NPs . Determining whether this is the case will require characterizing whether MxA homologs of the relevant species do in fact exert selection on influenza virus: there is evidence that swine MxA restricts influenza [48 , 49] , but the anti-influenza activity of avian MxA remains incompletely characterized across most bird species that are hosts for influenza [50–52] . Even if other MxA homologs exert selection , we would not expect non-human viruses to be optimally resistant to human MxA since MxA homologs have different specificities [53] . But similarities among the recognition mechanisms of MxA homologs could have driven fixation of resistant amino acids at many sites . Alternatively , perhaps many sites in NP have MxA-resistant amino acids due to some unknown selective pressure unrelated to MxA . In any case , our results demonstrate that it is important to establish the baseline when thinking about MxA resistance . While avian influenza strains are more sensitive to human MxA than human strains [16 , 17] , our results suggest that these avian strains still have amino acids that confer MxA resistance at most sites . Another question is whether MxA resistance comes at an inherent functional cost . Several studies have introduced resistance mutations from human NPs into avian NPs and found that the resulting viruses are attenuated [17 , 45] . One interpretation is that MxA resistance is inherently costly . But another interpretation is simply that amino-acid mutations are often detrimental , and that this is no more likely to be true of MxA-resistance mutations than MxA-sensitizing ones . In support of this idea , the MxA-sensitizing mutations that we identified at site 51 were all deleterious to viral polymerase activity and viral growth , as was the MxA-resistance mutation we identified at site 102 . We also identified mutations that increase both MxA sensitivity and resistance without attenuating viral growth . Therefore , a mutation’s effects on MxA sensitivity and viral replication are separable traits . In addition , the functional effects of mutations affecting MxA resistance may sometimes be idiosyncratic to the particular viral strain . For instance , the MxA-sensitizing D51N mutation was found to be beneficial for viral growth in an engineered avian influenza NP [45] , but was deleterious to both polymerase activity and viral growth in the Aichi/1968 NP that we used in our study . A related question is whether mutations that affect MxA resistance in one NP similarly affect resistance in NPs from more diverged viral strains . Our results suggest that the answer is yes . For instance , reverting each of the three mutations responsible for the MxA resistance of the 1918 virus [17] led to the expected decrease in MxA resistance in the Aichi/1968 NP used in our study . Similarly , our finding that D51N greatly increases MxA sensitivity agrees with another study [45] that reported this mutation also increased the MxA sensitivity of an engineered avian NP . Therefore , it appears that mutational effects on MxA sensitivity are at least somewhat conserved across NPs . We also find great consistency in the regions of NP where mutations affect MxA sensitivity . Despite the fact that most sites that we identified as contributing to MxA resistance are new , many of these sites map to the same regions of NP as previously characterized resistance mutations . About half the sites that we identified clustered at the base of the NP body domain , which is also the location of resistance mutations in the 1918 and 2009 pandemic viruses , as well as H7N9 influenza [17 , 22] . This solvent-exposed region , which is distinct from the NP surfaces important for RNA binding or interactions with the polymerase proteins [54 , 55] , could possibly be a binding interface between MxA and NP . However , we also found that MxA sensitivity was affected by some sites in surface-exposed loops distal from the base of the NP body domain . So although our results confirm that certain regions of NP are disproportionately important determinants of MxA sensitivity , the details of the inhibitory interaction between NP and MxA remain unclear [10 , 56] . Overall , we have used a powerful new deep mutational scanning approach to identify sites that affect the inhibition of a virus by a host restriction factor . This approach complements the traditional strategy of characterizing mutations at specific sites that differ across viral strains . An advantage of our approach is that it enables unbiased identification of all sites where mutations affect a phenotype , regardless of whether these sites have substituted during evolution . We envision that this approach can be extended to systematically examine how viral mutations affect additional homologs of NP and MxA , as well as to understand the interplay between viruses and other less well-characterized restriction factors [57 , 58] .
FASTQ files for the deep mutational scanning experiment and viral competition experiment are on the NCBI Sequence Read Archive with accession number SRP082554 . The computer code and input data files necessary to reproduce all the analysis in this work are available in S1 File and also at https://github . com/orrzor/2016_NP_MxA_paper ( last accessed August 15 , 2016 ) . The differential selection values estimated at each site in NP , and for each mutation at each site in NP , are in S2 and S3 Files respectively . We used lentiviral transduction to engineer MDCK-SIAT1 cells ( Sigma Aldrich ) to constitutively express human MxA or MxA-T103A under control of a CMV promoter . We placed a FLAG tag followed by a GSG linker ( DYKDDDDKGSG ) after the methionine start codon of human MxA . Downstream of MxA , we included an internal ribosome entry site ( IRES ) followed by the red fluorescent protein mCherry to act as a reporter for lentiviral transduction . At 48 hours after lentiviral transduction of the MDCK-SIAT1 cells , we single-cell cloned variants by serial dilution in 96-well plates . Wells with clonal transduced cells were identified by finding wells with single clusters of cells expressing mCherry . To verify that the recovered cell lines expressed MxA , we seeded the cells at 2 . 5 × 105 cells/well in D10 media ( DMEM supplemented with 10% heat-inactivated FBS , 2 mM L-glutamine , 100 U of penicillin/ml , and 100 μg of streptomycin/ml ) in 12-well dishes , and 20 h later , we collected the cells and performed Western blots to detect the FLAG tag on the MxA or β-actin as a loading control . To detect FLAG , we stained with a 1:5000 dilution of mouse anti-FLAG ( Sigma , F1804 ) followed by a 1:2500 dilution of Alexa Flour 680-conjugated goat anti-mouse ( Invitrogen , A-21058 ) . To detect β-actin , we stained with a 1:5000 dilution of rabbit anti-β-actin ( Abcam , ab8227 ) followed by a 1:2000 dilution of Alexa Flour 680-conjugated goat anti-rabbit ( Invitrogen , A-21109 ) . Blots were imaged using a Li-Cor Odyssey Infrared Imaging System . We measured polymerase activity for different NP mutants in the MDCK-SIAT1 cells expressing or not expressing MxA . For these assays , we co-transfected 500 ng of the pICR2-PB1flank-eGFP reporter plasmid along with 15 ng of indicated mutant pHW2000-Aichi68 NP and 125 ng each of HDM-Nan95-PA , HDM-Nan9-PB1 , and HDM-Nan95-PB2 into wells of 24-well dishes of MDCK-SIAT1 or MDCK-SIAT1-MxA cells . We chose this amount of NP plasmid because it was near the midpoint of the polymerase activity dose-response curve when holding the amounts of all other plasmids fixed ( S6 Fig ) . Transfections were performed with Lipofectamine 3000 ( ThermoFisher Scientific ) and the transfection mixes of DNA and lipofectamine were prepared according to the manufacturer’s protocol . The transfection itself was done using the modified protocol below , which was designed to increase transfection efficiency in MDCK-SIAT1 cells [66] . The transfection mix was incubated at room temperature for 1 hour and added to a well of a 24-well plate , and then 500 μL cells at 2 . 5 × 105 cells/mL was added to this well . After 4 hours , we changed the media to fresh D10 . At 20 hours post-transfection , we quantified the geometric mean of the GFP fluorescence by flow cytometry . We performed three biological replicates for each NP mutant , and each replicate used NP plasmid from an independent mini-prep . | During viral infection , human cells express proteins that can restrict virus replication . However , in many cases it remains unclear what determines the sensitivity of a given viral strain to a particular restriction factor . Here we use a high-throughput approach to measure how all amino-acid mutations to the nucleoprotein of influenza virus affect restriction by the human protein MxA . We find several dozen sites where mutations substantially affect the sensitivity of influenza virus to MxA . While a few of these sites are known to have fixed mutations during past adaptations of influenza virus to humans , most of the sites are broadly conserved across all influenza strains and have never previously been described as affecting MxA resistance . Our results therefore show that the known historical evolution of influenza has only involved substitutions at a small fraction of the sites where mutations can in principle affect MxA resistance . We suggest that this is because many sites are already broadly fixed at amino acids that confer high resistance . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"microbial",
"mutation",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"influenza",
"pathogens",
"microbiology",
"dna-binding",
"proteins",
"vertebrates",
"orthomyxoviruses",
"animals",
"viruses",
"polymerases",
"rna",
"viruses",
"... | 2017 | Deep mutational scanning identifies sites in influenza nucleoprotein that affect viral inhibition by MxA |
The European & Developing Countries Clinical Trials Partnership ( EDCTP ) is a partnership of European and sub-Saharan African countries that aims to accelerate the development of medical interventions against poverty-related diseases ( PRDs ) . A bibliometric analysis was conducted to 1 ) measure research output from European and African researchers on PRDs , 2 ) describe collaboration patterns , and 3 ) assess the citation impact of clinical research funded by EDCTP . Disease-specific research publications were identified in Thomson Reuters Web of Science using search terms in titles , abstracts and keywords . Publication data , including citation counts , were extracted for 2003–2011 . Analyses including output , share of global papers , normalised citation impact ( NCI ) , and geographical distribution are presented . Data are presented as five-year moving averages . European EDCTP member countries accounted for ~33% of global research output in PRDs and sub-Saharan African countries for ~10% ( 2007–2011 ) . Both regions contributed more to the global research output in malaria ( 43 . 4% and 22 . 2% , respectively ) . The overall number of PRD papers from sub-Saharan Africa increased markedly ( >47% ) since 2003 , particularly for HIV/AIDS ( 102% ) and tuberculosis ( TB ) ( 81% ) , and principally involving Southern and East Africa . For 2007–2011 , European and sub-Saharan African research collaboration on PRDs was highly cited compared with the world average ( NCI in brackets ) : HIV/AIDS 1 . 62 ( NCI: 1 . 16 ) , TB 2 . 11 ( NCI: 1 . 06 ) , malaria 1 . 81 ( NCI: 1 . 22 ) , and neglected infectious diseases 1 . 34 ( NCI: 0 . 97 ) . The NCI of EDCTP-funded papers for 2003–2011 was exceptionally high for HIV/AIDS ( 3 . 24 ) , TB ( 4 . 08 ) and HIV/TB co-infection ( 5 . 10 ) compared with global research benchmarks ( 1 . 14 , 1 . 05 and 1 . 35 , respectively ) . The volume and citation impact of papers from sub-Saharan Africa has increased since 2003 , as has collaborative research between Europe and sub-Saharan Africa . >90% of publications from EDCTP-funded research were published in high-impact journals and are highly cited . These findings corroborate the benefit of collaborative research on PRDs .
The European & Developing Countries Clinical Trials Partnership ( EDCTP ) , created in 2003 , is a partnership of 14 participating European Union ( EU ) Member States plus Norway and Switzerland , with sub-Saharan African countries . EDCTP aims to accelerate the development of new or improved drugs , vaccines , microbicides and diagnostics against poverty-related diseases ( PRDs ) including HIV/AIDS , tuberculosis ( TB ) , malaria and neglected infectious diseases ( NIDs ) [1] . Like other organisations that support research to generate new knowledge for translation into policy and practice , there is a need for EDCTP to assess the output and potential impact of the research that it funds . Although there are many indicators that can be used to measure research progress and knowledge , biomedical publications and their citations are widely used . Bibliometric methods have been used to analyse publication output and impact for specific disease areas , to quantify the volume of research output and compare the contributions from different institutions , countries and regions [2] . These methods can also be used to map research collaboration at the national , regional and international level and to compare its potential impact . The methods are based on mathematical and statistical techniques and therefore can provide a quantitative assessment of research performance [2] . Journal papers report research work . These papers refer to or ‘cite’ earlier relevant work and the new papers will be cited later , in their turn . The more citations a paper accumulates the more it is considered having ‘impact’ . Therefore , citation counts are recognised as a measure of impact , and can be seen as an indicator of the strength of the innovative research from a group , an institution , a country or a region . Most impact indicators do not use simple counts of citations; they use average ( normalised ) citation counts for defined groups of papers , as individual papers may have varying or unusual citation profiles . Citation rates differ across subject areas and over time . For example , citations in biological sciences occur more rapidly and plateau at a higher level than citations in physical sciences or mathematics , and citation rates are generally higher for natural sciences than for social sciences [3] . In addition , older papers have more time to accumulate citations than more recent ones . The main objective of this article is to assess the geographical and temporal trends in the publication of research papers on HIV/AIDS , TB , malaria and NIDs by European EDCTP member countries and sub-Saharan African countries through a bibliometric analysis . The secondary objectives of this article are to describe collaboration patterns and to assess the citation impact of research funded by EDCTP .
Publication data were drawn from the Thomson Reuters Web of Science that annually index the contents of over 12 , 000 journals worldwide and contains the Science Citation Index . We first drew out all relevant research publications on PRDs using search terms ( Table 1 ) anywhere in their title , abstract or keywords for the period 1 January 2003 to 31 December 2012 ( “publication date field” ) with citation data up to 31 December 2011 . Publications from substantive journal articles , reviews and proceedings papers published in peer-reviewed journals were included in the data extraction; hereafter referred to as ‘papers’ . We excluded editorials , meeting abstracts and other types of publications . All identified relevant papers from Web of Science in each disease area were included in the research and matched to PubMed using meta-data ( i . e . , digital object identifier , author , and source title ) [4] [5] to identify relevant epidemiological or clinical research papers . Epidemiological papers were identified using the PubMed Medical Subject Headings ( MeSH ) where the qualifier contained the term ‘epidemiology’[5] . Clinical trials research was identified using PubMed specific publication types ( ‘clinical trial’ ( including phase I-phase IV ) ; ‘controlled clinical trial’ or ‘randomized controlled trial’ ) or MeSH Headings containing either ‘clinical trial’ or ‘controlled trial’[5] . To assess the citation impact of research funded by EDCTP , the EDCTP publication database , which includes publications arising to EDCTP-funded research and available in-house ( n = 244 papers ) , was matched to the Web of Science . Subsequently , EDCTP- related papers were also identified by using funding acknowledgment data in Web of Science . This iterative process resulted in a final count of 437 EDCTP-related papers matched to the Web of Science and 258 papers were matched to the 2003–2012 main dataset . Of these , 237 papers were used in the citation analysis . To conduct region specific analysis , the European EDCTP member countries and sub-Saharan African partner countries were grouped into the respective regions as detailed in Table 2 . The data are presented as five-year moving averages ( i . e . , 2003–2007 , 2004–2008 , 2005–2009 , 2006–2010 , and 2007–2011 ) to smooth out the yearly fluctuations in the number of papers and their citations , allowing for accurate trend analysis . A paper was assigned to each country and each organisation that appeared at least once for any author on the paper . A paper was counted only once for each country and organisation , irrespective of the number of variations in the addresses . For example , a paper with four authors , all from the same country , with two authors from the same organisation and the other two from two different organisations would be assigned to the country only once and once to each of the three different organisations to avoid double-counting of either institutional or country-level data . Of note , international , governmental and non-governmental organisations were assigned according to the country given by the author . This includes organisations such as the World Health Organization ( WHO ) ( headquartered in Geneva , Switzerland ) . The source of funding has only been indexed in Web of Science consistently since mid-2008 , although it is generally acknowledged in publications . For this reason , and because authors do not always present their organisation in the same way , algorithms were used to unify the variants used for the various funding sources and organisations . All papers that acknowledged EDCTP were attributed to EDCTP; as the authors may have chosen to acknowledge EDCTP without having received funding support from EDCTP , these papers were compared with EDCTP’s records to confirm if funding had been received from EDCTP . Papers identified in the initial extraction were matched with publication data from EDCTP to produce a subset of papers for which EDCTP funding could be confirmed . Only papers for which EDCTP financial support was confirmed were included in analysis of the impact of EDCTP-funded research . Five bibliometric indicators were considered in our analysis; the first four indicators are based at paper level and the fifth at journal level . Co‐authorship is an index of research collaboration . The share of output that was collaborative across the disease areas were analysed using the address data associated with the publication . The software Wolfram Mathematica was used to create the maps and to produce a visual representation of the citation impact between EDCTP member countries ( institutions ) and sub-Saharan Africa and within sub-Saharan Africa . Disease burden data were obtained from the World Health Organisation ( WHO ) Global Burden of Disease estimates [9] . The overall burden of disease was assessed using the disability-adjusted life year ( DALY ) . DALY rates are expressed per 100 , 000 population using 2004 population estimates . The updated WHO global health estimates were only published after the completion of this study and were therefore not used .
Although the output for HIV/AIDS increased over the study period in Europe ( 17 . 2% ) and sub-Saharan Africa ( 102 . 4% ) , the percentage of the overall world output decreased for Europe ( from 36 . 4% in 2003–2007 to 34 . 0% in 2007–2011 ) , while the percentage of the overall world output for sub-Saharan African countries increased ( from 8 . 2% in 2003–2007 to 13 . 2% in 2007–2011 ) , particularly for Southern Africa ( from 4 . 5% in 2003–2007 to 7 . 8% in 2007–2011 ) ( Fig 1 ) . The output from European collaboration with sub-Saharan African countries nearly doubled over the study period and the share of this collaborative research as percentage of the overall world production increased from 3 . 3% in 2003–2007 to 5 . 2% in 2007–2011 ( Fig 1 ) . The normalised citation impact ( NCI ) of this collaborative research was higher ( NCI: 1 . 62 ) than for European ( NCI: 1 . 30 ) or sub-Saharan African countries ( NCI: 1 . 33 ) separately . Overall , institutions in the United Kingdom ( UK ) ( 2080 papers ) and France ( 706 papers ) were the leading European collaborating partners with sub-Saharan Africa in HIV/AIDS research and this research was well-cited ( NCI: 1 . 83 and 1 . 64 , respectively ) followed by collaborative research with institutions in the Netherlands ( 448 papers , NCI: 1 . 48 ) and Belgium ( 389 papers , NCI: 1 . 51 ) ( Fig 2 ) . Output , collaboration and normalised citation impact in HIV/AIDS research within sub-Saharan Africa was led by Southern ( NCI: 1 . 55 ) and East Africa ( NCI: 1 . 42 ) whereas research output , collaboration and normalised citation impact was lower in and between West ( NCI: 0 . 91 ) and Central Africa ( NCI: 0 . 82 ) ( Fig 2 ) . These results partly reflect the disease burden as countries with higher DALYs ( disability-adjusted life year ) were more active in this research area although some regional variation was observed ( Fig 2 ) . During the study period sub-Saharan African TB research output increased by 81% and the region’s share of global TB research increased from 8 . 0% in 2003–2007 to 10 . 3% by 2007–2011 . The normalised citation impact of this research was well above the world average rising from 1 . 18 in 2003–2007 to 1 . 67 in 2007–2011 . Only the normalised citation impact of papers from Central Africa ( NCI: 0 . 86 , 2007–2011 ) was below the world average . This increase in sub-Saharan African output and impact was mainly driven by Southern Africa , which accounted for nearly 64% of the entire sub-Saharan African TB research output and which was highly cited ( NCI: 1 . 97 , 2007–2011 ) . It would seem that a shift in focus towards HIV/TB co-infection was responsible for the trend of more research output from sub-Saharan Africa , particularly in more recent years ( Fig 3 ) . In Europe , the TB research output increased by 29 . 7% but the percentage share of world output decreased by 7 . 4% during the study period . Nevertheless , Europe produced over a third of the global research output in TB and this research was highly cited ( NCI: 1 . 37 , 2007–2011 ) . The collaboration between Europe and sub-Saharan Africa increased by 84 . 6% between 2003 and 2011 , driven by multidisciplinary HIV/AIDS and TB research which was well-cited ( NCI: 2 . 11 ) . Similar to HIV/AIDS research , leading European collaborating partners with sub-Saharan Africa ( mainly Southern and East Africa ) in TB research were institutions in the UK ( 809 papers , NCI: 2 . 19 ) , France ( 267 papers , NCI: 1 . 80 ) , the Netherlands ( 222 papers , NCI: 1 . 96 ) , and Switzerland ( 224 papers , NCI: 3 . 63 ) ( Fig 4 ) . As shown in Fig 4 , TB research efforts are not correlated with burden of disease , except for South Africa , and the research collaboration on TB within sub-Saharan Africa is weak . Over the study period , the European share of the world output for malaria research decreased slightly ( from 46 . 7% to 43 . 4% ) , while it remained stable in sub-Saharan Africa ( ~21% ) , despite increased numbers of papers in both regions ( 27 . 2% in Europe; 47 . 6% in sub-Saharan Africa ) ( Fig 5 ) . East Africa ( 9 . 1% ) and West Africa ( 8 . 1% ) lead in sub-Saharan Africa in terms of share of global malaria research output; the normalised citation impact for East Africa ( NCI: 1 . 75 ) was the highest but Southern Africa ( NCI: 1 . 65 ) was almost as high ( Fig 5 ) . At the end of the study period the normalised citation impact for collaborative African-European research was 1 . 81 , which was higher than the respective figures for either European ( NCI: 1 . 51 ) or sub-Saharan African research ( NCI: 1 . 47 ) ( Fig 5 ) . In malaria research , the leading European collaborating partners with sub-Saharan Africa were institutions in the UK ( 1 , 836 papers , NCI: 2 . 19 ) , France ( 739 papers , NCI: 1 . 37 ) , Switzerland ( 552 papers , NCI: 2 . 29 ) , and Germany ( 494 papers , NCI: 1 . 24 ) ( Fig 6 ) . As shown in Fig 6 , there is little correlation between burden of disease and malaria research output in sub-Saharan Africa . Compared with HIV , TB and malaria the research output on NIDs was lower during the study period and the research was less well cited ( Fig 7 ) . Again , despite increased numbers of papers in both regions ( Europe 18 . 8%; sub-Saharan Africa 45 . 4% ) , the share of the world output for NID research decreased in Europe ( 37 . 1% to 32 . 6% ) while it remained stable , at about 7% , in sub-Saharan Africa over the study period ( Fig 7 ) . Although the citation rates of NID research in sub-Saharan Africa were modest ( NCI: 1 . 10 vs . 1 . 25 for European countries ) , the normalised citation impact of research from Central Africa in particular increased , but also from West and East Africa , suggesting more recent research activity in these regions . The normalised citation impact for European collaboration with sub-Saharan Africa was 1 . 34 at the end of the study period ( Fig 7 ) . Leading European collaborating partners with sub-Saharan Africa in NID research were institutions in the UK ( 882 papers , NCI: 1 . 53 ) , France ( 411 papers , NCI: 1 . 20 ) , Switzerland ( 273 papers , NCI: 1 . 98 ) , Germany ( 245 papers , NCI: 1 . 07 ) and Belgium ( 238 papers , NCI: 1 . 61 ) . Some collaborative research between institutions in West , Central and East Africa exists but these links were not very strong ( Fig 8 ) .
Our results show that since 2003 the total research output in PRDs and their associated normalised citation impact and world share from sub-Saharan Africa has risen substantively , particularly in HIV/AIDS and TB , and principally involving Southern and East Africa . The research output in PRDs from sub-Saharan Africa mainly includes clinical and epidemiological research . Furthermore , although European research output generally has increased , its share of the global output for PRDs has dropped . However , the overall output from European research collaboration with sub-Saharan Africa for PRDs has increased over the period studied and the normalised citation impact of this collaborative research is generally higher than that for either European or sub-Saharan African research not involving North-South collaboration . The research output from EDCTP-funded research projects has a higher normalised citation impact for both HIV/AIDS and TB and more than 90% of the papers from EDCTP-funded projects are published in high impact journals ( i . e . in the first and second quartile of journals by journal impact factor in their Web of Science journal category ) . Although sub-Saharan Africa assumes the heaviest burden of PRDs its share of global PRD research output is still relatively modest compared with European EDCTP member countries that have a much higher research output in this area [10] . For the period 2007–2011 , Europe accounted for approximately one-third of the world’s research output on HIV/AIDS , TB and NIDs and 43% on malaria . However , the fall in the European share of the global research output on PRDs is due to the rising research output from BRICK ( i . e . Brazil , Russia , India , China and South Korea ) economies , in addition to the growth in research output from sub-Saharan Africa [11–13] . For the period 2007–2011 , sub-Saharan Africa’s share of global research was higher for malaria ( 22% ) than for the other PRDs ( HIV/AIDS 13 . 2%; TB 10 . 3%; NIDs 7 . 6% ) . Since 2003 , the African research output on PRDs has steadily increased , although not uniformly for all diseases or geographic areas . Sub-Saharan Africa’s research output on HIV/AIDS and TB grew faster than that for malaria and NIDs and while Southern Africa dominated the sub-Saharan African HIV/AIDS and TB research output , Eastern Africa dominated the sub-Saharan African malaria and NIDs research output . In contrast , during the study period , Central Africa had a low level of research output for PRDs despite being a region with a high disease burden , particularly for malaria and NIDs [14] . However , although there were a low number of papers , their normalised citation impact for malaria ( NCI: 1 . 31 vs . 1 . 22 ) and NIDs ( NCI: 1 . 17 vs . 0 . 97 ) is above the world average and thus not insignificant . The observed trend in research output from sub-Saharan Africa suggests that sustained support and funding by governments , development partners , private foundations and public partnerships , such as EDCTP , to address Millennium Development Goal 6 to combat HIV/AIDS , malaria and other diseases is having a positive impact progressively . A very large share of sub-Saharan African research on PRDs is a result of international collaboration . For 2007–2011 , 40% , 53% , 59% , and 60% of clinical research on HIV/AIDS , TB , malaria and NIDs , respectively was conducted in collaboration with European EDCTP member countries . The United Kingdom and France have traditionally been part of these collaborations due to their historic colonial ties with many countries in sub-Saharan Africa . The collaboration trends show that several European countries are increasing their research interests in PRDs with increasing collaboration among themselves and sub-Saharan African countries , particularly Belgium , Denmark , the Netherlands , Sweden and Switzerland . This could be attributed partly to funding instruments like the EDCTP programme that actively promotes cross-national research through North-South , North-North and South-South networking . As highlighted by the World Bank’s recent publication , a high collaboration rate reflects the “noteworthy effort and interest of academia outside of Africa to support sub-Saharan Africa’s research capacity” [15] . This report pointed out that “international collaboration is highly instrumental in raising the impact of sub-Saharan Africa’s publications” [15] . Our results corroborate this statement as illustrated in the example of Central Africa , where few research groups operate , our results show that their research activities are highly collaborative ( HIV/AIDS: ≥40%; TB; ≥53%; malaria and NIDs ≥60% ) resulting in normalised citation impact for malaria and NIDs that are above the world average . This trend was also seen for research activities funded by EDCTP . The normalised citation impact of EDCTP-funded collaborative research on PRDs between European EDCTP member countries and sub-Saharan African countries was higher than the respective figures for either European or sub-Saharan African research . These findings suggest that collaborative research is of mutual benefit to researchers in sub-Saharan Africa and Europe . However , a point of concern is that a large percentage of the sub-Saharan African research output results from collaborative projects , meaning that there is still very limited output that comes from purely sub-Saharan African partnerships and our analysis of South-South collaboration links supports this finding . This is in part due to a lack of critical mass of researchers in sub-Saharan Africa , and where local scientific leadership exists , this is mainly established through external research funding that dictates the collaborating partners . This underpins the need to bolster funding for health research and development by African governments as major contributors and to improve policymakers’ understanding of the value of research to drive national health priority setting [16] , [17] . Clearly , more political commitment and increased capacity building is needed across the board to enable existing and new research communities in sub-Saharan Africa to not only sustain the current research output but to promote increased research output from South-South collaborations on diseases and needs relevant to the specific regions . We observed an increasing trend for research output to be published in open-access journals for 2007–2011 . Since the early 1990s , scientific peer-reviewed publication has been revolutionised with an increasing number of open-access only journals and , more recently , traditional subscription journals offering an open-access option ( hybrid journals ) [18] , [19] . In addition , many major research funders such as the European Union , Medical Research Council ( MRC ) , Wellcome Trust and the National Institutes of Health ( NIH ) now require open-access publication of results from the projects that they fund , either in an open-access journal ( gold open-access ) or in a full-text archives , such as the National Library of Medicine’s PMC ( green open-access ) [20–23] . According to the European Commission: “open access to scientific peer-reviewed publications has been anchored as an underlying principle of the Horizon 2020 [the European Union Framework programme for Research and Innovation] Regulation and Rules of Participation” [23] , [24] . Although , there have been some doubts about the quality of peer-review in some open-access journals , a recent study found that among newer journals , open-access journals were being cited as often as subscription-based journals and , in some subcategories , they are being cited more [18] , [25] . Even though some studies have reported no evidence of a citation advantage for open-access journals compared with subscription journals , publishing research results in open-access journals should increase its accessibility to more readers , which is important in sub-Saharan Africa [26] , [27] . We used the Web of Science database to analyse the research output on PRDs from European EDCTP member countries and sub-Saharan African countries . This database contains over 12 , 000 of the highest impact journals , including open-access journals , selected using rigorous editorial and quality criteria . This database has several advantages over others , as it provides data on both scientific productivity and impact through the citation count and includes information on all institutions participating in Thomson’s Reuters Web of Science and their country of origin , which is not included in Medline , allowing for the quantification of collaboration between countries that publish these research activities . However , while coverage of English-language journals is very comprehensive , one limitation of the Web of Science is that coverage of non-English-language journals is less extensive , although this has recently increased with the inclusion of French and Portuguese journals in particular . Another potential limitation of this analysis is the choice of the Web of Science as our only data source . This database was the first to be established , but other databases are now available , such as Scopus , which includes nearly 22 , 000 journals [28] . One study that compared the Web of Science and Scopus reported a higher citation rate for several tropical diseases ( including malaria and some NIDs ) with Scopus , a significantly higher citation rate for TB with the Web of Science and no significant difference for HIV/AIDS [29] . The author suggested that this was because Scopus abstracts more from biomedical journals than the Web of Science and journals from developing countries are more likely to be included in Scopus . However , the authors did not provide information on how the search strategy was optimised for both databases nor if the results were normalised and both of these factors can have an important impact on the reliability of the results . A general limitation of our analysis is that bibliometrics is not an ‘exact science’ and relies upon interpretation and re-iteration to achieve a ‘best fit’ dataset that will adequately describe the research area whilst excluding papers of marginal relevance . The results must be interpreted with this in mind . Another potential limitation of our analysis is the method used to assign papers to organisations . Authors often report their affiliations in different ways for different publications , so we used an algorithm to unify these affiliations . International governmental and non-governmental organisations , including organisations such as WHO were assigned to the country given by the author . This may have resulted in an inflated research output for some countries ( e . g . Switzerland for WHO , Luxembourg for Médecins sans Frontières ) . Additionally , we were looking at collaboration patterns based on papers only , whereas some research collaboration takes time to set up and may not produce papers for many years . Given that there is a time lag between research funding and publication and between publication and citation more recently funded research is less likely to be published and less likely to be cited compared with research that has had more time to accumulate citations . As less than 30% of the EDCTP-funded clinical trials had been completed at the time of this analysis , our results inevitably show an incomplete picture . Nonetheless , the results have already revealed an increase in high impact papers across the three main PRDs . We expect the observed trends to continue as suggested by the number of recent , high impact relevant papers that have been published since 2013 and , therefore , were not included in the current analysis [30] , [31] . In the future , with more EDCTP-associated papers published , we hope to establish more exact estimates of the difference in research performance and relevant benchmarks as observed in the Impact Profiles . In conclusion , the normalised citation impact of research from sub-Saharan-Africa increased substantially from 2003 to 2011; collaborative research had a higher impact and was more highly cited than non-collaborative research and EDCTP-funded research was particularly highly cited and published in high impact journals . Since one of the core elements of EDCTP is to promote collaborative research , the association between highly collaborative research and high citation supports the inference that the partnership plays an important role in promoting sub-Saharan African leadership in PRDs research . | The European & Developing Countries Clinical Trials Partnership ( EDCTP ) was created in 2003 as a European response to the global health crisis caused by the three main poverty-related diseases ( PRDs ) of HIV/AIDS , tuberculosis and malaria . EDCTP funds research focusing on clinical trials for diagnosing , preventing and treating these diseases . We conducted a bibliometric analysis to 1 ) measure research output and citation impact from European and African researchers working on PRDs , 2 ) describe collaboration patterns , and 3 ) assess the citation impact of research funded by EDCTP . Citation analysis is a commonly used bibliometric tool to analyse scientific literature . Overall , the volume and citation impact of papers from sub-Saharan Africa has increased since 2003 , as has collaborative research between Europe and sub-Saharan Africa . Papers arising from collaborative research had a higher citation impact than non-collaborative research and >90% of publications from EDCTP-funded research projects were published in high-impact journals . These results suggest that research on PRDs in sub-Saharan Africa is growing and that the EDCTP partnership contributes to high-impact , collaborative research published in high-impact journals . By providing research funds and supporting activities to strengthen the research environment , the partnership contributes to sub-Saharan African researchers taking the lead in PRD research . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Bibliometric Assessment of European and Sub-Saharan African Research Output on Poverty-Related and Neglected Infectious Diseases from 2003 to 2011 |
Calorie restriction ( CR ) robustly extends the lifespan of numerous species . In the yeast Saccharomyces cerevisiae , CR has been proposed to extend lifespan by boosting the activity of sirtuin deacetylases , thereby suppressing the formation of toxic repetitive ribosomal DNA ( rDNA ) circles . An alternative theory is that CR works by suppressing the TOR ( target of rapamycin ) signaling pathway , which extends lifespan via mechanisms that are unknown but thought to be independent of sirtuins . Here we show that TOR inhibition extends lifespan by the same mechanism as CR: by increasing Sir2p activity and stabilizing the rDNA locus . Further , we show that rDNA stabilization and lifespan extension by both CR and TOR signaling is due to the relocalization of the transcription factors Msn2p and Msn4p from the cytoplasm to the nucleus , where they increase expression of the nicotinamidase gene PNC1 . These findings suggest that TOR and sirtuins may be part of the same longevity pathway in higher organisms , and that they may promote genomic stability during aging .
In the budding yeast Saccharomyces cerevisiae , replicative lifespan is measured by the number of divisions that a mother cell undergoes before senescing [1–3] . A primary cause of aging in this organism is homologous recombination between ribosomal DNA ( rDNA ) repeats , resulting in the formation of extrachromosomal rDNA circles ( ERCs ) that accumulate to toxic levels in mother cells [4] . Sir2p and a closely related homolog , Hst2p , belong to the sirtuin family of NAD+-dependent deacetylases [5] that can forestall aging by stabilizing the rDNA locus [4 , 6 , 7] . Although rDNA recombination is not known to play a role in the aging of metazoans , the function of Sir2p enzymes in lifespan determination appears to be conserved . In Caenorhabditis elegans and Drosophila melanogaster , additional copies of the SIR2 gene or pharmacological modulation of the Sir2p deacetylase also extend lifespan [8–11] . The diet known as calorie restriction ( CR ) prolongs the lifespan of numerous species , including fungi , invertebrates , and mammals [1 , 12 , 13] . Whether or not Sir2p enzymes play a role in CR-mediated lifespan extension is hotly debated . In support of their playing a role , additional copies of either SIR2 or HST2 suppress rDNA recombination and extend yeast replicative lifespan , whereas strains lacking SIR2 and HST2 fail to live longer when subjected to CR [14 , 15] . Similarly , CR diets or genetic mimics of CR fail to extend the lifespan of D . melanogaster and C . elegans lacking Sir2p [12 , 16] . However , other researchers favor a model in which Sir2p plays no role in CR-mediated lifespan extension , and instead the TOR ( target of rapamycin ) pathway is proposed to play the central role [17 , 18] . TOR is a nutrient-responsive phosphatidylinositol-kinase-related kinase that regulates protein synthesis and cell growth , and is inhibited by rapamycin , an immunosuppressive and anticancer drug that specifically inhibits TOR [19] . It has recently been discovered that lifespan can be extended in a variety of species by inhibition of TOR signaling , including S . cerevisiae , C . elegans , and D . melanogaster [17 , 20 , 21] . The mechanism by which inhibition of TOR signaling extends lifespan is unclear , but it has been proposed that it may act by altering ribosome assembly and translation [17 , 21–23] . Similar to CR , inhibition of TOR signaling can extend yeast lifespan in the absence of SIR2 [17] , but whether SIR2 plays a role in lifespan extension during inhibition of TOR signaling is not known . We have previously investigated the pathways by which CR operates in yeast . The enzymatic activity of Sir2p is regulated by endogenous levels of nicotinamide ( NAM ) , a sirtuin inhibitor [24] . Yeast strains grown on standard 2% glucose medium have an intracellular concentration of ∼50 μM NAM , which is almost precisely the IC50 of Sir2p [24–28] . We and others have shown that CR and other mild stresses , including heat stress and osmotic shock , extend yeast lifespan by increasing expression of the PNC1 gene , which encodes a nicotinamidase [27 , 29] . Recent evidence indicates that mammalian Nampt/PBEF , a putative functional ortholog of PNC1 that is required for the conversion of NAM to NAD+ , also regulates sirtuin activity [30 , 31] . PNC1 is an intriguing longevity gene because its expression is regulated by environmental stimuli that extend lifespan , such as heat , osmotic stress , low amino acids , and CR . Binding sites for the stress-responsive zinc-finger transcription factors Msn2p and Msn4p have been identified in the PNC1 promoter [32] . Msn2p and Msn4p have previously been shown to regulate chronological lifespan extension in response to deletion of the yeast Akt homolog SCH9 by controlling the expression of the superoxide dismutase SOD2 [33 , 34] . Msn2p/4p are therefore good candidates to regulate PNC1 expression in response to nutrient availability , but a previous study , which used a cdc25–10 mutant strain to mimic CR , concluded that MSN2 and MSN4 play no role in CR-mediated lifespan extension [14] . Here we show that Msn2p/4p are important regulators of yeast replicative lifespan and that they relocalize from the cytoplasm to the nucleus during CR , where they bind to and activate the PNC1 gene . Moreover , inhibition of TOR signaling acts via this same pathway to promote the expression of PNC1 and suppress rDNA recombination . These data provide evidence for a pathway from the cell's environment to an actual cause of aging , via which both CR and TOR signaling modulate lifespan .
We began by asking whether MSN2/4 are mediators of yeast lifespan extension by CR ( 0 . 5% glucose ) . In contrast to previous work using a genetic mimic of CR [14] , we found that lifespan extension by CR was completely MSN2/4-dependent ( Figure 1A ) . Single deletions of MSN2 or MSN4 did not block the ability of CR to extend lifespan , consistent with their known redundancy ( Figure 1B ) . Interestingly , MSN2 and MSN4 were also absolutely necessary for lifespan extension resulting from TOR inhibition ( Figure 1C ) . The requirement of MSN2/4 for CR- and TOR-mediated lifespan extension indicated that CR and TOR might extend lifespan via the same mechanism . We and others have presented evidence that CR works by increasing Sir2p activity [8 , 14 , 29] , as indicated by increases in telomeric silencing and a reduction in rDNA recombination [7 , 15 , 35] . We therefore examined whether rapamycin increased telomeric silencing and reduced rDNA recombination , and whether this was altered by the presence or absence of MSN2/4 or PNC1 . We observed that rapamycin increased telomeric silencing in a wild-type strain , but not in strains lacking MSN2/4 or PNC1 ( Figure S1 ) . Treatment with rapamycin also suppressed rDNA recombination and , again , this effect required MSN2/4 and PNC1 ( Figure 2A and 2B ) . The effect of rapamycin on silencing and rDNA recombination was also completely blocked by treating cells with NAM ( Figure 2B ) , indicating that a sirtuin is likely required for this effect . Neither CR nor rapamycin increased protein levels of Sir2p ( Figure 2C ) . Taken together , these data indicate that inhibition of TOR signaling increases the activity of Sir2p and/or another sirtuin . In addition to Sir2p , S . cerevisiae contains four HST ( homolog of sir two ) genes , HST1–4 . We and others have uncovered considerable redundancy in the sirtuin family , with Hst1p and Hst2p able to substitute for Sir2p during CR [15] . Recent studies in our laboratory have demonstrated that overexpression of any one of the sirtuin genes HST1–4 can suppress rDNA recombination ( D . Lamming and M . Latorre-Esteves , unpublished data ) . Thus , we were not surprised to find that rapamycin could suppress rDNA recombination in a sir2Δ fob1Δ strain ( Figure S2 ) . This result supports a previously published report that inhibition of TOR signaling can extend lifespan in the absence of SIR2 [17] . Although TOR inhibition can suppress rDNA recombination in the absence of SIR2 , deletion of both SIR2 and HST2 blocked the ability of rapamycin to suppress rDNA recombination ( Figure S2 ) . Consistent with these data , we found that rapamycin could extend the lifespan of a sir2Δ fob1Δ strain but had no effect on the lifespan of a sir2Δ hst2Δ fob1Δ strain ( Figure S2 ) . Thus , rather than working through a single sirtuin , we favor a model in which inhibition of TOR signaling suppresses rDNA recombination and promotes longevity by activating multiple sirtuins , including Sir2p and Hst2p . In agreement with this model , PNC1 was required for lifespan extension by rapamycin ( Figure 2D ) , and overexpression of PNC1 was sufficient to suppress rDNA recombination and extend lifespan in an msn2Δ/4Δ background ( Figure 2E and 2F ) . Thus , within the framework of our model , TOR signaling is upstream of PNC1 and PNC1 is downstream of MSN2/4 . The simplest mechanistic explanation is that inhibition of TOR activates MSN2 and MSN4 , which then increase the expression of PNC1 . To test this , we examined the effect of CR and rapamycin on Pnc1p levels in the presence and absence of MSN2/4 . Both treatments induced Pnc1p-GFP ( Figure 3A ) as well as native Pnc1p ( Figure 3B ) . In contrast , there were considerably lower levels of Pnc1p in the untreated msn2Δ/4Δ strain , and Pnc1p levels remained below those of the untreated wild-type strain , even in response to CR or rapamycin ( Figure 3B ) . Msn2p/4p are normally maintained in the cytoplasm by the activity of PKA . As cAMP levels fall , PKA activity decreases , and dephosphorylated Msn2p/4p relocalize to the nucleus [36] . A cdc25–10 mutant that has constitutively decreased cAMP/PKA signaling [37] expressed higher levels of Pnc1p than the wild-type strain , and this was MSN2/4-dependent ( Figure 3C ) . Upregulation of PNC1 expression by other stresses was also MSN2/4-dependent ( Figure 3D ) , and in agreement with the lifespan data in Figure 1B , single deletion of MSN2 or MSN4 did not greatly affect the ability of CR to induce PNC1 ( Figure S3 ) . We have previously shown that heat shock extends lifespan in a PNC1-dependent manner [29] , so we were curious whether this was an MSN2/4-dependent process . Heat shock induced the expression of PNC1 and extended lifespan even in the absence of MSN2/4 ( Figure 3D and 3E ) . To explore the MSN2/4-independent mechanism by which heat shock induces PNC1 , we analyzed the PNC1 promoter and found putative binding sites for the heat shock factor Hsf1p ( positions −251 to −275 and −319 to −353 , with respect to the start codon ) . Since HSF1 is an essential gene , we used an msn2Δ/4Δ strain containing doxycycline-repressible HSF1 to examine the role of Hsf1p in PNC1 regulation [38] . Repression of HSF1 largely blocked the ability of heat shock to upregulate PNC1 ( Figure 3F ) , but it did not affect the ability of CR or rapamycin to induce PNC1 ( data not shown ) . Under standard yeast growth conditions ( 2% glucose , 30 °C ) , Msn2p/4p localize predominately to the cytoplasm [36] , but in response to a variety of stresses , Msn2p/4p localize to the nucleus , where they activate stress-responsive genes [39] . We observed that CR induced the translocation of Msn2p into the nucleus , and the extent of the translocation was proportional to the degree of CR ( Figure 4A and 4B ) . CR also induced the nuclear localization of Msn4p , but quantification of the nuclear localization indicated that Msn4p was less sensitive to glucose restriction than Msn2p ( Figure S4 ) . The reasons for this difference are unknown , but it may allow for differential gene regulation as nutrients are depleted . Nuclear localization of Msn2p also increased in both tor1Δ and cdc25–10 strains ( Figure 4C and 4D ) , in agreement with previous reports [40 , 41] . Cells grown in moderate CR conditions ( 0 . 5% and 0 . 1% glucose ) showed heterogeneous localization patterns for Msn2p/4p , indicating that Msn2p/4p might be oscillating between the nucleus and cytoplasm , as has recently been noted for cells exposed to light stress or osmotic shock [42] . To determine if this was the case , time-lapse photomicrographs of cells expressing Msn2p-GFP were taken at 30-s intervals during growth under various glucose concentrations . The vast majority of cells grown in 2% glucose showed a cytoplasmic localization of Msn2p ( Figure 5A ) , while cells incubated in medium lacking glucose had exclusively nuclear localization of Msn2p ( Figure 5B ) . These patterns did not change over time . However , in cells grown under intermediate levels of CR ( 0 . 1% or 0 . 5% glucose ) , nucleo-cytoplasmic oscillations of Msn2p-GFP were observed ( Figures 5C and S5 ) . The GFP signal eventually became bleached ( data not shown ) , demonstrating that Msn2p was not simply being degraded in the nucleus and re-synthesized in the cytoplasm but was instead actively being transported in and out of the nucleus [43] . Msn4p-GFP showed a similar oscillatory behavior , and localization of Msn2p or Msn4p under 2% or 0 . 5% glucose was not altered by a lack of the other transcription factor ( data not shown ) . The PNC1 promoter contains four Msn2p/4p-binding sites known as stress response elements ( STREs ) [32] . To determine whether Msn2p/4p directly regulate PNC1 in response to CR , we tested several reporter constructs containing different regions of the PNC1 promoter ( Figure 6A ) . The full-length promoter construct was greatly induced in response to CR ( Figure 6B ) , whereas the reporters lacking one or more of the STREs were induced at significantly lower levels . The construct with no STREs showed no induction . The involvement of MSN2/4 was confirmed by the finding that CR completely failed to upregulate the reporter in an msn2Δ/4Δ strain ( Figure 6C ) . The ability of CR to induce the PNC1 reporter was largely unaffected in single msn2Δ or msn4Δ mutants ( Figure 6D and 6E ) , which is consistent with the ability of CR to extend the lifespan of single but not double msn mutants ( see Figure 1A and 1B ) . Interestingly , deletion of MSN4 did not affect the expression of the STRE4 reporter construct under 2% glucose , but significantly reduced its response to CR ( Figure 6E ) . In contrast , the STRE2 reporter construct was unaffected by the MSN4 deletion , indicating that the influence of Msn4p on PNC1 expression may be mediated by the distal two STRE elements . We have previously shown that while single deletion of MSN2 impairs the ability of PNC1 expression to be induced by stress , single deletion of MSN4 has little effect ( Figure S3 ) . Furthermore , Msn2p localizes to the nucleus more readily in response to decreased glucose concentration ( compare Figures 4 and S4 ) , deletion of MSN2 results in approximately 2-fold less expression of the reporter construct than deletion of MSN4 , and the STRE2 constructs were less responsive to CR in an msn2Δ strain than in an msn4Δ strain ( Figure 6D and 6E ) . Together these data support the conclusion that Msn2p is more important than Msn4p for regulating PNC1 expression in response to CR . Msn2p also appears to be more important for the response of PNC1 to heat stress ( 37 °C ) . Deletion of MSN2 resulted in an approximately 2-fold decrease in the expression of the STRE4 reporter construct in response to heat , relative to the wild-type strain ( compare Figure 6B and 6D ) . In contrast , single deletion of MSN4 actually leads to increased expression of the STRE 4 reporter in response to heat ( compare Figure 6B and 6E ) . This upregulation in the msn4Δ strain is likely to be due to MSN2 , because deletion of MSN2 in the msn4Δ strain resulted in the poorest response of the STRE4 reporter to heat ( Figure 6C ) . The data also indicated that elements present in the STRE4 reporter but not in the STRE2 reporter are responsible for the MSN2/4-independent induction of the promoter in response to heat stress . Consistent with this hypothesis , the putative binding sites for Hsf1p identified by our promoter analysis lie within this region . Given the relative importance of Msn2p , we next asked if the nuclear localization of Msn2p was sufficient to induce PNC1 . We expressed a constitutively nuclear mutant of Msn2p , Msn2p ( S4A ) [39] , in an msn2Δ and msn4Δ mutant , and observed robust induction of Pnc1p ( Figure 7A ) , demonstrating that nuclear localization of Msn2p is sufficient to induce PNC1 expression . We also noted that this strain grew poorly , in agreement with previous reports that constitutive nuclear localization of Msn2p/4p antagonizes growth [44] . The data thus far strongly supported a model in which CR promotes the binding of Msn2p directly to the PNC1 promoter , yet there remained the possibility that PNC1 was regulated by Msn2p indirectly . To distinguish between these two possibilities , we used chromatin immunoprecipitation to determine if Msn2p binds to the PNC1 promoter in response to CR ( Figure 7B ) . We detected Msn2p at the PNC1 promoter , and the apparent abundance of Msn2p was proportional to the degree of CR ( Figure 7C ) , which was consistent with the glucose-dependent nuclear localization of Msn2p that we previously observed . In order to better understand the target specificity of Msn2p/4p in response to CR , we utilized previously published microarray data [45 , 46] to examine the expression of 82 previously identified STRE-containing genes [47] . We were surprised to find that the genes varied greatly in their responsiveness to conditions of low glucose . Ten genes , including PNC1 , were highly responsive to small changes in glucose concentration , whereas other STRE-containing genes were responsive only to large changes in glucose concentration or were unresponsive ( Table S1 ) . In an effort to understand why some genes are more responsive to Msn2p/4p than others , we analyzed the promoters of these STRE-containing genes . The genes that were less responsive to low glucose averaged fewer STRE elements than the highly responsive genes and , on average , had binding sites approximately 50–60 base pairs further from the start of the coding sequence than genes that responded to low glucose ( Table S2 ) . We speculate that the placement of transcription factor binding sites at varying distances from the promoter may be a conserved mechanism for the differential regulation of stress-induced genes . Because our work indicated that two major longevity pathways , namely CR and TOR , promote lifespan by inducing the expression of PNC1 , we wondered whether other yeast longevity genes also modulate PNC1 expression . Deletion of the glycolysis pathway gene HXK2 extends lifespan and has been proposed as a genetic mimic of CR [15 , 17 , 37] . Consistent with this , an hxk2Δ strain had higher levels of PNC1 ( Figure S6 ) , placing it upstream of PNC1 . Similarly , there are abundant data linking the Snf1p/AMPK pathway to yeast longevity . The yeast homolog of AMPK , Snf1p , phosphorylates Msn2p in response to glucose deprivation , is regulated by TOR , and influences lifespan by modulating ERC formation [48 , 49] . Deletion of SNF4 , an activator of Snf1p , has been shown to extend lifespan , whereas deletion of SIP2 , a repressor of Snf1p , shortens lifespan [49] . Surprisingly , we saw no evidence for involvement of the SNF pathway ( SNF1 , SNF4 , or SIP2 ) in the regulation of PNC1 ( Figure S6 ) , indicating that the activity of Snf1p/AMPK regulates ERC formation independently of PNC1 . Furthermore , CR induced PNC1 expression equally well in wild-type and snf1Δ mutant strains ( Figure S6 ) . We also examined the potential role of ADR1 , a transcription factor that is regulated by SNF1/AMPK and that we suspected from our promoter analysis of PNC1 might regulate PNC1 [50] . We found that ADR1 was not required for the induction of PNC1 by CR ( Figure S6 and data not shown ) , and that overexpression of ADR1 , or expression of a constitutively active form of Adr1p , did not induce expression of PNC1 ( Figure S6 ) . Together , these data show that while attenuation of TOR signaling , PKA activity ( cdc25–10 ) , or glucose metabolism ( hxk2Δ ) extends replicative lifespan , ostensibly by mimicking the effects of CR , the SNF pathway regulates lifespan via a PNC1-independent mechanism .
How CR delays aging and extends the lifespan of various species is poorly understood . In this study , we have connected two sections of the yeast CR pathway , namely the cytoplasmic components ( the glucokinase/cAMP/PKA pathway ) and the nuclear components ( Pnc1p , Sir2p , and ERCs ) . Furthermore , we have shown that TOR signaling , which was previously thought to regulate lifespan independently of sirtuins and ERCs , actually governs the activity of the sirtuins and suppresses rDNA recombination ( Figure 8 ) . This provides additional support to the theory that CR extends replicative lifespan , at least in part , by activating sirtuins . We also demonstrate that the induction of PNC1 in response to numerous stresses is largely controlled by the transcription factors Msn2p and Msn4p . Under conditions of high salt or sorbitol , PNC1 expression is increased in a manner that is completely dependent on MSN2/4 ( Figure 3C ) . While we have linked the increase in Pnc1p levels during heat stress in an msn2Δ/4Δ strain to the transcription factor Hsf1p ( Figure 3F ) , we did not observe a role for Hsf1p in the response to CR or low amino acids . There must be additional factors that control the expression of PNC1 , because an increase in Pnc1p levels still occurs in an msn2Δ/4Δ strain grown in 0 . 5% glucose or in medium with low amino acids . One possibility is that PNC1 is co-activated by Gcr1p , a transcriptional activator with potential binding sites ∼500 base pairs upstream of the PNC1 start codon . GCR1 regulates glycolytic enzyme genes , ribosomal gene synthesis , and trehalose/glycogen metabolism [51 , 52] , making it an interesting candidate for future analysis , although we note that any such analysis is complicated by the severe growth defect of a gcr1Δ strain . In contrast to our study , a previous study utilizing a cdc25–10 mutant as a mimic of CR found that replicative lifespan extension of a PSY316 strain can occur in the absence of MSN2/4 [14] . A recent study has shown that PSY316 may differ substantially from other yeast strains in terms of Sir2p-mediated lifespan extension [18] , and our data may reflect that difference . We favor the notion that while the cdc25–10 mutation mirrors aspects of CR , such as lower PKA activity and increased lifespan , it does not fully replicate it . A previous study has shown that inhibition of TOR signaling can extend lifespan , even in the absence of SIR2 [17] . In agreement with this data , we find that treatment with rapamycin can suppress rDNA recombination and extend lifespan in a sir2Δ fob1Δ strain ( Figure S2 ) . Yeast contain four additional sirtuin genes ( HST1–4 ) , some of which can compensate for the lack of Sir2p during CR [15] . Under the conditions and with the strain used in this study , we have observed that rapamycin no longer lowers rDNA recombination or promotes longevity if both SIR2 and HST2 are deleted ( Figure S2 ) , implying that these two genes are primarily responsible for the effect . However , a W303 sir2Δ hst2Δ fob1Δ strain has a high rate of rDNA recombination and a short lifespan , which may serve to obscure the role of additional sirtuins or other mediators in the response to TOR inhibition . In fact , overexpression of PNC1 in a wild-type strain lowers rDNA recombination more than in a strain lacking MSN2/4 , which may indicate that genes downstream of MSN2/4 besides PNC1 also function to repress rDNA recombination . These alternative pathways may be especially important when glucose concentrations are extremely low [53] and may include pathways that directly regulate rDNA stability , such as RPD3-dependent loading of condensin onto the rDNA array in response to nutrient signaling [54] . TOR signaling also promotes the synthesis of ribosomal proteins , and downregulation of ribosomal biogenesis can extend the lifespan of both yeast [17] and C . elegans [22 , 23] . These data suggest that TOR signaling may act to promote lifespan via multiple pathways that act in parallel to promote longevity ( Figure 8 ) . Our analysis of the responsiveness of STRE-containing genes found ten genes , including PNC1 , that are upregulated more than 2-fold in response to a slight decrease in the glucose concentration ( 2% to 1 . 75% ) ( Table S1 ) . In general , the genes in this category are highly sensitive to environmental stresses , including heat shock and osmotic stress [46] . We speculate that other genes in this category , which includes both metabolic and heat shock genes , may also play a role in lifespan extension . Heat shock proteins in particular have been shown to promote longevity in numerous organisms , and are upregulated during CR in rodents [55 , 56] . Interestingly , MSN2/4 have also been shown to be required for the extension of yeast chronological lifespan [57] . MSN2/4 are responsible for the activation of numerous stress-responsive genes , including the superoxide dismutase SOD2 , a gene that promotes chronological lifespan [34] . Yet , overexpression of SOD2 shortens replicative lifespan , and it has been demonstrated that deletion of MSN2/4 can actually lead to increases in replicative lifespan [58] . Even though we saw no such effect ( Figure 1A ) , perhaps because of a difference in strain background , there may be a reciprocal relationship between replicative and chronological lifespan . A recent study showed that deletion of SIR2 can extend chronological lifespan in several strains [59] , and we have observed that overexpression of SIR2 or HST2 shortens chronological lifespan in W303 ( unpublished data ) . The identification of the stress response factors Msn2p/4p as key components of the CR pathway in yeast supports two theories about CR . The first is known as the hormesis hypothesis of CR , which states that CR is a mild biological stress that provides health benefits because it activates an organism's defenses against adversity [60 , 61] . The second hypothesis is that the promoter elements of key longevity genes are just as important as the longevity genes themselves [28 , 62] . These promoters serve as sensors of the organism's environment by accepting different and additive inputs from environmentally sensitive transcription factors . The existence of short DNA sequences that dictate longevity could explain how new lifespans evolve so rapidly in response to a new ecological niche . Theoretically , if a transcription factor binding site regulates a key longevity gene , then a single base change might be sufficient to alter how long a species lives in response to an environmental condition . In contrast to previous suggestions , we find that TOR and sirtuin signaling are components of the same longevity pathway that extends yeast replicative lifespan by stabilizing the repetitive rDNA ( Figure 8 ) . Given the high degree of functional conservation of TOR and sirtuins between yeast and higher organisms , and the recent discovery of a role for mammalian sirtuins in DNA repair [63] , the findings in this study raise the possibility that the mammalian TOR pathway influences sirtuin activity and that together they may promote the health and longevity of mammals .
W303AR MATa , W303AR MATa pnc1::kanr , W303AR PNC1-GFP::kanr , and W303AR SIR2-3xHA::kanr have been previously described [24 , 29 , 64] . Gene disruptions in W303AR MATa were achieved by replacing the wild-type genes with the kanr , hphr , or natr marker as described [65 , 66] and verified by PCR using oligonucleotides flanking the genes . PNC1 was overexpressed as previously described [29] . W303AR cdc25–10 was created by replacing the endogenous copy of CDC25 with a plasmid-borne copy of cdc25–10 ( the kind gift of S . J . Lin ) as previously described [37] . pPNC1-STRE . 4 , pPNC1-STRE . 2 , and pPNC1-STRE . 0 constructs and the pAdh-Msn2p-GFP/HA constructs were kindly provided by M . Ghislain [33] and C . Schuller [36] , respectively . Msn4p-GFP and Msn2p ( S4A ) -GFP constructs were the kind gift of M . Jacquet . Plasmids for expression of Adr1p were obtained from E . Young . BY4741 deletions in this background were from F . Winston and P . Silver ( Harvard ) , BY4742 and BY4742 hxk2::kanr were from B . Kennedy [18] , W303 msn2Δ/4Δ and W303 msn2Δ/4Δ tetO-HSF1 were from H . Nelson [38] . All primer sequences , strains , and plasmid maps are available upon request . Yeast were grown in yeast peptone dextrose ( YPD ) medium supplemented with an additional 0 . 015% w/v adenine , histidine , leucine , tryptophan , and uridine , and containing 2% w/v glucose during normal growth and 0 . 5% glucose for CR unless otherwise stated . For growth in low amino acid medium , synthetic complete medium containing 0 . 03 % w/v essential amino acids and 2% glucose was used . Strains were pre-grown overnight at 30 °C . The following day , cells were inoculated at an O . D . 600 = 0 . 1 and grown until log phase of growth was attained during the various conditions mentioned ( O . D . 600 = 0 . 7 ) . For treatment with rapamycin or heat shock , cells were grown for 2 h untreated , at which point rapamycin was added to a final concentration of 1 nM , or cells were moved to 37 °C , and cells were then grown for an additional 2 h . rDNA recombination rates were determined by determining the frequency of loss of ADE2 in the rDNA of strain W303AR as previously described [15 , 29] . For rapamycin recombination assays , cells were grown for 2 h without rapamycin , followed by growth with 1 nM rapamycin for 2 h . More than 6 , 000 colonies were examined for each strain . Results are average values and standard deviation of at least three experiments . For replicative lifespan analyses , strains were pre-grown overnight on YPD plates unless otherwise noted . All lifespan analyses were carried out by using micromanipulation as previously described [14] , and all micromanipulation dissections , including for cells grown under heat stress ( 37 °C ) , were carried out at laboratory temperature . For cells treated with rapamycin , yeast that growth-arrested in the G1 phase of the cell cycle due to toxicity [67] within the first nine divisions were excluded from the datasets . Statistical analysis was carried out using the JMP-IN statistics package ( SAS , http://www . sas . com/ ) . Wilcoxon rank-sum test p-values were calculated for each pair of lifespans , as shown in Table S3 . For the observation of nuclear migration of Msn2p-GFP or Msn4p-GFP , yeast were grown in YPD for 30 min at 1% , 0 . 5% , 0 . 1% , and 0 . 05% glucose ( w/v ) , after first pre-growing to log phase in 2% glucose . Nuclei in live cells were stained with Hoechst #33342 ( Sigma-Aldrich , http://www . sigmaaldrich . com/ ) . Time-lapse photomicrographs were captured every 30 s using 1-s exposures . Image analysis was performed using the imageJ software package ( National Institutes of Health , http://www . nih . gov/ ) in order to calculate the ratios of average nuclear intensity versus average cytoplasmic intensity . Whole cell extracts were used to assay β-galactosidase activity as described [33] . Enzymatic activity is expressed as nanomoles o-nitrophenol-β-D-galactopyranoside cleaved per minute per milligram total protein . Rabbit anti-Pnc1p polyclonal antibodies were generated by immunization of rabbits ( Covance , http://store . crpinc . com/ ) with recombinant protein , and fresh serum was used at a dilution of 1:5 , 000 . Mouse monoclonal anti-β-tubulin antibody ( MAB3408 , clone KMX-1 , Upstate , http://www . upstate . com/ ) , mouse monoclonal anti-actin antibody ( Upstate/Chemicon MAB1501 ) and polyclonal rabbit anti-HA antibody ( Abcam , http://www . abcam . com/ ) were used at a dilution of 1:1 , 000 . Anti-rabbit ( Amersham , http://www . amersham . com/ ) and anti-mouse ( Amersham ) horse radish peroxidase–conjugated antibodies were used at dilutions of 1:7 , 000 . The chromatin immunoprecipitation procedure was a modification of the method described by Strahl-Bolsinger et al . [68] . Changes to the protocol are as follows: 100 ml of cells ( 2 . 0 × 107 cells/ml ) was cross-linked with 2% formaldehyde for 15 min at room temperature . Glycine was added to a final concentration of 250 mM , and the incubation continued for an additional 5 min . The suspension was sonicated seven times for 10 s with the amplitude set at 30% using a Branson model 450 digital sonifier ( Branson , http://www . bransonultrasonics . com/ ) . The suspension was clarified by centrifugation for 5 min , maximum setting , at 4 °C in a microcentrifuge . Samples were incubated on ice for 2 min between pulses . Then 1 μl of RNase ( 10 μg/ul ) was added to samples , and they incubated for 30 min at 37 °C . Afterwards , sheared chromatin was purified using QIAquick spin columns ( Qiagen , http://www1 . qiagen . com/ ) . Then 250 μl of supernatant was incubated with 15 μl of rabbit anti-HA antibody ( Abcam ) . For the PCR analysis , the actin control primers used were as follows: ACT1-Chip-Fwd , GCCTTCTACGTTTCCATCCA [69] , and ACT1-Chip-Rev , GGCCAAATCGATTCTC AAAA [69] . The PNC1 promoter primers used were as follows: Pnc1p-Chip-Fwd , GATCAAGGTGGCACACAGGG , and Pnc1p-Chip-Rev , ATACATAGTGGGCCAAACGG . The PCR protocol used was one cycle with 2 min at 95 °C , 30 s annealing at 55 °C , and a 1-min extension at 72 °C , followed by 30 cycles with 30 s at 95 °C , 30 s annealing at 55 °C , and 1-min extension at 72 °C . A final extension was performed for 4 min at 72 °C . Specific binding of Msn2p-HA to the endogenous PNC1 promoter ( PNC1p ) was analyzed by calculating the ratio of the percent IP of PNC1p to the percent IP of ACT1 , using the V4 . 2 . 2 Quantity One 1-D analysis package ( Bio-Rad , http://www . bio-rad . com/ ) . | There are only a few techniques that reliably promote longevity in multiple , distantly related species . Perhaps the best known , caloric restriction ( CR ) , was first shown to promote lifespan in rodents in the 1930s and has since been shown to work in most species it has been tested on . We and others have previously proposed that CR extends lifespan in budding yeast by boosting the activity of sirtuin deacetylases , which work to extend lifespan by suppressing genomic instability . A competing theory is that CR works by suppressing the TOR ( target of rapamycin ) signaling pathway , which has recently been discovered to extend the lifespan of yeast and worms , but the downstream players are not yet known . We show that TOR inhibition and sirtuins are part of the same CR pathway that extends yeast lifespan by stabilizing the genome . CR and TOR inhibition promote longevity by relocalizing two transcription factors , Msn2p and Msn4p , from the cytoplasm to the nucleus , where they increase expression of the nicotinamidase gene PNC1 , a regulator of sirtuin activity . We propose that TOR signaling and sirtuins may also be part of the same CR pathway in mammals . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology",
"microbiology",
"molecular",
"biology",
"genetics",
"and",
"genomics",
"saccharomyces"
] | 2007 | MSN2 and MSN4 Link Calorie Restriction and TOR to Sirtuin-Mediated Lifespan Extension in Saccharomyces cerevisiae |
AGGF1 is an angiogenic factor with therapeutic potential to treat coronary artery disease ( CAD ) and myocardial infarction ( MI ) . However , the underlying mechanism for AGGF1-mediated therapeutic angiogenesis is unknown . Here , we show for the first time that AGGF1 activates autophagy , a housekeeping catabolic cellular process , in endothelial cells ( ECs ) , HL1 , H9C2 , and vascular smooth muscle cells . Studies with Atg5 small interfering RNA ( siRNA ) and the autophagy inhibitors bafilomycin A1 ( Baf ) and chloroquine demonstrate that autophagy is required for AGGF1-mediated EC proliferation , migration , capillary tube formation , and aortic ring-based angiogenesis . Aggf1+/- knockout ( KO ) mice show reduced autophagy , which was associated with inhibition of angiogenesis , larger infarct areas , and contractile dysfunction after MI . Protein therapy with AGGF1 leads to robust recovery of myocardial function and contraction with increased survival , increased ejection fraction , reduction of infarct areas , and inhibition of cardiac apoptosis and fibrosis by promoting therapeutic angiogenesis in mice with MI . Inhibition of autophagy in mice by bafilomycin A1 or in Becn1+/- and Atg5 KO mice eliminates AGGF1-mediated angiogenesis and therapeutic actions , indicating that autophagy acts upstream of and is essential for angiogenesis . Mechanistically , AGGF1 initiates autophagy by activating JNK , which leads to activation of Vps34 lipid kinase and the assembly of Becn1-Vps34-Atg14 complex involved in the initiation of autophagy . Our data demonstrate that ( 1 ) autophagy is essential for effective therapeutic angiogenesis to treat CAD and MI; ( 2 ) AGGF1 is critical to induction of autophagy; and ( 3 ) AGGF1 is a novel agent for treatment of CAD and MI . Our data suggest that maintaining or increasing autophagy is a highly innovative strategy to robustly boost the efficacy of therapeutic angiogenesis .
AGGF1 is an Angiogenic factor with a G-patch domain and a Forkhead-associated ( FHA ) domain . AGGF1 was initially identified by our laboratory through positional cloning analysis for a gene involved in development of Klippel–Trénaunay syndrome ( KTS ) , a congenital vascular disorder [1] . AGGF1 can induce angiogenesis and excessive angiogenesis , and increased AGGF1 expression is a cause of KTS [1–5] . We and others have also found that AGGF1 is critical to specification of veins [6 , 7] , specification of multipotent hemangioblasts [8] , and anti-inflammation [9] . However , the molecular mechanisms underlying these processes remain to be fully defined . Coronary artery disease ( CAD ) and its most severe manifestation , myocardial infarction ( MI ) , are the most common causes of death worldwide . Therapeutic angiogenesis has been proposed as an attractive new strategy to treat CAD and MI patients . Therapeutic angiogenesis can be defined as the utilization of angiogenic growth factors to promote neovascularization and growth of collateral blood vessels , which act as endogenous bypass conduits to improve blood flow and increase tissue perfusion in the ischemic extremity . However , there is currently no United States Food and Drug Administration ( FDA ) -approved therapeutic angiogenesis to treat CAD , MI , or other ischemic diseases [10 , 11] . Many challenges must be overcome before therapeutic angiogenesis becomes an applied patient therapy , including the critical identification of the most robust , effective angiogenic factor [10 , 11] . Importantly , lack of understanding of the fundamental molecular mechanisms underlying therapeutic angiogenesis has slowed advances in this field . Autophagy is an evolutionarily conserved dynamic catabolic process that removes damaged , dysfunctional organelles and long-lived protein aggregates [12] . It recycles amino acids and other substrates for protein synthesis and ATP generation [12] . However , excessive autophagy can also lead to cell death . Autophagy is initiated by the formation of the phagophore; this process is mediated by the class III PI3-K complex consisting of Vps34 , Vps15 , Atg14 , and beclin 1 [12] . The phagophore then elongates and engulfs cytoplasmic materials targeted for degradation , leading to the formation of autophagosome . During this process , the microtubule-associated protein 1 light chain 3 ( LC3 ) is cleaved and converted to LC3-I . LC3-1 , the soluble form of LC3 , is then activated and converted to LC3-II , which is the autophagic vesicle-associated form of LC3 . The autophagosome is fused with the lysosome to form the autolysosome , which leads to degradation of the vesicle content by lysosomal hydrolases , recycling of the degraded products into amino acids and lipids , and generation of ATP [12 , 13] . Here , we demonstrate that AGGF1 induces autophagy in endothelial cells ( ECs ) and all other cells analyzed as well as in mice with acute MI using a series of integrative in vitro and in vivo approaches . We show that an angiogenic factor can induce autophagy and that autophagy acts upstream of angiogenesis and is essential for therapeutic angiogenesis . We also demonstrate the robust potential of recombinant AGGF1 and AGGF1-mediated autophagy in therapeutic angiogenesis to treat acute MI .
Because AGGF1 is an angiogenic factor , we focused our studies of AGGF1 on ECs . To identify the role of AGGF1 in regulating autophagy , we treated human umbilical vein endothelial cells ( HUVECs ) with different doses of AGGF1 and then analyzed expression levels of LC3-II and p62 . AGGF1 strongly activated autophagy and the effect was concentration-dependent and saturable ( Fig 1A ) . As a positive control , serum starvation induced the activation of autophagy ( S1 Fig ) . To further demonstrate that AGGF1 induces autophagy , HUVECs were infected with a viral expression construct for Green Fluorescent Protein ( GFP ) -LC3 for 24 h and then treated with AGGF1 or control Immunoglobin G ( IgG ) . HUVECs with GFP punctuate ( autophagic cells with autophagic vesicle-bound LC3-II ) and overall GFP-staining were counted ( Fig 1B ) . The ratio of HUVECs with GFP punctuate was 31% ± 4% of the total GFP-positive cells with AGGF1 treatment but only 10% ± 2% for HUVECs incubated with IgG ( Fig 1B ) . Electronic microscopy directly demonstrated that AGGF1 induced formation of autophagosome ( Fig 1C ) . These data further support the finding that AGGF1 induces autophagy in ECs . To determine whether AGGF1-activated autophagy is specific to ECs , we analyzed several other types of cells . As shown in S2 Fig , AGGF1 activated autophagy in all cells examined . These data suggest that AGGF1 is a universal master regulator of autophagy . We created Aggf1 knockout mice with a gene-trapping allele in intron 11 . Homozygous Aggf1-/- mice are embryonically lethal , but heterozygous Aggf1+/- mice showed partial embryonic lethality , in which 60% of them can survive to adulthood . To further demonstrate the critical role of AGGF1 in autophagy , we isolated heart tissue samples from Aggf1+/- mice and performed western blot analysis for autophagy markers LC3-II and p62 . Compared to wild-type mice , Aggf1+/- mice showed a significantly decreased expression level of LC3-II and a significantly increased expression level of p62 ( Fig 1D ) . Moreover , cardiac ECs from wild-type or Aggf1+/- KO mice showed reduced activation of autophagy compared with wild-type ECs ( Fig 1E ) . These data indicate that reduced expression of AGGF1 by 50% inhibits autophagy , suggesting that AGGF1 is critically involved in autophagy . Western blot analysis was performed with heart tissue lysates from 8-wk-old wild-type and Aggf1+/- KO mice with apoptosis markers . Aggf1+/- KO mice showed an increased level of poly ( ADP-ribose ) polymerase ( PARP ) , caspase 3 , or Bax , but a reduced level of Bcl-2 compared with wild-type mice ( S3 Fig ) . No HIF1α was detected in either wild-type or Aggf1+/- KO mice ( S3 Fig ) . These data suggest that haploinsufficiency of Aggf1 induced apoptosis of cardiomyocytes . To distinguish the cause–effect relationship between AGGF1-promoted autophagy and AGGF1-induced angiogenesis , we performed a series of studies with two inhibitors of autophagy , bafilomycin A1 ( Baf ) and chloroquine ( CQ ) , and DMSO control . The effectiveness of Baf and CQ on inhibition of autophagy was verified by western blot analysis ( S4 Fig ) . Recombinant AGGF1 promoted HUVEC proliferation ( compare DMSO+IgG and DMSO+AGGF1 , Fig 2A ) . However , inhibition of autophagy by Baf or CQ completely blocked AGGF1-induced HUVEC proliferation ( Fig 2A ) . These data suggest that autophagy is required for AGGF1-mediated EC proliferation . EC migration is also one of the key processes involved in angiogenesis . Wound-healing scratch assays for migration showed that AGGF1 treatment significantly induced HUVEC migration , but this effect was blocked by Baf or CQ ( Fig 2B ) . Similarly , with Transwell migration assays , AGGF1 in the bottom chamber induced migration of 1 . 64-fold more HUVECs to the bottom of Transwells than IgG . However , this effect was blocked by Baf or CQ ( Fig 2C ) . These data suggest that autophagy is required for AGGF1-mediated EC migration . We also assessed the effect of autophagy on capillary tube formation mediated by AGGF1 using a matrigel-based angiogenesis assay . AGGF1 treatment increased the number of endothelial cell tubes , but this effect was blocked by Baf or CQ ( Fig 2D ) . Moreover , using an ex vivo endothelial cell sprout assay from aortic rings , it is obvious that sprouts emerged from the aorta rings and grew outward after 4 d in culture with AGGF1 ( 500 ng/ml ) ( compare AGGF1 and IgG in the DMSO group , Fig 2E ) . However , treatments with Baf or CQ resulted in a significant decrease in sprout length and density by AGGF1 as compared to IgG ( Fig 2E ) . These data suggest that autophagy is required for AGGF1-mediated angiogenesis . Similar to Baf or CQ , inhibition of autophagy by knockdown of Atg5 expression by siRNA ( S5 Fig ) also blocked EC proliferation ( Fig 3A ) , wound-healing migration ( Fig 3B ) , Transwell migration ( Fig 3C ) , capillary tube formation ( Fig 3D ) , and sprout angiogenesis from aortic rings ( Fig 3E ) . We analyzed the expression level of AGGF1 in response to acute MI in a mouse model with ligation of the left anterior descending ( LAD ) coronary artery . The expression levels of both the AGGF1 mRNA and AGGF1 protein in heart tissue increased significantly in mice with MI compared to mice with sham operation ( p < 0 . 05 ) ( Fig 4A and 4B ) . Immunostaining with an anti-AGGF1 antibody showed that AGGF1 expression increased in the border zone and the infarct area of the left ventricle in mice with MI compared to the non-infarct area or similar areas in mice with sham operation ( Fig 4C ) . To determine the mechanism by which MI induces AGGF1 expression , we hypothesized that hypoxia after MI may be critical . To test the hypothesis , we examined the expression level of the AGGF1 protein in HUVECs in response to hypoxia . AGGF1 expression increased in HUVECs challenged with 1% oxygen for 30 min up to 8 h compared to cells cultured under a normal level of oxygen ( S6A Fig ) . Similar findings were made for the AGGF1 mRNA ( S6B Fig ) . Western blot analysis showed that AGGF1 expression was increased in response to hypoxia in mouse cardiac ECs ( S7 Fig ) . Together , these data suggest that hypoxia induces expression of AGGF1 , providing a mechanism for ischemia-induced AGGF1 up-regulation in MI mice . In addition , the activity of AGGF1-promoted autophagy , i . e . , the expression level of LC3-II or decreased expression of p62 , increased as the culture time of HUVECs under hypoxia increased ( Fig 4D ) . This may be related in part to the earlier finding that hypoxia induces increased AGGF1 expression ( S6 Fig ) . When AGGF1 expression was knocked down in HUVECs with transfection of siRNA , hypoxia-induced autophagy was inhibited ( Fig 4E ) . The data suggest that AGGF1 is critical to induction of autophagy under hypoxia . To explore the functional effects of up-regulation of AGGF1 expression in response to MI , we tested the purified recombinant AGGF1 protein as a therapy for acute MI . The survival rate of MI mice with AGGF1 protein therapy was 80% 2 wk after treatment , a significant increase over the 60% rate for the group with IgG treatment ( p < 0 . 01 ) ( Fig 4F ) . Similarly , 4 wk after treatment , AGGF1 remarkably improved the post-MI survival rate to 75% from 53% for treatment with IgG ( p < 0 . 01 ) ( Fig 4F ) . Echocardiography showed that AGGF1 dramatically improved cardiac functions after acute MI ( Fig 4G ) . Left ventricular ejection fraction ( LVEF ) dramatically improved with AGGF1 treatment after MI . The significant improvement was obvious even at 14 d after AGGF1 treatment with LVEF of 50 . 29% ± 8 . 47% compared with 23 . 69% ± 7 . 49% with treatment of control IgG ( p < 0 . 01; S8 Fig ) . At 28 d ( end of study ) , LVEF more than doubled with AGGF1 treatment compared to control IgG treatment ( 55 . 57% ± 9 . 65% versus 22 . 77% ± 7 . 79% , p < 0 . 01 ) and reached almost the normal range of LVEF for mice ( Fig 4H ) . Similar dramatic improvement was observed for LV fraction shortening ( LVFS ) with AGGF1 treatment compared to IgG treatment ( 29 . 42% ± 6 . 37% versus 14 . 35% ± 2 . 53% at 28 d after treatment , p < 0 . 01 ) ( Fig 4H ) . Moreover , significant decreases were found with AGGF1 treatment compared to IgG treatment for the LV end-diastolic diameter ( LVEDD ) ( 4 . 17 ± 0 . 42 mm versus 4 . 89 ± 0 . 41 mm , p < 0 . 01 ) and LV end-systolic diameter ( LVESD ) ( 3 . 26 ± 0 . 58 mm versus 3 . 92 ± 0 . 57 mm , p < 0 . 01 ) ( Fig 4H ) , indicating improvement of cardiac structure with AGGF1 protein therapy . The new speckle tracking function of Vevo-2100 echocardiography can be used to quantify global and regional ( e . g . , infarct region ) cardiac functions and estimate the size of infarct [14 , 15] . The peak longitudinal strain and strain ratios serve as markers for myocardial contractile states at earlier time points after MI and predict the later development of adverse LV remodeling ( S9A Fig ) . Four weeks after treatment , the longitudinal strain of the global heart in the MI group with AGGF1 treatment was -8 . 03% ± 2 . 38% , a dramatic increase compared to that of IgG treatment ( -2 . 85% ± 0 . 86% , p < 0 . 01 ) ( S9B Fig ) . Identical results were observed for the longitudinal strain ratio of the global heart ( -5 . 14% ± 1 . 56% for AGGF1 treatment versus -2 . 72% ± 0 . 74% for IgG treatment , p < 0 . 01 ) ( S9C Fig and S1 Table ) . Myocardial contractile function in the regional infarct areas was also assessed . The longitudinal strain and longitudinal strain ratio in the infarct areas significantly improved with AGGF1 treatment compared to control IgG treatment ( longitudinal strain: -4 . 39% ± 2 . 68% versus -1 . 625% ± 0 . 63% , p < 0 . 01; longitudinal strain ratio: -3 . 52% ± 1 . 54% versus -1 . 22% ± 0 . 11% , p < 0 . 01 ) ( S9C Fig ) . These data suggest that AGGF1 dramatically improves the global cardiac function as well as myocardial contractile function in the regional infarct areas . AGGF1 protein therapy inhibited cardiac hypertrophy after MI . At 28 d after treatment , AGGF1 treatment inhibited early cardiac hypertrophy associated with MI . As shown in S10 Fig , MI induced an increase of the ratio of heart weight ( HW ) to tibia length ( TL ) or body weight and the ratio of lung weight ( LW ) to TL or body weight , indicative of cardiac hypertrophy . However , AGGF1 protein therapy significantly attenuated MI-induced hypertrophy of the heart ( S10 Fig ) . To further confirm whether AGGF1 protein therapy reduces mortality and dramatically improves cardiac function after MI through autophagy , we determined the functional role of AGGF1 in autophagy in the infarcted heart . We found that , compared to sham-operated mice , mice with acute MI showed an increased expression level of LC3-II over control β-actin or LC3-I or decreased expression of p62 ( Fig 5A ) . Surprisingly , the AGGF1-treated group showed a dramatic increase of LC3-II expression or decreased expression of p62 compared with the IgG-treated group ( Fig 5A ) . These data indicate that AGGF1 induces autophagy in mice after MI in vivo . To elucidate the mechanism by which AGGF1 protein therapy drastically improves myocardial function and contraction after acute MI , we examined whether AGGF1 promoted therapeutic angiogenesis . Immunostaining with an anti-CD31 antibody , an endothelial cell marker , was performed with cryo-sections from left ventricles . MI mice treated with AGGF1 showed much more CD31-positive capillaries than MI mice treated with control IgG ( 543 ± 143 versus 272 ± 72 capillaries/high-power field [HPF] , p < 0 . 01 ) or sham mice ( Fig 5B ) . We performed α-SMA staining for heart sections and showed that AGGF1 treatment increased the density of α-SMA-positive vessels ( S11A Fig ) . These data suggest that AGGF1 promotes therapeutic angiogenesis after MI , resulting in improved myocardial contractile function . Cardiac apoptosis increases in response to MI and is responsible for myocardial fibrosis after MI . We assessed the cyto-protective activity of AGGF1 after MI by TUNEL staining of left ventricles . Administration of the recombinant AGGF1 protein enhanced myocardial cell survival by significantly decreasing the number of TUNEL-positive cells after MI ( p < 0 . 01 ) ( Fig 5C and S12 Fig ) . Twenty-eight days after AGGF1 treatment , myocardial tissue samples were collected and activity of caspase-3 was measured with myocardial lysates . AGGF1 protein reduced caspase-3 activity by 44 . 79% ( p < 0 . 01 ) ( Fig 5C ) . Cardiac apoptosis is a major cause of cardiac fibrosis . H&E staining with Masson trichrome revealed that MI increased cardiac fibrosis in the left ventricle and reduced LV wall thickness ( top panel , S13 Fig ) . AGGF1 protein therapy decreased MI-induced cardiac fibrosis and increased LV wall thickness ( top panel , S13 Fig ) . The size of cardiac fibrotic areas reflecting the size of the infarct areas was increased by MI , but AGGF1 treatment dramatically reduced the infarct size ( 22 . 54% ± 1 . 25% for AGGF1 treatment versus 50 . 94% ± 7 . 74% for IgG treatment , p < 0 . 01 ) ( bottom panel , S13 Fig ) . Moreover , AGGF1 treatment increased LV wall thickness compared to IgG treatment ( 510 . 47 ± 96 . 27 μm versus 266 . 89 ± 57 . 38 μm , p < 0 . 01 ) ( bottom panel , S13 Fig ) . H&E staining showed that MI induced necrosis , but AGGF1 administration reduced the MI-induced necrosis ( S11B Fig ) . These data are consistent with the echocardiographic data . To further confirm that AGGF1 inhibited myocardial apoptosis after MI , western blot analysis with myocardial lysates was performed for cleaved PARP , Bax , and Bcl-2 . MI increased the levels of cleaved PARP and Bax , indicating activation of Bax and PARP , but their activation was inhibited by AGGF1 protein therapy ( Fig 5D ) . Consistent with the protective role of AGGF1 , expression of anti-apoptotic protein Bcl-2 was increased by AGGF1 treatment ( Fig 5D ) . Real-time RT-PCR analysis showed that the expression level of Bcl-2 mRNA in myocardial tissue also increased in AGGF1-treated mice ( data not shown ) . The ratio of Bax/Bcl-2 , an indicator for apoptosis , decreased during AGGF1 protein therapy ( Fig 5D ) . To confirm whether AGGF1 could promote autophagy , leading to increased survival after MI , we created an MI model using Aggf1+/- mice . The survival rate of Aggf1+/- mice decreased 4 wk after MI compared with wild-type mice ( Fig 6A ) . Echocardiography showed that LVEF and LVFS were significantly lower in Aggf1+/- mice than in wild-type mice with or without MI ( Fig 6B ) , indicating a compromised myocardial contraction by AGGF1 haploinsufficiency . Compared with wild-type mice , LVEDD and LVESD decreased in Aggf1+/- mice with sham operation but increased in mice with LAD ligation ( Fig 6B ) . The infarct area ( fibrotic size ) after MI was larger in Aggf1+/- mice than in wild-type mice ( Fig 6C ) . The CD31-positive vessel density was significantly less in Aggf1+/- mice than in wild-type mice with or without MI ( Fig 6D ) , indicating a critical role of AGGF1 in angiogenesis in vivo . To identify the mechanism for the reduced angiogenesis , reduced survival , and compromised cardiac function and contraction in Aggf1+/- mice , we analyzed autophagy in these mice . After MI , the expression level of LC3-II was reduced , whereas p62 expression was increased in Aggf1+/- mice compared to wild-type mice ( Fig 6E ) , indicating a correlation between reduced angiogenesis and inhibition of autophagy in Aggf1+/- mice . Because autophagy is required for AGGF1-promoted EC proliferation , migration , and capillary tube formation as demonstrated above , we hypothesized that autophagy is essential for AGGF1-mediated angiogenesis and recovery of myocardial function in mice after MI . To test this hypothesis , we carried out a study with bafilomycin A1 in mice . Immunostaining with CD31 showed that AGGF1 protein therapy increased CD31-positive vessel density compared with IgG treatment ( compare AGGF1 and IgG in the vehicle group , p < 0 . 01 ) , but bafilomycin A1 treatment eliminated the angiogenic effect of AGGF1 treatment ( compare AGGF1 and IgG in the Baf group , p > 0 . 05 ) ( Fig 7A ) . These data indicate that autophagy is required for angiogenesis in vivo . AGGF1 protein therapy increased the survival rate of MI mice compared to IgG treatment in the vehicle group ( Fig 7B ) . However , preconditioning by daily intraperitoneal injection of bafilomycin A1 ( 0 . 3 mg/kg ) for five consecutive days after MI and before AGGF1 treatment eliminated the pro-survival function of AGGF1 ( compare AGGF1 and IgG in the Baf group , Fig 7A ) . Bafilomycin A1 eliminated the therapeutic effects of AGGF1 on recovery of LVEF , LVFS , LVEDD , and LVESD ( Fig 7C and S14 Fig ) . H&E staining with Masson trichrome showed that AGGF1 treatment reduced the infarct size and increased LV wall thickness in the infarct areas compared to IgG treatment ( the vehicle group , Fig 7D ) , but bafilomycin A1 treatment blocked the therapeutic effects of AGGF1 on cardiac fibrosis ( the Baf group , Fig 7D ) . To further test the hypothesis that autophagy is essential for AGGF1-mediated angiogenesis and recovery of myocardial function after MI , we assessed the effects of AGGF1 protein in an autophagy-compromised background in mice . Beclin 1 , a mammalian homology of yeast Atg6/Vps30 , is an initiator of autophagy and is required for autophagosome formation [16] . Heterozygous Becn1+/- mice showed defective autophagy [17] . Immunostaining with CD31 showed that AGGF1 did not have any effect on angiogenesis ( CD31-positive vessel density ) in Becn1+/- mice either with sham operation or after MI surgery ( Fig 7E ) . These data indicate that without autophagy , the angiogenic activity of AGGF1 is completely lost , suggesting that autophagy is essential for angiogenesis . As shown in Fig 7F , AGGF1 protein therapy did not affect the survival rate of MI in Becn1+/- mice compared to IgG treatment . MI mice with IgG treatment ( n = 44 ) demonstrated a 79 . 54% 2-wk survival rate , whereas MI mice with AGGF1 treatment ( n = 43 ) had an 83 . 72% 2-wk survival rate ( p = 0 . 34 ) ( Fig 7F ) . At 4 wk , there remained no significant difference between the survival rates of Becn1+/- MI mice with AGGF1 treatment ( 39 . 53% ) and with IgG treatment ( 36 . 36% ) ( p = 0 . 28 ) . These data suggest that the effectiveness of AGGF1 protein therapy on MI survival requires beclin 1 and autophagy . Echocardiography showed that compared to IgG treatment , AGGF1 treatment did not improve LVEF , LVFS , LVEDD , LVESD , cardiac fibrosis , infarct sizes , and LV wall thickness in Becn1+/- mice ( Fig 7G and 7H ) . Altogether , these data indicate that beclin 1 and autophagy are essential for therapeutic recovery of myocardial functions after MI with AGGF1 protein therapy . To further validate the critical role of autophagy in AGGF1-mediated angiogenesis and cardiac repair , we performed AGGF1 therapy for MI in mice lacking Atg5 expression in the myocardium . Atg5 forms a complex with Atg12 and Atg16L1 , which can act as an E3 ligase to facilitate conjugation of LC3 to phosphaidylethanolamine in the autophagic membrane . We injected adenoviruses with AAV9-CMV-Cre into the myocardium of Atg5flox/flox mice to generate Atg5 KO mice ( AAV9-GFP as control ) . Western blot analysis with cardiac extracts showed that Atg5 expression is absent and the level of autophagy was low in Atg5 KO mice 2 wk after virus injection compared with AAV-GFP mice ( S15 Fig ) . MI was created in Atg5 KO mice and control mice . In AAV9-GFP control mice , AGGF1 protein therapy significantly increased the density of CD31-positive vessels in the heart after MI compared with IgG treatment ( p < 0 . 01 ) ( Fig 8A ) . In AAV9-CMV-Cre Atg5 KO mice with deficient autophagy , AGGF1 protein therapy failed to increase the density of CD31-positive vessels in the heart after MI compared with IgG treatment ( p > 0 . 05 ) ( Fig 8A ) . These results further demonstrate that autophagy is required for angiogenesis in vivo . In AAV9-GFP control mice , AGGF1 treatment increased the survival rate of MI mice compared to IgG treatment ( Fig 8B ) . However , the therapeutic effect of AGGF1 was completely lost in AAV9-CMV-Cre Atg5 KO mice with deficient autophagy ( Fig 8B ) . Echocardiography showed that AGGF1 treatment significantly increased LVEF and LVFS , reduced LVEDD , LVESD , cardiac fibrosis , and infarct sizes , and increased LV wall thickness in AAV9-GFP control mice with MI compared with IgG treatment ( Fig 8C and 8D ) , but the therapeutic effect of AGGF1 was completely lost in AAV9-CMV-Cre Atg5 KO mice with deficient autophagy ( Fig 8C and 8D ) . Together , these data indicate that autophagy is essential for therapeutic recovery of myocardial functions after MI with AGGF1 protein therapy . To identify the molecular signaling pathway for autophagy induced by AGGF1 , we analyzed the activation/phosphorylation of c-Jun N-terminal protein kinase ( JNK ) . AGGF1 markedly induced JNK activation in HUVECs in a dose-dependent manner , with peak effect at a dose of 500 ng/ml ( Fig 9A ) . Moreover , AGGF1 induced phosphorylation of JNK in a time-dependent manner ( Fig 9B ) . Interestingly , as time increased , AGGF1-induced JNK phosphorylation increased ( Fig 9B ) . Two different JNK inhibitors , JNK inhibitor II and JNK inhibitor III , blocked JNK activation by AGGF1 . Consistent with this finding , AGGF1 increased the rate of autophagy as indicated by the increased level of LC3-II; however , the two JNK inhibitors inhibited autophagy induced by AGGF1 ( Fig 9C ) . The Becn1-Vps34-Atg14L complex is involved in autophagy [12] . Since beclin 1 is required for autophagy by AGGF1 as shown in Becn1+/- mice , we analyzed the effect of AGGF1 on the assembly of the Becn-1-Vps34-Atg14L complex . Co-immunoprecipitation assays showed that AGGF1 increased the complex formation between beclin 1 and Vps34 and between beclin 1 and Atg14L ( Fig 9D ) . JNK inhibitor II reduced the complex formation among beclin 1 , Vps34 , and Atg14L ( Fig 9D ) . Vps34 is a class III PI3K and the activity ( leading to PtdIns3P production ) is required for autophagosome formation [18] . Thus , we measured the activity of Vps34 lipid kinase . Compared with IgG , AGGF1 significantly increased the Vps34 lipid kinase activity as measured for conversion of PtdIns to PtdIns ( 3 ) P by an ELISA assay ( Fig 9E ) . In the presence of JNK inhibitor II , the increased Vps34 PI3K activity by AGGF1 disappeared ( Fig 9E ) . Moreover , in an assay for Vps34 PI3K-dependent intracellular localization of a GFP-FYVE reporter protein , predominantly localized to membranes of endocytic compartments ( the FYVE domain preferably binds to Ptdlns ( 3 ) P to other phosphoinositides ) , AGGF1 treatment increased the number of GFP-2-FYVE dots , an indication of increased Vsp34 PI3K activity ( Fig 9F ) . However , JNK inhibitor II significantly decreased the number of AGGF1-induced green GFP-2-FYVE dots ( Fig 9F ) . The last step in autophagy is the degradation of autophagic cargos in the lysosomes . Autophagic cargos are mostly long-lived proteins , and their rate of degradation , an indicator for autophagic flux , can be measured [18] . AGGF1 significantly increased the degradation of long-lived proteins in HUVECs , and this effect was blocked with JNK inhibitor II ( Fig 9G ) . To further confirm the involvement of the JNK pathway in AGGF1-mediated autophagy , we used siRNA to knock expression of JNK1 down and determine the effect of AGGF1 on activation of autophagy . AGGF1 failed to increase the LC3-II/LC3-I ratio ( activation of autophagy ) when JNK1 expression was knocked down ( Fig 9H ) . JNK1 knockdown also reduced the AGGF1-induced formation of the beclin1-Atg14L-VPS34 complex ( Fig 9I ) . Combined with the data obtained on JNK inhibitors , these data demonstrate the involvement of the JNK pathway in AGGF-mediated autophagy . Because Atg14L interacts with VPS34 , we used siRNA to knock expression of Atg14L down and then determined the effect of AGGF1 on activation of autophagy . AGGF1 failed to increase the LC3-II/LC3-I ratio ( activation of autophagy ) when Atg14L expression was knocked down ( Fig 9J ) . Atg14L knockdown also reduced the AGGF1-induced formation of the beclin1-Atg14L-VPS34 complex ( Fig 9K ) . After autophagy activation , DFCP1 , an Endoplasmic Reticulum ( ER ) -residing PI3P binding molecule , is recruited to a membrane compartment related to autophagosome biogenesis . We found that the number of DFCP1-positive puncta was increased by AGGF1 treatment in HUVECs ( S16 Fig ) . Knockdown Atg14L significantly decreased the number of DFCP1 puncta and blocked the effect of AGGF1 ( S16 Fig ) . These data suggest that AGGF1-induced PI-3K activation resulted in an increased level of PI3P , which was required for DFCP1 recruitment . As a whole , these data demonstrate that Atg14L is required for AGGF-mediated autophagy . The above data suggest that AGGF1 induces phosphorylation/activation of JNK , which increases the activity of Vps34 and the assembly of the Becn-1-Vps34-Atg14L complex , resulting in the formation of autophagosomes and onset and completion of autophagy ( Fig 10 ) .
In this study , we found that angiogenic factor AGGF1 can activate autophagy in ECs in vitro and in MI mice in vivo , thereby identifying a new and critical upstream regulator of autophagy . Interestingly , angiogenic factor AGGF1 has been found to induce autophagy during angiogenesis . First , we demonstrated that mice with acute MI ( subject to LAD ) showed an increased expression level of either AGGF1 mRNA or protein , which was associated with up-regulated autophagy ( Fig 4 ) . Second , autophagy was inhibited in heterozygous Aggf1+/- knockout mice ( Fig 6 ) . Third , treatment of HUVECs with recombinant AGGF1 dramatically induced autophagy ( Fig 1 ) . These data indicate that AGGF1 is a strong promoter for autophagy . Identification of the important role of an angiogenic factor in autophagy is a novel aspect of the study . The data also revealed a novel in vivo function for AGGF1 in the regulation of autophagy . Mechanistically , we found that AGGF1 activates autophagy by stimulating JNK activation . Two JNK inhibitors and JNK1 siRNA blocked induction of autophagy by AGGF1 ( Fig 9 ) . AGGF1-induced JNK activation increased the activity of Vps34 and facilitated co-assembly of the Becn-1-Vps34-Atg14L complex required for autophagy . Consistently , these functions were disrupted by JNK inhibitors and JNK1 and Atg14L siRNAs ( Fig 9 ) . We have previously reported that AGGF1 activates AKT to regulate specification of veins in zebrafish [6] . However , increased AKT activation was shown to inhibit autophagy [19] . Therefore , it is the AGGF1-JNK signaling pathway , not the AGGF1-AKT signaling pathway , that modulates autophagy . This conclusion is supported by a recent finding that activation of JNK1 by FGF18 induced autophagy in bone growth [20] . During the initiation of autophagy , several other key modulators of the Becn-1-Vps34-Atg14L complex were identified , including Unc-51 Like Autophagy Activating Kinase 1 ( ULK1 ) and Autophagy And Beclin 1 Regulator 1 ( AMBRA1 ) [21] . Future studies may examine whether ULK1 or AMBRA1 affects the regulation of the complex of BECN-VPS34-Atg14 by the AGGF1-JNK pathway . Atg14L was shown to be a unique subunit of the autophagy-specific PI3K complex involved in autophagosome formation . Atg14L cooperated with the ER-resident Soluble N-ethylmaleimide-sensitive factor Attachment protein Receptor ( SNARE ) protein syntaxin 17 ( STX17 ) to assemble autophagosome at ER-mitochondria contact sites [22] . It may be interesting to investigate whether AGGF1 modulates the Atg14L-STX17 interaction involved in autophagosome formation . Another major novelty of this work involves the unexpected finding that autophagy is essential for therapeutic angiogenesis . The data in the present study demonstrated the essential role of autophagy in angiogenesis in vivo in mice . The molecular mechanism by which autophagy activates angiogenesis is unknown . One possible mechanism is that autophagy produces metabolites which are required for efficient EC proliferation , migration , and angiogenesis . Recently , fatty acid carbons and deoxyribose nucleoside triphosphate ( dNTP ) synthesis were shown to be required for EC proliferation and vascular sprouting [23] . Interestingly , AGGF1 activates autophagy not only in ECs , but also in all cells examined , including cardiac cells HL1 and H9C2 cells and vascular smooth muscle cells ( VSMCs ) . These data suggest that the AGGF1 has a much broader biological function . Because the function of AGGF1 is thought to be more related to angiogenesis , the present study focused exclusively on ECs . The dramatic therapeutic effects of AGGF1 on mice with CAD and MI may be related to the beneficial effects of AGGF1-activated autophagy not only in ECs , but also in cardiomyocytes and other cells . Future studies with EC-specific , cardiomyocyte-specific , and VSMC-specific knockout mice may distinguish specific roles of AGGF1 in ECs , cardiomyocytes , and other cells . Autophagy occurs at a low basal level under physiological conditions in the heart and is up-regulated under ischemia/reperfusion [24 , 25] . The basal level of autophagy in the heart plays an important role in maintaining cellular homeostasis; however , the role of increased autophagy in response to ischemia/reperfusion is controversial [12] . To date , it is not clear whether autophagy is protective ( cell survival ) or detrimental ( cell death ) to the heart during ischemia/reperfusion [13 , 25] . In some cases , up-regulation of autophagy was associated with a cardio-protective effect , but the cause–effect relationship was not conclusively demonstrated . In other cases , up-regulation of autophagy in response to ischemia/reperfusion was detrimental and exacerbated cardiomyocyte death [12] . In this study , we provide evidence to demonstrate that autophagy is essential for functional recovery of the heart from acute MI in mice . Preconditioning of MI mice with autophagy inhibitors bafilomycin A1 and chloroquine eliminated the therapeutic effects of AGGF1 protein on increased survival and improved myocardial contraction and recovery on cardiac structure ( Fig 7 ) . Similarly , AGGF1 protein completely lost its therapeutic effects in autophagy-deficient mice ( Becn1+/- knockout mice and AAV9-CMV-Cre Atg5 KO mice ) ( Figs 7 and 8 ) . These data suggest that autophagy plays a protective role under ischemia and is required for cardiac repair after acute MI . During the MI survival studies of Becn1+/- knockout mice and AAV9-CMV-Cre Atg5 KO mice , we made another interesting observation . Autophagy deficiency in Becn1 KO and Atg5 KO mice increased the survival of MI mice at the first 2 wk , but dramatically reduced MI survival starting after the first 2 wk ( Figs 7 and 8 ) . The data suggest that autophagy is protective to MI survival in the long run , and reduced autophagy leads to worsened survival . However , in the short run , autophagy may be damaging to MI survival , potentially due to strong stress response to ischemia . Instead , reduced autophagy leads to improved survival . The use of angiogenic factors for generating new blood vessels from existing vasculature , i . e . , therapeutic angiogenesis , has been proposed as an attractive strategy for treatment of CAD and MI patients . The VEGF ( vascular endothelial growth factors ) family of growth factors and fibroblast growth factor 2 ( FGF2 ) are the widely studied angiogenic factors for therapeutic myocardial angiogenesis . For VEGF , the therapeutic dose was difficult to establish as lower doses did not show therapeutic effect , whereas high doses caused aberrant vascular growth and vascular permeability . For FGF2 , long-term effects were not observed . We have found that vascular permeability increased in Aggf1+/- KO mice and AGGF1 administration blocked vascular permeability in mice . Therefore , AGGF1 protein therapy has a unique advantage for therapeutic angiogenesis over VEGFA . To date , no therapeutic angiogenesis treatment has been approved by the US FDA and many important issues must be resolved before therapeutic angiogenesis becomes a practical patient therapy . One solution is to identify other angiogenic factors to achieve robust therapeutic angiogenesis . In this study , we show that in a mouse model for acute MI , AGGF1 protein therapy significantly reduced mortality and dramatically improved overall cardiac function and myocardial contraction by inhibiting cardiac hypertrophy , reducing infarct size , and preventing cardiac apoptosis and fibrosis in vivo ( Fig 4 ) . AGGF1 protein therapy enhanced myocardial angiogenesis in the mouse model for acute MI ( Fig 5B ) , an effect that is relevant to the dramatically improved myocardial function and contraction . Our data have demonstrated that angiogenic factor AGGF1 is a new growth factor with promising therapeutic potential in the treatment of CAD and acute MI . One of the challenges for therapeutic angiogenesis is its efficacy . Considering the essential role of autophagy in therapeutic angiogenesis , maintaining autophagy is required for effective therapeutic angiogenesis , and increasing the levels of autophagy may be a potential strategy to robustly increase the efficacy of therapeutic angiogenesis . For example , utilization of pharmacological agents such as rapamycin [12] that increase autophagy may increase the efficacy of therapeutic angiogenesis for CAD and MI during AGGF1 protein therapy . In conclusion , the present study demonstrates that autophagy is essential for therapeutic angiogenesis , suggesting that manipulation of autophagy may serve as a potential strategy to robustly boost the efficacy of therapeutic angiogenic therapy for MI and other ischemic diseases . We show that AGGF1 activates autophagy , whereas haploinsufficiency of AGGF1 inhibits autophagy . AGGF1 increases survival of mice with acute MI , reduces infarct areas , inhibits cardiac apoptosis and fibrosis , and leads to dramatic recovery of left ventricular function and myocardial contraction . Inhibition of autophagy by an autophagy inhibitor bafilomycin A1 or in Becn1+/- and Atg5 KO mice with deficient autophagy eliminated all therapeutic effects of AGGF1 . Together , these data uncover new fundamental molecular mechanisms underlying autophagy and therapeutic angiogenesis and provide a novel treatment strategy for CAD and MI , the leading causes of sudden death worldwide .
Both animal care and experimental procedures were performed according to the guidelines by the National Institutes of Health ( NIH ) and the Guide for the Care and Use of Laboratory Animals by the National Research Council of the United States of America and approved by the Institutional Animal Care and Use Committee ( IACUC ) of Cleveland Clinic ( 2012–0899 ) and according to the Guide for the Care and Use of Animals for Research by the Ministry of Science and Technology of the P . R . China ( 2006–398 ) and approved by the Ethics Committee on Animal Research of College of Life Science and Technology of Huazhong University of Science and Technology ( [2014]IEC [S089] ) . Human umbilical vein endothelial cells ( HUVECs ) were cultured in the EGM-2 medium containing 5% ( v/v ) fetal bovine serum ( FBS ) and EGM-2 Single Quots ( Lonza ) . A polyclonal antibody against AGGF1 was from Proteintech . Antibodies against CD31 , LC3 , p62 , cleaved caspase-3 , Bcl-2 , Bax , phospho-JNK , total JNK , Vps34 , Atg14L , beclin-1 , and β-actin were from Cell Signaling . The antibody against cleaved poly ( ADP-ribose ) polymerase ( PARP ) was from Santa Cruz . Cardiac endothelial cells ( ECs ) were isolated from AGGF1+/- or wild-type mice as described previously [26] and cultured as for HUVECs described above . HUVECs were transfected with siRNA using the electroporation P5 Primary Cell Nucleofector Kits ( Lonza ) . The transfection efficiency of siRNA was determined by real-time RT-PCR analysis of a target gene . Cells were discarded if contaminated by mycoplasma or other microorganisms . HUVECs were treated with 25 μM chloroquine ( Sigma ) or 100 nM bafilomycin A1 ( Sigma ) for 1 h before AGGF1 or IgG treatment . The dosage for autophagy inhibitors was as reported previously [27 , 28] . Western blot analysis was carried out with different antibodies as described previously [29 , 30] . In brief , 4 wks after treatment , mice were anesthetized and left ventricles were dissected out for extraction of total proteins . Cultured HUVECs were washed three times with PBS and used for extraction of total proteins . Proteins were extracted on ice in 20 mM Tris-HCl , pH 7 . 6 , 150 mM NaCl , 0 . 1% DOC , 0 . 5% NP-40 , 10% glycerol , 1 mM glycerophosphate , 1 mM NaF , 2 . 5 mM Na pyrophosphate , 1 mM Na3VO4 , and a cocktail of protease inhibitors ( Calbiochem ) . Extracted proteins were then mixed with reducing laemmLi sample buffer , boiled for 10 min , separated by SDS-polyacrylamide gel electrophoresis ( PAGE ) , transferred to nitrocellulose membranes , and blotted with a primary antibody and appropriate secondary antibodies . Images from western blot were captured and quantified using 1-D Analysis Software and Quantity One ( Bio-Rad ) . The full-length human AGGF1 cDNA was cloned into a bacterial expression vector pET-28b ( Novagen ) , resulting in an overexpression construct for 6xHis-tagged AGGF1 , pET-28-AGGF1 [2] . The pET-28-AGGF1 expression construct was transformed into Escherichia coli BL21 ( DE3 ) Star , and 6xHis–AGGF1 protein was overexpressed and purified using a Ni-NTA agarose column according to the manufacturer's instructions ( Qiagen ) . The eluted protein was dialyzed , and quality of purification was examined by SDS-PAGE by coomassie blue staining and western blot analysis using an anti-AGGF1 antibody [2] . Male C57BL/6 mice were used for all studies . The Aggf1+/- KO mice with a gene-trap allele were created by us . The autophagy-deficient Becn1+/- KO mice ( Becn1tmeBlev ) were from the Jackson Laboratory . Atg5flox/flox mice [31] were kindly provided by Dr . Noboru Mizushima at University of Tokyo , RIKEN BRC , and Dr . Quan Chen at the Chinese Academy of Sciences . To create Atg5 KO in mice , Atg5flox/flox mice were anesthetized and injected with 0 . 2 ml of the adenovirus AAV9-GFP or AAV9-CMV-Cre ( Vector Biolabs ) into the left ventricular myocardium using a 35-gauge needle at multiple sites . Animal care and experimental procedures were approved by the Institutional Animal Care and Use Committee ( IACUC ) of Cleveland Clinic and the Ethics Committee on Animal Research of College of Life Science and Technology at Huazhong University of Science and Technology . The acute MI model was created by ligation of the left anterior descending ( LAD ) coronary artery as described previously [32] with male mice at the age of 10–12 wk ( about 25 g ) . The mice were anesthetized with an intraperitoneal injection of sodium pentobarbital ( 50 mg/kg ) and then intubated using a fine polyethylene cannula connected to a small animal ventilator . The body temperature of the mouse was monitored using a rectal sensor at all times during the surgical procedure and maintained by a heated surgical plate . After respiration of the mouse was controlled by the ventilator , a thoracotomy incision was made in the second intercostal space , and the heart was exteriorized out of the chest . The LAD coronary artery was ligated permanently with a 7–0 nonabsorbable surgical suture and the heart was then returned inside the chest . The chest wall was closed in layers and skin incision was closed by sutures . The mouse was then removed from the ventilator and kept warm in a cage at 37°C overnight . Sham-operated mice were subjected to the same surgical treatment , but the LAD was not ligated . C57BL/6N mice were assessed for baseline cardiac function with echocardiography prior to surgery and then subjected to LAD ligation . One day after the surgery , each animal was examined with echocardiography to confirm the success of the MI surgery . One week after the surgery , the mice were injected intravenously with AGGF1 ( 0 . 25 mg/kg body weight ) or the same dose of nonimmune control IgG ( 0 . 25 mg/kg body weight , R&D systems ) twice a week for 2 wk or until death . The IgG was used as the control as reported in other studies [33–35] and was considered to be a better control than saline without any protein . In the set of experiments for bafilomycin A1 ( an autophagy inhibitor ) , we divided the mice that survived the LAD ligation procedure for 1 wk and were ensured to have successful MI by echocardiography into four groups: an IgG-treated group with vehicle pretreatment ( IgG with 0 . 1% DMSO ) , an IgG-treated group with pretreatment with bafilomycin A1 ( 0 . 3 mg/kg body weight , Sigma ) , an AGGF1-treated group with vehicle pretreatment , and an AGGF1-treated group with pretreatment with bafilomycin A1 . Bafilomycin A1 is a vascular H+-ATPase inhibitor and a membrane-permeable lysosomal inhibitor that blocks autophagosome-lysosome fusion to prevent the final digestion step in the process of autophagy [36] . In these experiments , one day after surgery , the mice were assessed by echocardiography for the success of the MI procedure and then treated with bafilomycin A1 or vehicle control by daily intraperitoneal injection for 5 d . One week after the surgery , the AGGF1 protein or control IgG was then injected through tail vein twice a week for 2 wk . All mice were randomized in all experimental protocols . Echocardiography was performed with a Vevo 2100 High-Resolution Micro-Ultrasound System ( Visual Sonics Inc . , Toronto , Canada ) at five different time points: pre-surgery , 1 d after LAD ligation and administration of AGGF1 protein or IgG ( 1 d ) , 1 wk , 2 wk , and 4 wk after the surgery to determine the baseline heart function and ventricular dimensions in different experimental groups of mice . The mouse under study was anesthetized with 1% isoflurane , placed on a heat pad in the supine position , and kept at 37°C to minimize confounding of data by fluctuating body temperatures . Hairs were removed by depilatory cream on the left chest before echocardiography . A 30 MHz variable frequency transducer was used to capture two-dimensional echocardiographic images of the mid-ventricular short axis and parasternal long axes when the heart rate of the mouse was between 450 and 550 beats per minute . Echocardiographic analysis was performed with digital images using a standard formula as previously described [14 , 15] . All procedures of echocardiography , including data acquisition and analysis , were performed by a researcher who was blind of the experimental treatments to avoid biases . Four weeks after AGGF1 or IgG treatment , mice were anesthetized , euthanized , and the hearts were excised and fixed overnight . The fixed hearts were sectioned and immunohistochemical staining was performed on paraffin-embedded sections with a primary antibody against AGGF1 or CD31 , which was followed by incubation with a biotinylated secondary antibody as described previously [29 , 37] . The sections were then treated with peroxidase-conjugated biotin–avidin complex using VECTASTAIN ABC-AP and visualized by DAB . Slides were counterstained with H&E staining . Vessel density was evaluated by counting the number of neovessels and arterioles in five random and non-repeated high power fields . Myocardial apoptosis was measured first by the TUNEL assay ( terminal deoxynucleotidyltransferase ( Tdt ) -mediated dUTP nick-end labeling ) [29] . Four weeks after treatments with AGGF1 or IgG control , mouse hearts were excised and fixed in 4% paraformaldehyde , embedded in paraffin , cut into 5 μm-thickness sections and used for the TUNEL assay using the In Situ Cell Death Detection kit ( Roche Diagnostics GmbH ) . The images were visualized under a fluorescence microscope and captured . The border zone of fibrosis was examined carefully . More than five fields in three different sections were examined for each mouse by a researcher who was blinded to the treatments . The percentage of the number of TUNEL-positive cells over the total number of nuclei as determined by DAPI was calculated . Heart sections incubated with the label solution but without terminal transferase were used as negative controls . A TUNEL-positive control was included by incubating sections with DNase I ( 3000 U/ml in 50 mm Tris-HCl , pH 7 . 5 , 1 mg/ml BSA ) for 15 min at 15–25°C to induce DNA strand breaks prior to the labeling procedure . Myocardial apoptosis was also measured by the caspase-3 activity [29] . We used a caspase-3 colorimetric assay kit and followed the manufacturer’s instruction ( Promega ) . The absorbance of p-nitroaniline cleaved by caspase-3 was measured at 405 nm using FlexStation3 . Data on the caspase-3 activity were standardized over the sham group and the fold change of the caspase-3 activity ( relative caspase activity ) was compared among different groups . HUVECs were seeded on coverslips in 24-well plates . The AGGF1 protein at different concentrations or an equal dose of negative control IgG in PBS was added to the wells at the next day . After 2 h of incubation , HUVECs were incubated under a hypoxic condition with 1% O2 . After 12 h , the coverslips were washed in PBS and fixed in 4% paraformaldehyde for 1 h at room temperature . After blocking and permeabilization , the slides were used for the TUNEL assay as described above . HUVECs were seeded in 6-cm plates at 1 × 106/ml . The AGGF1 protein at different concentrations or an equal dose of IgG in PBS was added to the wells at the next day . After hypoxic treatment , cells were lysed and the caspase-3 activity was determined using the caspase-3 fluorescence kit as described above . Apoptosis of HUVECs was also assessed by western blot analysis using anti-cleaved caspase-3 , anti-Bax , anti-Bcl-2 , and anti-cleaved PARP as described above . For cell proliferation , HUVECs were cultured in 96-well plates , incubated with AGGF1 or control IgG for 2 h , and cultured in a hypoxia chamber equilibrated with 1% O2 for different time points . The cells were then used for cell proliferation assays with the 2- ( 2-methoxy-4-nitrophenyl ) -3- ( 4-nitrophenyl ) -5- ( 2 , 4-disulfophenyl ) -2H-tetrazolium , monosodium salt ( WST-8 ) kit ( Cell Counting Kit-8 , Dojindo Laboratories ) . The number of living cells in triplicate wells was directly proportional to the amount of the WST-8formazan dye generated by measuring the absorbance at 450 nm . Two different types of cell migration assays were carried out as described by us [2] . The first is the scratch wound assay . HUVECs were grown in 6-well plates as a confluent monolayer and mechanically scratched with a pipette tip . The cells were incubated with EBM containing the AGGF1 protein ( 500 ng/ml ) or control IgG and other treatment agents . Migration was quantified as the ratio of the area covered with cells over the cell-free area . The second is the Transwell migration assay . HUVECs were plated into the upper compartment and were allowed to migrate towards AGGF1 ( 500 ng/ml ) or IgG in the lower chamber ( Corning Costar , MA , USA ) . After 6 h of incubation , HUVECs on the bottom of the Transwell membrane were fixed with 4% paraformaldehyde at 37°C for 20 min and stained with hematoxylin at 37°C for 5 min . The membranes were washed three times with PBS and photographed . The migrated cells on the bottom of the surface were counted in eight standardized fields . Approximately 5 × 104 HUVECs in EBM were seeded onto matrigel ( Corning ) with the AGGF1 protein ( 500 ng/ml ) or negative control IgG . After 6 h of incubation , 2 μg/ml of Calcein AM ( Invitrogen , USA ) was added directly to the well and incubated for 20 min . The well was washed with PBS three times . The images were visualized under a microscope and captured . The number of mature vessel tubes formed was counted as described previously [1 , 2] . Only completely closed tubes were regarded as mature tubes . The mouse aorta was removed under aseptic conditions . The dissected aorta was transferred to a dish containing cold EBM . To avoid contamination of other cell types , excessive fat tissue was quickly removed by forceps , and the aorta was then embedded into matrigel . The AGGF1 protein or PBS was added into the media . After 4 d of incubation in 37°C with 5% CO2 , images of newly formed vessels were captured under a microscope and analyzed . HUVECs were transfected with an expression plasmid for HA-tagged Vps34 using nucleofection ( Lonza ) and then incubated with different amounts of AGGF1 or control IgG . Exogenous Vps34 expressed in HUVECs was immunoprecipitated using an anti-HA antibody , measured using the class III PI3K assay , and was carried out using the Class III PI3-kinase ELISA kit ( Echelon , UT , US ) . HUVECs were transfected with an overexpression plasmid for pGFP-2xFYVE by nucleofection ( Lonza ) . Twenty-four hours later , HUVECs were treated with or without JNK inhibitors for 1 h and then incubated with AGGF1 for 12 h . Images were captured with a florescence microscope and analyzed for the number of green GFP-FYVE dots with endocytic membrane Vps34 per cell . The assay was performed following the protocol described previously [18 , 38] . Briefly , HUVECs were plated and incubated with 3H-labeled leucine for 36 h . Then , cells were washed three times with PBS and incubated in unlabeled medium for 24 h to release short-lived proteins . After the chase period , cells were washed again three times with PBS and cultured in medium with AGGF1 or IgG for 12 h . The supernatant was collected and precipitated with trichloroacetic acid ( TCA ) . The TCA-soluble fraction was measured by liquid scintillation counting . Total cell radioactivity was analyzed after lysis with 0 . 1 M NaOH . The degradation of long-lived protein was calculated as a percentage of the radioactivity in TCA-soluble fraction to the total cell radioactivity . HUVECs were transfected with an overexpression plasmid for GFP-LC3B by nucleofection ( Lonza ) . After 24 h of culture , HUVECs were treated with AGGF1 or control IgG for 12 h . Images were captured with a florescence microscope and analyzed for the ratio of green punctuate cells over all green cells with successful transfection . Autophagy was also monitored by electron microscopy as described previously [39] . Total RNA was extracted from cultured cells or mouse hearts using Trizol ( Invitrogen ) according to the manufacturer’s instruction . A total of 0 . 5 μg of RNA samples was reverse-transcribed using M-MLV Reverse Transcriptase according to the manufacturer’s protocol ( Promega ) . Quantitative real-time PCR analysis was then performed using the FastStart Universal SYBR Green Master ( Roch ) and a 7900 HT Fast Real-Time PCR System ( ABI ) as described previously [40–42] . Experiments were performed in triplicate . Two-group comparisons were analyzed by a Student’s t test or nonparametric Wilcoxon rank test whenever appropriate ( e . g . , when the sample size was small and/or the distribution was not normal ) . For comparisons of more than two groups , one-way ANOVA or the generalized linear regression approach was employed for normal distributions and the Kruskal Wallis test for non-normal or small samples . Bonferroni correction was used to adjust for multiple testing after the overall F or Kruskal Wallis test showed statistical significance . The extract testing was used when the sample size was small ( e . g . , when all group sizes were <10 ) . The survival curve analysis was analyzed using the Kaplan–Meier product-limit approach and compared by the log-rank test . Statistical significance is indicated by * p < 0 . 05 and ** p < 0 . 01 . | Coronary artery disease is the number one killer disease worldwide . Recently , therapeutic angiogenesis has been proposed as an attractive new strategy for treating this and other ischemic diseases . This study establishes the angiogenic factor AGGF1 as a novel target and agent that can successfully treat coronary artery disease and acute myocardial infarction and dramatically improve survival and cardiac function in mouse models . We present the unexpected finding that AGGF1 has these effects via activating autophagy , and that autophagy is essential for therapeutic angiogenesis in animals . We find that AGGF1 is a novel master regulator of autophagy not only in endothelial cells but also in all other cell types examined in the study . Mechanistically , AGGF1 activates autophagy by activating JNK , which leads to activation of the Vps34 lipid kinase and assembly of the Becn1-Vps34-Atg14 complex involved in the initiation of autophagy . The study thus provides a link connecting the therapeutic angiogenesis and autophagy pathways in heart disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"death",
"autophagic",
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"diagnostic",
"radiology",
"ultrasound",
"imaging",
"cardiovascular",
"physiology",
"cardiovascular",
"anatomy",
"gene",
"regulation",
"cell",
"processes",
"echocardiography",
"angiog... | 2016 | Angiogenic Factor AGGF1 Activates Autophagy with an Essential Role in Therapeutic Angiogenesis for Heart Disease |
The mechanisms by which RNA-binding proteins control the translation of subsets of mRNAs are not yet clear . Slf1p and Sro9p are atypical-La motif containing proteins which are members of a superfamily of RNA-binding proteins conserved in eukaryotes . RIP-Seq analysis of these two yeast proteins identified overlapping and distinct sets of mRNA targets , including highly translated mRNAs such as those encoding ribosomal proteins . In paralell , transcriptome analysis of slf1Δ and sro9Δ mutant strains indicated altered gene expression in similar functional classes of mRNAs following loss of each factor . The loss of SLF1 had a greater impact on the transcriptome , and in particular , revealed changes in genes involved in the oxidative stress response . slf1Δ cells are more sensitive to oxidants and RIP-Seq analysis of oxidatively stressed cells enriched Slf1p targets encoding antioxidants and other proteins required for oxidant tolerance . To quantify these effects at the protein level , we used label-free mass spectrometry to compare the proteomes of wild-type and slf1Δ strains following oxidative stress . This analysis identified several proteins which are normally induced in response to hydrogen peroxide , but where this increase is attenuated in the slf1Δ mutant . Importantly , a significant number of the mRNAs encoding these targets were also identified as Slf1p-mRNA targets . We show that Slf1p remains associated with the few translating ribosomes following hydrogen peroxide stress and that Slf1p co-immunoprecipitates ribosomes and members of the eIF4E/eIF4G/Pab1p ‘closed loop’ complex suggesting that Slf1p interacts with actively translated mRNAs following stress . Finally , mutational analysis of SLF1 revealed a novel ribosome interacting domain in Slf1p , independent of its RNA binding La-motif . Together , our results indicate that Slf1p mediates a translational response to oxidative stress via mRNA-specific translational control .
The control of translation in response to external stimuli plays an important role in the regulation of gene expression . Indeed , some estimates of the relative contributions of different molecular mechanisms to the overall control of gene expression highlight a dominant role for translational control [1] , [2] . Inhibition of translation initiation in particular forms a focus for much of this regulation . For example , in response to external stimuli , such as amino acid starvation or hydrogen peroxide stress , global translation initiation is normally reduced whilst significant numbers of specific mRNAs continue to be translated [3] . A variety of mechanisms exist to reduce the translation of most mRNAs e . g . through eIF2α phosphorylation , matched with complementary mechanisms to allow certain mRNAs to escape such global controls . One mechanism described to facilitate escape from global controls is via upstream ORFs; for example on the GCN4 and ATF4 mRNAs in yeast and mammals , respectively . In addition to intrinsic mRNA properties , a large number of RNA binding proteins ( RBPs ) are known to bind specific mRNAs in order to either activate or repress their translation [4] , forming a cellular network of post-transcriptional regulation above that exerted at the transcriptional level . Over 600 proteins encoded by the yeast genome are predicted to bind RNA [5] but the mechanisms by which RBPs control the translation of subsets of mRNAs are not yet clear . The La-motif ( LaM ) is an RNA binding domain which defines a superfamily of RNA-binding proteins conserved across eukaryotes [6] . Most organisms generally possess a true La protein ortholog with a LaM and one or more adjacent RNA-recognition ( RRM ) domains , which function in the nucleus binding RNA polymerase III primary transcripts . Human La was first identified as an autoantigen in patients suffering from autoimmune disorders . In addition there are a larger number of La related proteins ( LARPs ) most of which share the conserved adjacent LaM and RRM domains , but these proteins function in diverse processes , with both human LARP1 and LARP4 being implicated in binding polyA mRNAs and ribosomes [6]–[9] . S . cerevisiae has three LaM proteins; Lhp1p , Slf1p and Sro9p . Lhp1p is a true La protein ortholog , while Slf1p and Sro9p are atypical-LARPs that have a central LaM but lack any currently known RRM [10] . Slf1p and Sro9p appear evolutionarily most closely related to the LARP1 and 4 families [6] . In common with LARP1 and 4 family proteins , Slf1p and Sro9p preferentially associate with translating ribosomes and the polyA binding protein [7]–[11] and are believed to stimulate protein synthesis and/or promote mRNA stability of their bound mRNAs . Slf1p and Sro9p are homologous , sharing 30% identity at the amino acid level , but outside of the La domain there is little sequence similarity between Lhp1p and Slf1p/Sro9p . Cells deleted for SRO9 display a slight slow growth phenotype , although null alleles lacking SRO9 , SLF1 and LHP1 do not show any additive effects suggesting that they are not functionally redundant [10] , [12] . Taken together , these data suggest that Slf1p and Sro9p are not required for protein synthesis but may have a role in the regulation of translation , possibly in an mRNA-specific manner . Interestingly , the SLF1 mRNA , but not the SRO9 mRNA , has a Puf3p binding site in its 3′-untranslated region , via which Puf3p is believed to repress translation of the SLF1 mRNA [13] . As Puf3p primarily binds many mRNAs encoding mitochondrial functioning proteins this suggests that Slf1p may also have a role in mitochondrial function . Increased SLF1 mRNA translation is also thought to promote respiration and the extension of yeast chronological life span . To gain insight into the functions of this intriguing protein family , we have investigated the roles of the yeast LARPs using a full range of genome-scale techniques , including at the transcriptome , translatome and quantitative proteome level . Our studies have revealed a key role for Slf1p in the activation of translation of mRNAs critical for reprogramming gene expression to facilitate the cellular response to oxidative stress . We show that Slf1p has a critical role in mediating the coordinated cellular oxidative stress response to reactive oxygen species and SLF1 is required for resistance to oxidative stress . In addition , mutational analysis of Slf1p reveals that it does have a novel ribosome-interaction domain independent of its mRNA binding LaM . Taken together our results provide a system-wide analysis of the role of the LARP Slf1p in an important cellular defence mechanism , highlighting it as a key player in the translational control of gene expression under oxidative stress conditions .
Slf1p and Sro9p are related LARPs that both associate with translating ribosomes . We therefore decided to assess their roles in RNA biology using a range of post-genomic techniques . Slf1p and Sro9p share 37% overall amino acid identity , 57% within the LaM , suggesting they likely have similar or overlapping roles . We used RNA sequencing ( RNA-Seq ) to assess the impact of deletion of each LARP independently , by comparing the relative total transcript levels in slf1Δ and sro9Δ mutant strains with an isogenic wild-type strain . Triplicate samples were processed using a standard workflow ( see Materials and Methods ) that revealed 204 mRNAs were significantly increased ( FDR<0 . 05 ) and 253 mRNAs decreased in abundance compared to the parental strain in an slf1Δ mutant ( S1 Table and S1A Fig . ) . In comparison , an SRO9 deletion mutant showed a greater impact , with 702 mRNAs increased and 666 mRNAs decreased in an sro9Δ strain ( S2 Table and S1A Fig . ) . Although more transcripts alter following loss of SRO9 , in general the degree of change appears more modest with only 35 transcripts increasing and 85 decreasing by more than 2 fold . In contrast the variation in fold change is much greater in the slf1Δ strain ( Fig . 1A ) . RNA-binding proteins that mediate post-transcriptional control can interact with functionally related mRNAs [14] . We therefore searched for functional categories enriched among the differentially expressed slf1Δ and sro9Δ mRNAs using MIPS category classifications . Classes including respiration , protein synthesis and the oxidative stress response were statistically over represented among slf1Δ down-regulated transcripts , while only protein synthesis was similarly affected following sro9Δ ( Fig . 1B ) . There are 141 mRNAs down-regulated by both gene deletions suggesting some overlap in the targets of each LARP and as expected they are enriched in genes involved in protein synthesis according to MIPS ( Fig . 1B ) . Thus , both Sro9p and Sfl1p appear to contribute to the regulation of mRNAs involved in protein synthesis , while Slf1p has targets in additional pathways . The functional classes enriched in transcripts that were up-regulated following the loss of each factor are largely distinct ( Fig . 1B ) , as expected , since fewer mRNAs ( only 71 ) were up-regulated in both datasets ( S1B Fig . ) suggesting that these proteins influence the mRNA architecture of the cell in a related but distinct manner . A recent study investigated the effect of overexpressing SLF1 on mRNA abundance and identified 852 mRNAs that increase in abundance and 599 mRNAs that decrease in abundance using a microarray-based approach [12] . We compared this dataset with our slf1Δ dataset and found a highly significant overlap between those mRNAs that increase in abundance when SLF1 is overexpressed and those mRNAs that decrease in abundance in an slf1Δ strain ( Fig . 1C ) . Indeed , the 99 transcripts that appear in both experiments are significantly enriched for ribosome biogenesis , ribosomal proteins and oxidative stress response functions ( Fig . 1C ) . The functional categories of transcripts that alter in abundance in an sro9Δ strain are different compared with those in the slf1Δ , with the exception of transcription ( up-regulated ) and protein synthesis ( down-regulated ) . When comparing the two transcriptomes to each other , there are 71 transcripts that increase in both the slf1Δ strain and the sro9Δ strain and 141 transcripts that decrease in abundance in both strains ( S1B Fig . ) . However , there is also a modest crossover between the transcript sets that increase in one of the two mutant strains but decrease or don't change in the other ( S1B Fig . ) . Beyond the shared role of the yeast LARPs in regulating mRNAs involved in protein synthesis , particularly mRNAs encoding ribosomal proteins , we noted Slf1p's additional potential role in mediating the responses to oxidative stress . Since it has been shown previously that Slf1p promotes copper detoxification [12] , which is related to oxidative stress tolerance and which require regulations and reprogramming of protein synthesis [15] , it is this role that we explore further in this study . We developed a rapid RIP-Seq approach to identify RNAs bound by TAP-tagged proteins , using strains bearing genomically-integrated C-terminal TAP tags . Our strategy involved minimally disturbing cells and processing them as rapidly as possible to maintain physiological interactions . This used swift cell freezing in liquid nitrogen and cell lysis , followed by an immunoprecipitation step using IgG conjugated to paramagnetic beads . Using paramagnetic beads enabled rapid immunoprecipitation and washes and resulted in sample processing that generated significantly reduced background binding in comparison to approaches relying on extended incubations such as cross-linking protocols or employing agarose beads which are prone to non-specific interactions . In our protocol total RNA and RNA isolated from the Slf1p-TAP and Sro9p-TAP immunoprecipitated fractions was depleted for rRNA and then converted into cDNA sequencing libraries using standard methods ( see Materials and Methods ) . Triplicate Slf1p-TAP and Sro9p-TAP RIP-Seq experiments identified 488 and 1433 mRNAs , respectively , that are significantly enriched above total RNA ( corrected FDR<0 . 05 ) ( S3–S4 Tables ) . When the two datasets were compared , only 264 transcripts were identified as being significantly enriched by both Slf1p and Sro9p ( Fig . 2A ) . Despite this , a Gene Ontology analysis of the independent Slf1p and Sro9p mRNA-target sets shows common enrichment above the genomic background for mRNAs involved in protein synthesis and mitochondrial functions ( Fig . 2B ) . Functional analysis of the 264 transcripts bound by both Slf1p and Sro9p also identified enrichment for protein synthesis and mitochondrial functions . Sro9p mRNA targets were also enriched for ‘cell cycle and DNA processing’ and ‘biogenesis of cellular components' categories . During the course of our study , mRNA targets for both Slf1p and Sro9p were additionally reported by an independent study using a RIP-Chip approach [12] . Although the overlap between the studies appears to be modest , a significant number of transcripts are common to both datasets and the functional classes enriched are the same ( S2 Fig . ) . Both studies identify protein synthesis , particularly mRNAs encoding ribosomal proteins as significant targets of Slf1p and Sro9p . When comparing functionally enriched gene classes of the Slf1p and Sro9p mRNA targets ( Fig . 2B ) with those mRNAs that change transcriptionally in the corresponding mutant strains ( Fig . 1B ) , the enrichment in common functional themes is further reinforced . Genes linked to protein synthesis are both transcriptionally down-regulated in the deletion strains and bound by both factors . Similarly , genes within the ‘mitochondrion’ MIPS category are enriched in both Sro9p targets and are up-regulated in the sro9Δ mutant . We therefore examined the specific overlap in transcripts between the two sets , comparing transcriptionally regulated genes with targets identified in our RIP-Seq experiment . Notably , Slf1p-mRNA targets are also down-regulated in the slf1Δ mutant ( 56 mRNAs , P = 3 . 67×10−11; Fisher's Exact test ) , whereas there is little crossover with those mRNAs ( 6 mRNAs ) that increase in abundance ( S3A Fig . ) , suggesting that Slf1p is required for maintaining steady state target mRNA levels . In contrast , there were far fewer than expected Sro9p-TAP bound mRNAs whose transcript levels are altered in an sro9Δ mutant ( S3B Fig . ) . The origin of this effect is clearly evident in Fig . 2C , which shows the distribution of log2 transcriptional fold changes in the deletion strains , highlighting the distributions of transcripts also bound by the equivalent TAP-tagged protein in the RIP experiment; Slf1p targets are clearly less abundant in slf1Δ cells while Sro9p target abundance is apparently unchanged . Applying increasing FDR cut off stringencies to our Rip-Seq data to restrict our analysis to the most significant hits maintains these trends ( S4 Fig . ) . To gain further insight into Slf1p and Sro9p functions , we compared the RIP-Seq targets with other recently published genome wide measurements of mRNA half-life , PolyA tail length and ribosome occupancy by ribosome footprinting [16] . The only significant finding was that Slf1p mRNA targets are enriched for mRNAs that are actively translated and therefore have a higher translational efficiency ( Fig . 2D ) . We conclude that the LARPs bind both overlapping and distinct sets of mRNAs including highly translated mRNAs such as those encoding ribosomal proteins . Unexpectedly , loss of Sro9p does not significantly alter mRNA target levels , while loss of Slf1p does . This suggests that Slf1p targets may be under greater dynamic control than those bound by Sro9p , or that other factors can more easily compensate for loss of Sro9p than for Slf1p . Our transcriptome analyses suggest a role for SLF1 in mediating the oxidative stress response . We further examined this finding by testing the sensitivity of slf1Δ and sro9Δ mutants to a range of stress conditions . We first confirmed that the growth of both mutant strains is inhibited by copper as previously described [17] and found that slf1Δ mutants , and to a lesser extent sro9Δ mutants , are sensitive to hydrogen peroxide stress ( Fig . 3A ) . Mutants deleted for SLF1 also showed a modest sensitivity to cadmium , which like hydrogen peroxide causes oxidative stress ( Fig . 3A ) . Sensitivity to stress conditions is not a general property of slf1Δ and sro9Δ mutants since little or no sensitivity was found with various other stress conditions , including growth at elevated or lower temperatures ( 37°C or 16°C ) , at pH 5 , or high salt ( 1 M NaCl ) ( S5A Fig . ) . Plasmid-borne SLF1 complements the slf1Δ mutant sensitivity to hydrogen peroxide , confirming that Slf1p is important for oxidative stress tolerance ( S5B Fig . ) . Our data indicate that Slf1p is important for oxidative stress tolerance . Because many stress responsive genes are transcriptionally and/or translationally activated in response to stresses such as hydrogen peroxide [15] , [18] we assessed Slf1p RNA targets by RIP-Seq following treatment of cells with 0 . 4 mM hydrogen peroxide , a concentration sufficient to induce a robust and rapid reprogramming of gene expression [15] . Hydrogen peroxide treatment increased the number of significantly bound mRNAs to 1053 compared to 488 in the untreated Slf1p RIP-Seq experiment ( S5 Table ) . Reassuringly , there was still a highly significant overlap between both datasets and 358 transcripts were bound by Slf1p during normal and oxidative stress conditions ( Fig . 3B ) . Functional enrichment analysis of the stress-bound mRNAs again highlighted the oxidative stress response , mitochondrial function , electron transport and protein synthesis as significantly enriched MIPS categories ( Fig . 3C ) . The expanded set of RNA-targets retain very high ribosome occupancy ( Fig . 2D ) . This suggests that Slf1p is binding actively translated mRNAs that are required for the cellular response to oxidative stress . Both Sro9p and Slf1p associate with translating ribosomes [10] and oxidative stress is known to cause a global reprogramming of protein synthesis [15] , [19] . Our data show that under these conditions , Slf1p binds oxidative stress regulated mRNAs and that slf1Δ cells are sensitive to oxidative stress conditions . We therefore investigated the impact of SLF1 and SRO9 deletions on the global translational response to hydrogen peroxide stress using polyribosomal profiling . Deletion of SLF1 does not affect the polyribosome profile in unstressed cells ( Fig . 4A ) . In contrast , following a 15 minute treatment with 0 . 25 mM hydrogen peroxide , the slf1Δ strain exhibited a more dramatic inhibition of translation initiation than the parental strain; as detected by an increase in the 80S monosome peak compared to the polysome ribosomal peaks ( Fig . 4A ) . A similar but less pronounced effect was observed for the sro9Δ strain ( S5C Fig . ) . Repeating polyribosomal profiling experiments over a range of hydrogen peroxide concentrations ( S6 Fig . ) revealed that the slf1Δ mutant strain exhibits maximum translation inhibition at a lower hydrogen peroxide concentration than the wild-type strain ( quantification shown in Fig . 4B ) . Based on these findings , we suggest that the enhanced inhibition of translation in response to oxidative stress conditions may account for the growth sensitivity of this mutant . As noted previously , Slf1p and Sro9p co-sediment with ribosomal subunits across polysome gradients , suggesting that these proteins interact with actively translating ribosomes [10] . To track Slf1p and Sro9p across gradients , we used the Slf1-TAP strain employed in our RIP-Seq experiments . We confirmed that the addition of a TAP-tag does not affect the growth or stress sensitivity of the yeast strains ( S5A Fig . ) . Treating cells with 0 . 4 mM hydrogen peroxide causes an inhibition of translation initiation , but Slf1p and Sro9p still associate with polysomes ( Fig . 4C ) , indicating that they remain associated with the fraction of ribosomes still actively translating mRNAs . Cap-dependent eukaryotic translation initiation requires the eukaryotic translation initiation factors eIF4E and eIF4G and is enhanced by the poly ( A ) binding protein , Pab1p . The cap is bound by eIF4E and Pab1p binds the poly ( A ) tail of the mRNA . The scaffold protein eIF4G binds to both eIF4E and Pab1p forming a ‘closed loop’ complex that is thought to promote protein synthesis [20] . If our hypothesis that the yeast LARPs remain associated with actively translating mRNAs following stress is correct , these translation factors should also be associated with the LARPs following stress . To test this , we used the Slf1p-TAP and Sro9p-TAP strains and performed TAP affinity purifications and Western blotting with specific antibodies to assess whether translation initiation factors remain associated with each LARP . In purifications from unstressed cells , both LARPs immunoprecipitated a fraction of key closed-loop proteins eIF4E , eIF4G and Pab1p , as well as markers for the 40S ( Rps3p ) and 60S ( Rpl35p ) ribosomal subunits ( Fig . 5A lane 8 and 5C lane 8 ) . This is consistent with previous work identifying interactions between Slf1p and eIF4E or Pab1p [21] . However , RNase I treatment diminished co-immunoprecipitation of the closed loop factors with both LARPs ( Fig . 5 panels A and C , lanes 9 ) implying that these interactions are mRNA mediated . Repeating the experiments following treatment with 0 . 4 mM hydrogen peroxide for 15 minutes , largely maintained these interactions ( Fig . 5 B and D , comparing lanes 10 and 11 ) , although the interaction of eIF4G with Slf1 appears more sensitive to hydrogen peroxide than the other factors . Again these interactions were RNase I sensitive ( Fig . 5 B and D compare lanes 11 and 12 ) . In summary , the interactions of both LARPs with initiation factors that are components of the closed loop complex , as well as ribosomal proteins , suggests that they interact with actively translated mRNAs following stress . Our data so far , strongly suggest that Slf1p has a significant role in promoting or protecting the translation of genes necessary for the cellular response to oxidative stress . If so , we reasoned that at least some of the oxidative stress induced changes in gene expression manifest at the translational level would be dependent upon Slf1p . Therefore , to examine oxidative stress induced proteome changes , we used a label-free quantitative mass spectrometry ( LC-MS ) approach comparing the total cell extract proteome during normal growth conditions and following addition of hydrogen peroxide . Five replicate wild-type stressed and unstressed samples were analysed , enabling quantitation of 1565 proteins in the wild type strain ( see Materials and Methods and S6 Table ) , of which 315 altered significantly ( 249 up and 66 down ) in response to peroxide stress ( FDR p<0 . 05 ) . Significantly , 97 of these are encoded by Slf1p mRNA targets identified by our RIP-Seq following oxidative stress . By repeating the proteome analysis in an slf1Δ strain , we identified 2140 proteins , of which only 2 increased in abundance significantly ( FDR p<0 . 05 ) in response to hydrogen peroxide ( S7 Table ) , suggesting that the oxidative stress induced reprogramming of the proteome is significantly muted in slf1Δ cells . It is possible that that some of the decrease in bulk translational activity in the slf1 mutant might arise due to decreased mRNA abundance in the mutant strain . However , of the 248 proteins which showed an attenuated protein induction in the slf1 mutant , only 33 were found to decrease at the mRNA level in the slf1 mutant strain . Of the 97 proteins identified as altered by oxidative stress in the wild-type strain that are encoded by Slf1p mRNA targets , 83 of these proteins were also quantified in our slf1Δ proteomics experiment ( red and yellow symbols in Fig . 6A ) . Fourteen of these proteins are involved in the oxidative stress response ( red symbols in Fig . 6A ) as are a further 22 proteins that are not encoded by Slf1p-mRNA targets ( blue symbols in Fig . 6A ) . It is clear from the plot in Fig . 6A that the induction of oxidative stress related proteins is significantly attenuated in slf1Δ cells ( red and blue symbol positions deviate significantly below the dotted X = Y line shown in Fig . 6A ) . Comparing Slf1p mRNA-targets identified following stress conditions with the proteome data obtained with wild-type cells identified 109 proteins induced by oxidative stress that are encoded by Slf1p mRNA targets . Although 100 of these proteins were also identified and quantified in our slf1Δ proteomics experiment , only 13 significantly increased in abundance after peroxide stress in a slf1Δ strain . Importantly , functional classification of the 87 proteins that were no longer stress induced in the mutant strain showed significant enrichment for proteins involved in the oxidative stress response and detoxification and repair of oxidant damage ( red dots , Fig 6A ) , highlighting Slf1p's role in mediating translation of these key transcripts . Fig . 6B shows an overview of the antioxidants and stress repair and detoxification proteins which comprise the yeast oxidative stress response . Proteins indicated in red correspond to Slf1p-mRNA targets where their oxidative stress protein induction is attenuated in an slf1Δ strain ( red circles in Fig . 6A ) . These include a number of key antioxidants including superoxide dismutase ( Sod1p and Sod2p ) , thioredoxins ( Trx1p and Trx2p ) , thioredoxin reductase ( Trr1p ) , peroxiredoxins ( Tsa1p , Ahp1 ) , glutaredoxin ( Grx2p , Grx5p ) , glutathione peroxidases ( Gpx2p , Gpx3p ) and the stress protective enzyme sulfiredoxin ( Srx1p ) . A number of proteins were also identified where their oxidative stress induction was attenuated in the slf1Δ mutant but they were not identified as direct mRNA targets of Slf1p in the RIP-Seq analysis . These proteins are indicated in blue on Fig . 6B ( and blue circles in Fig . 6A ) . Additional proteins are highlighted which were identified as Slf1p-mRNA targets but where we were unable to detect any high confidence peptide identifications for their parent proteins in the proteomics analysis ( Fig . 6B , in green ) . When taken together our series of ‘omics studies reveal the importance of Slf1p in mediating translational control of the expression of key oxidative stress genes . We propose that Slf1p activates translation of its target mRNAs . To gain more insight into the mechanism of Slf1p regulation of protein synthesis , we further examined its ribosome binding activity . Treating whole cell extracts with EDTA dissociates 80S monosomes and polyribosomes into 40S and 60S ribosomal subunits , and we noted that Slf1p-TAP and Sro9p both co-sediment with a small ribosomal subunit marker , Rps3p ( Fig . 4C ) . Similarly , in our TAP-IP experiments , interactions between each LARP and Rps3p appear resistant to RNase I treatment ( Fig . 5 , lanes + RNase ) . These studies suggest that both yeast LARPs interact with 40S ribosomal subunits in an RNA independent manner . To further examine Slf1p-ribosome association , we used a sucrose cushion assay , which is simpler than a full polysome analysis and useful for screening purposes . Here , cell lysates were resolved into light and heavy fractions on sucrose cushion gradients . In untreated cells , Slf1p is mainly present in the heavy ribosome associated fraction along with the majority of the 40S and 60S ribosomal subunit markers ( Fig . 7A ) . After treatment with hydrogen peroxide ( 0 . 4 mM , 15 min ) , although a significant proportion of the 40S and 60S ribosomal subunit markers shifted from the heavy ribosome-associated fraction into the lighter fraction , Slf1p is retained in the heavy fraction ( Fig . 7A ) . Coupled with Fig . 4 , we interpret these results as suggesting that , in this assay , ribosomes associated with mRNA remain in the heavy fraction while mRNA-free 80S monosomes are present in the lighter fraction . As a further proof that Slpf1p associates with actively translating ribosomes , cell extracts were treated with puromycin prior to separation using sucrose cushion assays [22] , [23] . Puromycin is an aminonucleoside antibiotic that causes premature chain termination during translation and the collapse of translating heavy polysomes . Puromycin caused a shift in the distribution of Slf1p from the heavy to light fractions in cell extracts from both control and peroxide treated cells ( Fig . 7B ) . Similarly , the initiation factor eIF4E was shifted from heavy to light fractions in response to puromycin treatment . This is consistent with Slf1p associating with actively translating ribosomes in both control and stressed yeast cells . A previous study identified the Slf1p LaM as necessary for mRNA binding [12] . Outside the LaM , Slf1p has no other recognised domains . Sequence alignments reveal two short regions towards the N-terminus with similarity to Sro9p and their homologs in other yeasts , along with two separate lysine and asparagine rich sequences ( Fig . 7C ) . To determine if the ribosome association of Slf1p similarly relies upon the LaM or if other regions are important , five truncation mutants of Slf1p were constructed deleting different regions of SLF1 ( Fig . 7C ) . In addition , a missense allele was created where four LaM residues key for RNA binding [24] are altered to alanines in the context of the full length protein ( LaM-PM ) . Each construct was C-terminally TAP tagged and introduced into slf1Δ cells as the sole source of Slf1p . The sucrose cushion assay was used to analyse the impact of each mutation on the interaction of Slf1p with the ribosome ( Fig . 7D and E ) . Deletion of the Asn-rich region between the amino terminus and the La motif ( ΔM ) had little effect on ribosome association . Deletion of the extreme C-terminus ( ΔC ) of Slf1p was difficult to interpret since it significantly reduced the levels of Slf1p and so was not considered further . In contrast , both N-terminal deletions ( ΔN , ΔN+ ) and La motif mutations ( ΔLaM , and LaM PM ) significantly decreased the level of Slf1p present in the heavy fraction compared with the corresponding wild type ( Fig . 7D and E ) . This experiment suggests that both of these regions are functionally important in maintaining Slf1p with mRNA-associated ribosomes . As a more robust test of our interpretations , the sedimentation of both ΔN+ and ΔLaM constructs were assessed across full polysome gradients . Their sedimentation patterns were significantly altered relative to that of the wild type protein , with a higher proportion co-sedimenting in light fractions away from the ribosomal material ( Fig . 8A ) . As a control , the ΔM deletion was observed to co-sediment with the translating ribosomes similar to wild type Slf1p ( Fig . 8A ) . These data indicate that there may be two ways that Slf1p can interact with the ribosome; through its N-terminal region or through the LaM . The LaM likely acts to promote the interaction of Slf1p with ribosomes indirectly via its interaction with mRNA [12] . We tested this idea by RNAse I treatment to disrupt polysomes and re-assessed the ribosomal-association of the Slf1p mutants ( Fig . 8B ) . Following RNase I treatment , wild type Slf1p remained associated with the resulting 80S ribosomes and ΔLaM had a modest impact on Slf1p ribosome association . In contrast , however , removal of the N terminal region ( ΔN+ ) significantly disrupted ribosome binding by Slf1p . Taken together with other data presented here these findings are consistent with idea that there are separable functional domains within Slf1p: the N terminus of Slf1p acting as a 40S ribosome binding domain , whereas the La motif facilitates mRNA interaction .
We set out to characterize the roles of the yeast LARPs Slf1p and Sro9p via an integrated set of post-genomic global analyses . These experiments have confirmed that both these homologous RNA-binding proteins have similar functional roles , with overlapping sets of mRNA targets that they bind and regulate in terms of abundance at the mRNA level , including many ribosomal proteins ( Fig . 1 and 2 ) . Slf1p target mRNAs are among the most actively translated mRNAs , identified by ribosome profiling ( Fig . 2D ) . This and other data strongly suggests Slf1p is a translational activator . Our experiments reveal that Slf1p has a critical role in mediating the coordinated cellular oxidative stress response to reactive oxygen species . Several lines of evidence show that Slf1p remains bound to actively translating mRNAs during oxidative stress ( Fig . 4 , 5 and 7 ) and that some of the stress mRNA targets encode many key antioxidant enzymes including thioredoxins , glutaredoxins and peroxiredoxins that are all critical to the cellular defence against hydrogen peroxide and whose expression is enhanced following stress ( Fig . 3 and 6 ) . Oxidative stress leads to a general down-regulation of protein synthesis initiation , caused by phosphorylation of eIF2 , as well as defects in the elongation phase of protein synthesis [15] . Yet , stress response proteins are apparently able to overcome this inhibition and increase or maintain their protein levels following stress by as yet unknown mechanisms . Our experiments offer one possible explanation , as they show that Slf1p plays a critical role in enhancing translation of many of these proteins , including many that are necessary for the cellular stress response ( Fig . 6 ) . As a consequence , slf1Δ cells are hyper-sensitive to hydrogen peroxide both in terms of growth and overall protein synthesis , as measured by polysome profiles ( Fig . 3 and 4 ) . Finally , we present evidence that Slf1p binding to the small ribosomal subunit is not solely dependent on the LaM , but instead optimally requires a novel motif within the N-terminal region of Slf1p ( Fig . 7 and 8 ) . Thus , we suggest that Slf1p acts as an adapter protein between specific mRNAs and the ribosome , promoting translation of key mRNAs during stress conditions by binding both 40S ribosomes ( via the N-terminal ribosome binding domain ) and specific target mRNAs ( via the LaM ) , with both domains critical for resistance to ROS . Notably , Sro9p is approximately six times as abundant as Slf1p according to Pax-DB [25] , which is reflected both by the increased number of its target mRNAs and the increased number of mRNAs whose levels are altered in its absence , although the change in abundance observed is generally less than two-fold . Intriguingly , and despite this , Slf1p has a greater impact on steady state mRNA levels of its targets than does Sro9p . Slf1p-target mRNAs are reduced in abundance in slf1Δ cells , while Sro9p does not significantly influence its target mRNA abundance . This provides further support to the idea that these LARPs are not functionally equivalent despite sharing many mRNA targets . Specificity may be achieved by binding other distinct partners . Distinctions between the yeast LARPs that we identified were that Sro9p ( i ) forms an RNA-dependent complex with Caf20p , while Slf1p does not , and , ( ii ) that Pab1p interaction with Sro9p appears less sensitive to RNase than does the Pab1p-Slf1p interaction In addition hydrogen peroxide treatment apparently reduced levels of eIF4G binding Slf1p . Possible implications for these observations are described below . Efficient translation of mRNAs involves capping of the 5′-end and polyadenylation of the 3′-end of the mRNA . The 5′ methyl cap is bound by eIF4E , the polyA tail is bound by multiple Pab1p proteins and eIF4G binds to both eIF4E and Pab1p forming a ‘closed-loop’ complex that is thought to promote translation [20] . As expected for factors promoting translation and interacting with ribosomes , Slf1p and Sro9p co-immunoprecipitate initiation factors which are part of the closed-loop complex , as well as components of the small and large ribosomal subunits . These interactions are largely RNA dependent . Caf20p competes for the eIF4G binding site of eIF4E , preventing the formation of the closed-loop complex and thus suppressing translation of certain mRNAs [26] , [27] . RNA-dependent co-purification of Caf20p with Sro9p may indicate that some of the mRNAs bound by Sro9p are not translationally active . In accord with this idea , a proportion of the Sro9p signal was found migrating in non-ribosomal fractions of polysome gradients . The Sro9p-Caf20p interaction is reduced following hydrogen peroxide stress although we do not know if this is significant . Similarly Sro9p-Pab1p interactions appear more resistant to RNase than Slf1p-Pab1p interactions . This may imply direct binding between Sro9p and Pab1p . We have not explored this possibility further , although studies of related proteins found similar interactions . The human LARPs 4 and 4b were also found to bind 40S subunits and PABP ( polyA binding protein ) [8] , [11] . LARP4 interacts with PABP through a PABP interaction motif 2 ( PAM2 ) found in its extreme N-terminus and which is shared with some other unrelated PABP interacting proteins and a second region downstream of the LaM and RRM domains [8] . It remains to be determined whether the continued mRNA-dependent interaction between the cap-binding complex factors and Slf1p under oxidative stress conditions , reflects Slf1p remaining bound to actively translating mRNAs . An alternative possibility is that both the cap-binding complex and Slf1p remain bound to repressed mRNAs during oxidative stress conditions since translation initiation can be blocked at several distinct steps , some of which lie downstream of the cap-binding complex . The interaction between Slf1p and eIF4G is diminished following oxidative stress . At present the significance of this observation also remains unresolved . Interactions between Slf1p and Pab1p/eIF4E are maintained , suggesting that at least some mRNAs bound by Slf1p may specifically lose eIF4G after hydrogen peroxide stress . Slf1p does not directly bind to eIF4E after hydrogen peroxide treatment , as this interaction remains RNA dependent , ruling out the possibility that Slf1p acts as a direct eIF4E binding protein . Our sucrose density gradient and immunoprecipitation data both clearly indicate that Slf1p and Sro9p associate largely with the 40S small ribosomal subunit in a manner that is resistant to EDTA and/or RNase treatment . In agreement with a previous study , the LaM [12] of Slf1p is responsible for mRNA binding , while here we identify a novel 40S binding domain in the Slf1p N-terminus that is shared with Sro9p and their close homolog's . This split in generic and specific recognition is not unprecedented in LARPs . The human LARPs 4 and 4b also have 40S ribosome interacting domains that are distinct to their LaM . The C-terminus of LARP4b was shown to interact with the 40S protein RACK1 [11] . LARP4 also binds to RACK1 [8] . RACK1 is located on the head of the 40S ribosomal subunit close to the mRNA exit channel [28] [29] . Therefore it is ideally placed to act as an adapter for RNA-binding proteins . In yeast RACK1 is called Asc1p and it is known to act as a ribosome binding site for the RBP Scp160p [30] and can regulate translation [31] . A recent study used genome-wide ribosome profiling to analyse the translational response to oxidative stress induced by hydrogen peroxide exposure [32] . This study provides translation efficiency ( TE ) data ( amount of footprint normalized to underlying mRNA abundance ) following treatments with 0 . 2 mM hydrogen peroxide for five or 30 minutes . We have compared this data with our RIP-Seq and proteomics analyses , which treated yeast cells with 0 . 4 mM hydrogen peroxide for 15 minutes . In order to investigate any possible association between the ribosome footprinting results and our RIP-Seq and proteomics analyses , following [32] , we classified mRNAs and proteins as being up , down or unchanged in our experiments . S7 Fig . shows that the distribution of TE values is not the same across both the transcript and protein abundance subsets ( Kruskal-Wallis test; FDR<0 . 01 ) . In particular , we found an enrichment for mRNAs that are significantly down in the RIP-Seq experiment and have low TE following 30-minute stress ( χ2 test; Bonferroni corrected p-value = 0 . 0009 ) . In addition , there is also an association between proteins with increased abundance in the proteomics experiment and mRNAs with increased TE in the ribosome footprinting experiment after 5-minute stress ( χ2 test; Bonferroni corrected p-value = 0 . 044 ) , and 30-minute stress ( χ2 test; Bonferroni corrected p-value = 0 . 034 ) . In summary , the main conclusions from comparing these datasets are: 1 ) that being an Slf1p mRNA target does not increases TE under stress conditions , but protects against a decrease in TE; 2 ) that this effect does not appear to be immediate; and , 3 ) that mRNAs that have an increased TE under stress conditions , tend to have increased protein production . It was recently shown that human LARP1 is necessary to enhance translation of 5′ TOP mRNAs [33] . 5′TOP mRNAs are an abundant class of mRNAs in mammalian cells that include many ribosomal protein and translation factor mRNAs . Each mRNA possesses an oligo pyrimidine sequence at or near their 5′ termini . It was proposed that LARP1 specifically promotes expression of 5′TOP mRNAs . Yeast ribosomal RNAs do not possess 5′TOP sequences , so our findings here that the yeast LARP Slf1p promotes translation of ribosomal proteins implies that there may be more than one mechanism for LARPs to promote ribosomal protein synthesis , and suggests that both human and yeast LARPs function in similar ways . It is also interesting that these proteins share some functional parallels with the eubacterial ribosomal protein S1 . Similar to the LARPs , S1 is a large protein ( ∼68 KDa ) that interacts with the small ribosomal subunit of ribosomes . S1 also binds single stranded mRNA including to a subset of mRNA 5′ leaders and can promote mRNA-ribosome interactions that activate translation initiation [34] . Analysis of the Slf1p target mRNAs has identified an AU repeat motif in the 3′UTR of Slf1 target mRNAs . We used MEME to search for motifs within the 3′ and 5′UTR regions of the mRNAs that were significantly enriched in the unstressed Slf1p RIP-Seq experiment ( FDR<0 . 05 ) . The UTR regions of mRNAs that were significantly decreased in the RIP-seq experiment were used as a negative control set and were not found to contain any enriched motifs . Nothing was identified for the 5′UTR region , but a 21-nt motif ( mostly , alternate As and Us ) was identified in the 3′UTR ( Fig . 3D ) . This motif is not found in all sequences , but in 112 out of 414 . This motif is very similar to a previously identified Pub1p motif [35] which is also an alternating AU element in the 3′UTRs of mRNAs bound by Pub1p . There was no significant overlap between our Slf1p target mRNAs to those of Pub1 [35] . A larger set of Pub1 targets have been described [36] and when comparing these to our Slf1 mRNA targets there is a significant overlap of 178 mRNAs ( p = 10−4 ) . However , within this overlap there is an under enrichment for the motif suggesting that the presence of the motif in the Slf1p target mRNAs is not due to shared target mRNAs with Pub1p . The motif persists in Slf1 mRNA targets after hydrogen peroxide treatment ( Fig . 3D ) . The number of Slf1 target mRNAs containing the motif increases after oxidative stress and therefore the presence of the motif in this dataset is not simply due to Slf1p continuing to bind those target mRNAs that Slf1 binds under unstressed conditions . The physiological importance of this motif is unknown at present and will form the basis of future studies . Taken together , these data indicate that Slf1p plays a role in mRNA-specific regulation of translation during oxidative stress conditions and is necessary to promote the translation of stress-responsive mRNAs . It does this via mRNA interactions with the well-characterised LaM [12] , [24] and a novel 40S ribosome interaction region defined here . Given that Slf1p is only one of the 600 yeast proteins that are predicted to bind RNA [5] it is likely that many other RBPs will add to the complexity of mRNA-specific translational control .
BY4741 was the parental strain for all deletion mutants . A BY4741 HIS3+ strain was generated for use as the parental strain for TAP tag immunoprecipitations . This was generated by replacing the his3Δ1 allele in BY4741 with HIS3 . Slf1p-TAP and Sro9p-TAP tagged strains were obtained from Open Biosystems . The BY4741 slf1Δ mutant was generated by replacing SLF1 with a KanMX cassette using standard yeast genetic techniques . BY4741 sro9Δ was obtained from Euroscarf . All strains were grown in SCD media at 30°C to exponential phase ( OD600 0 . 5–0 . 7 ) . Cultures were exposed to 0 . 4 mM hydrogen peroxide for 15 or 60 minutes to induce oxidative stress . Wild-type SLF1 and SLF1 with point mutations within the La motif , both containing a C-terminal TAP tag , were synthesised ( Epoch Life Sciences ) . SLF1 constructs contained 289 nt upstream of the ATG and 234 nt downstream of the stop codon and were cloned into plasmid pRS416 . pRS416-SLF1 was used as a template to generate truncation mutations [37] . Polyribosomal profiling was performed as previously described [38] . Briefly , S . cerevisiae was grown to OD600∼0 . 7 , cycloheximide was added to a final concentration of 0 . 1 mg/ml and yeast cells were harvested by centrifugation . When cells were stressed with hydrogen peroxide , cultures were split into two 50 ml cultures and one of these was treated with hydrogen peroxide and incubated at 30°C for 15 minutes . S . cerevisiae were lysed in polyribosomal buffer containing cycloheximide and 2 . 5 OD260 units were loaded onto a sucrose gradient . 15–50% sucrose gradients were poured as previously described [38] . 5–25% sucrose gradients were poured in six separate fractions increasing in 5% sucrose intervals from 5–25% sucrose . For RNAse treatment , 12 units of RNAse I was added to polysome extracts and incubated for 1 h at 21°C prior to loading onto a sucrose gradient . Sucrose cushion gradients were performed as previously described [39] . For puromycin treatment , extracts were prepared in the absence of cycloheximide , and incubated with 1 mg/ml Puromycin for 10 minutes on ice prior to loading onto gradients TAP tagged strains were grown to exponential phase , centrifuged and washed in 3% glucose and 2x amino acids and snap frozen in liquid nitrogen . Yeast were lysed in L Buffer ( 20 mM Tris-HCl pH 8 , 140 mM NaCl , 1 mM MgCl2 , 0 . 5% NP40 , 0 . 5 mM DTT , 1 mM PMSF , EDTA free Protease Inhibitor cocktail tablet ( Roche ) , NaV3O4 , NaF and 40 units/ml RNAsin ) using a 6870 Freezer mill ( Spex ) . Lysates were cleared by centrifuging twice at 15 , 000 g . Beads were prepared as previously described [40] . Beads were pre-washed three-times with L Buffer and then added to 4 mg/ml of grindate . Immunoprecipitations were performed for 20 minutes at 4°C and washed five times with L buffer containing 10 units/ml RNAsin , changing tubes at least twice during the washes and the final two washes were performed for 15 minutes each . Where RNAse treatment was performed , RNAsin was omitted from the L buffer and 200 units of RNAse I was added during the 20 minute immunoprecipitation . For RNA isolation after the final wash , the beads were resuspended in 250 µl L Buffer and treated with Trizol . The aqueous phase was mixed with 70% ethanol and the RNA was purified using the RNeasy minikit ( Qiagen ) . For protein isolation after the final wash , protein was eluted from beads using 0 . 5 M sodium hydroxide . Eluted protein was concentrated using Amicon concentrator columns and analysed by immunoblotting . Once isolated , all RNA samples were processed in an identical manner . rRNA was depleted from the RNA samples using the Ribominus Eukaryote Kit for RNA-Seq ( Invitrogen ) . Total RNA samples were normalised to the amount of RNA isolated from the corresponding IP sample . Depleted samples were precipitated with 2 . 5x volumes ethanol , 1/10th volume 3 M sodium acetate and 1 µl glycogen , washed twice with 70% ethanol and re-suspended in 10 µl DEPC water . rRNA depletion was checked on a 2100 Bioanalyzer ( Agilent Technologies , Palo Alto , CA ) using an RNA nano-chip and the remaining RNA stored at -80°C . Sequencing libraries were generated using the whole Transcriptome Library preparation protocol provided with the SOLiD Total RNA-Seq Kit . Briefly , rRNA depleted samples were fragmented using RNase III , and subsequently cleaned up using the RiboMinus Concentration Modules ( Invitrogen ) . Fragmentation was assessed on a 2100 Bioanalyzer using the RNA pico-chip . Fragmented RNAs were reverse transcribed and size selected on a 6% TBE-Urea gel ( Novex ) , selecting for 150–250 nt cDNA . cDNA was then amplified and barcoded with the SOLiD RNA barcoding Kit . Samples were subsequently purified using the PureLin PCR Micro Kit ( Invitrogen ) and assessed on a 2100 Bioanalyzer using the High Sensitivity DNA chip . Samples were sequenced on an ABI SOLiD 4 at either The University of Manchester or at BGI . Reads were mapped to the S . cerevisiae genome ( genome assembly EF4 downloaded from ENSEMBL ) using Bowtie; sequences were then assigned to genomic features using HTseqcount ( mapping against the corresponding EF4 GTF file ) . Statistical significant enrichments of transcripts in the protein IPs relative to TAP-tag whole extracts were determined using the Generalized Linear Model ( GLM ) functionality within edgeR to produce a comparison with a paired statistical design [41] and generate gene lists at a FDR<0 . 05 . In addition , the GLM functionality was used to measure protein specific variance between experiments through the use of an interaction model [42] . Fold changes are presented as log2 ratios of the reads per million counts ( transcripts with fewer than twenty reads in each of the pertinent total extract samples were excluded from the plots ) . Sequencing data are publicly available on ArrayExpress , E-MTAB-2567 ( Slf1p ) and E-MTAB-2568 ( Sro9p ) . Functional categorisation of mRNAs and proteins was performed using MIPS Functional Catalogue ( mips . helmholtz-muenchen . de/proj/funcatDB/ ) . 5′ and 3′ UTR regions of mRNAs bound by Slf1p in the presence or absence of hydrogen peroxide were searched for common sequence motifs using the MEME Suite [43] . In all cases , the equivalent regions of the depleted mRNAs in the RIP-seq experiment were used as a negative set . In an additional control , the letters in the positive set were shuffled in order to check that the motif did not come out because of the relative frequencies of nucleotides . The parental and mutant strains were grown in SCD to exponential phase and treated as described above . RNA was isolated from cleared lysates using Trizol and used to generate sequencing libraries . To enable a comparison of the transcriptomes of both mutants , transcriptomes were binned into 0 . 25-fold bins based on fold enrichment above the parental strain . These data were then expressed as a percentage frequency of transcripts within each bin . Data were not filtered on FDR prior to binning . Quintuplicate repeats of the wild-type and slf1Δ strains were grown in SCD media to exponential phase , split in two , and half treated with 0 . 4 mM hydrogen peroxide for 1 h . Cultures were harvested , washed in 3% glucose with 2x amino acids and snap frozen in liquid nitrogen . Cell pellets were lysed using the 6870 freezer mill ( Spex ) into 8 ml of 25 mM ammonium bicarbonate buffer containing a protease inhibitor cocktail tablet ( Roche ) . Ground samples were defrosted , cleared by centrifugation ( 15 , 000 g 10 minutes ) , and 100 µg of cleared lysate was diluted to a final volume of 160 µl containing 1% ( w/v ) RapiGest ( Waters Corporation ) . Samples were incubated at 80°C for 10 minutes , reduced using a final concentration of 3 . 5 mM DTT in 25 mM ammonium bicarbonate and incubated at 60°C for 10 minutes . Iodoacetamide was added to a final concentration of 10 mM and incubated at room temperature for 30 minutes . A final concentration of 0 . 01 µg/µl trypsin in 10 mM acetic acid was added and samples were digested for 4 . 5 h at 37°C . Hydrochloric acid was added to a final concentration of 13 mM and a second identical trypsin digest was performed overnight at 37°C . 0 . 5 µl of trifluoroacetic acid was added and incubated at 37°C for 2 h . 7 . 5 µl of acetonitrile:water ( 2∶1 ) was added and incubated at 4°C for 2 h and centrifuged at 13 , 000 g for 15 minutes . Supernatant was removed and desalted using OLIG R3 reversed-phase media on a microplate system . Peptides were eluted in three cycles of 50% acetonitrile and dried by vacuum centrifugation , and reconstituted to 10 µL with 5% acetonitrile and 0 . 1% formic acid . Digested samples were analysed by LC-MS/MS using an UltiMate 3000 Rapid Separation LC ( RSLC , Dionex Corporation , Sunnyvale , CA ) coupled to an Orbitrap Elite ( Thermo Fisher Scientific , Waltham , MA ) mass spectrometer . Peptide mixtures were separated using a gradient from 92% A ( 0 . 1% FA in water ) and 8% B ( 0 . 1% FA in acetonitrile ) to 33% B , in 44 min at 300 nL min−1 , using a 250 mm×75 µm i . d . 1 . 7 mM BEH C18 , analytical column ( Waters ) . Peptides were selected for fragmentation automatically by data dependant analysis . The acquired MS data from the five replicates were analysed using Progenesis LC-MS ( v4 . 1 , Nonlinear Dynamics ) . The retention times in each sample were aligned using one LC-MS run as a reference , then the “Automatic Alignment” algorithim was used to create maximal overlay of the two-dimensional feature maps . Features with charges ≥+5 were masked and excluded from further analyses , as were features with less than 3 isotope peaks . The resulting peaklists were searched against the Saccharomyces Genome Database ( SGD , version 3rd February 2011 ) proteome using Mascot v2 . 4 ( Matrix Science ) . Search parameters included a precursor tolerance of 5 ppm and a fragment tolerance of 0 . 5 Da . Enzyme specificity was set to trypsin and one missed cleavage was allowed . Carbamidomethyl modification of cysteine was set as a fixed modification while methionine oxidation was set to variable . The Mascot results were imported into Progenesis LC-MS for annotation of peptide peaks . The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium ( http://www . proteomexchange . org ) via the PRIDE partner repository with the dataset identifier PXD000887 and DOI 10 . 6019/PXD000887 . | All organisms must respond to changes in their external environment such as exposure to different stresses . The availability of genome sequences and post-genomic technologies has enabled the analysis of these adaptive responses at the molecular level in terms of altered gene expression profiles . However , relatively few studies have focused on how cells regulate the translation of mRNA into protein in response to stress , despite its fundamental role in gene expression pathways . In this study , we show that a previously identified RNA-binding protein called Slf1p plays a major role in mRNA-specific regulation of translation during oxidative stress conditions and is necessary to promote the translation of stress-responsive mRNAs . This protein is a member of the so-called “La-related” family of proteins that have not been well characterized , although they are conserved throughout evolution . Exposure to oxidants is known to cause a general down-regulation of protein synthesis , although many stress response proteins are able to overcome this inhibition and increase their protein levels following stress by as yet unknown mechanisms . Our experiments offer one possible explanation , as they show that Slf1p plays a critical role in enhancing translation of many of these proteins , including many that are necessary for the cellular stress response . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry",
"rna",
"cell",
"biology",
"oxidative",
"stress",
"gene",
"expression",
"genetics",
"biology",
"and",
"life",
"sciences",
"proteomics",
"molecular",
"cell",
"biology",
"oxidative",
"damage",
"antioxidants",
"ribosomes"
] | 2015 | The Yeast La Related Protein Slf1p Is a Key Activator of Translation during the Oxidative Stress Response |
Arf4 is proposed to be a critical regulator of membrane protein trafficking in early secretory pathway . More recently , Arf4 was also implicated in regulating ciliary trafficking , however , this has not been comprehensively tested in vivo . To directly address Arf4’s role in ciliary transport , we deleted Arf4 specifically in either rod photoreceptor cells , kidney , or globally during the early postnatal period . Arf4 deletion in photoreceptors did not cause protein mislocalization or retinal degeneration , as expected if Arf4 played a role in protein transport to the ciliary outer segment . Likewise , Arf4 deletion in kidney did not cause cystic disease , as expected if Arf4 were involved in general ciliary trafficking . In contrast , global Arf4 deletion in the early postnatal period resulted in growth restriction , severe pancreatic degeneration and early death . These findings are consistent with Arf4 playing a critical role in endomembrane trafficking , particularly in the pancreas , but not in ciliary function .
Primary cilia are microtubule-based organelles , which perform sensory functions important for health and development in vertebrates . Severe defects in primary cilia lead to embryonic lethality . Mild defects cause a wide range of syndromic diseases , termed ciliopathies , which manifest as a spectrum of features including obesity , retinal degeneration , cerebral anomalies and renal disease [1] . Each cilium contains a set of specific proteins , some shared across cell types and others adapted to perform unique cell-specific functions . This specialization relies on robust intracellular trafficking mechanisms whose malfunctions underlie a variety of pathological conditions [2 , 3] . These mechanisms remain a subject of active investigation , with many proteins being delivered into the cilium through intraflagellar transport ( IFT ) [4–7] . Less is known about the molecular players responsible for designating proteins for ciliary trafficking upon their exit from the biosynthetic membranes . Many ciliary proteins contain short sequences used for their specific targeting [8 , 9] . One of the better studied cases is the visual pigment rhodopsin , which contains a four amino acid ( V[S/A]PA ) targeting sequence at its C-terminus known as the VXPX motif [10 , 11] . Human patients containing mutations in either the V or P residues exhibit autosomal dominant retinitis pigmentosa [12] , and deletion of these residues in mice results in rhodopsin mislocalization followed by photoreceptor cell death [13–15] . These residues were also found to be important in an in vitro assay for the formation of rhodopsin carrier vesicles in the Golgi [16] . A similar RVXP motif is present in other ciliary-localized proteins such as the CNGB1b subunit of the olfactory CNG channel , polycystin-1 , polycystin-2 and presenlin-2 [17–20] . The P-to-A replacement in the VXPX motif of presenilin-2 was also shown to mislocalize this protein from the basal body of epidermal suprabasal cells [18] . The work by Deretic and colleagues suggested that the C-terminal VXPX sequence in rhodopsin is recognized by Arf4 [21] . Arf4 is a small GTPase regulating protein trafficking in the early secretory pathway . It is typically localized in the ER/Golgi intermediate compartment and cis-Golgi [22 , 23] . Deretic and colleagues hypothesized that , in addition to this well-established function , Arf4 directs ciliary trafficking of rhodopsin in trans-Golgi . Most importantly , they showed that Arf4 interacts with rhodopsin’s cytoplasmic C-terminus in vitro [21] and that expression of a dominant-negative Arf4 mutant in frog rods causes a partial rhodopsin mislocalization from the cilium [16] . Subsequent work suggested that Arf4 is required for trafficking other ciliary proteins , including polycystins and fibrocystin [17 , 24] . These findings implicated Arf4 as a key player in sorting transmembrane proteins to the cilium , suggesting that its malfunction or loss would lead to human diseases such as retinal degeneration and polycystic kidney disease . To address the functional role of Arf4 in vivo , an Arf4 knockout mouse was generated , but this mutation resulted in embryonic lethality between days 9 and 10 [24] . Although embryonic lethality occurs in mice with severe defects in ciliary assembly , the embryonic node of the Arf4 knockout mouse had normal cilia , which were functional since all embryos broke left-right symmetry properly and formed a D-looped heart . In wild type mice , Arf4 is highly expressed in the visceral endoderm starting at embryonic day 7 . In Arf4 mutants , the microvilli and lysosomes of the visceral endoderm cells were disrupted and the localization of the endocytic receptor megalin was altered . Since the visceral endoderm is the major secretory and absorptive tissue of the developing embryo it is likely that these defects are the cause of lethality in the Arf4 knockout mice . This is consistent with the established function of Arf4 in mediating ER/Golgi protein trafficking , but does not imply a defect in ciliogenesis as no cilia were detected on these cells . One approach to circumvent the embryonic lethality of germline Arf4 knockout , reported in a recent study [18] , was to knockdown Arf4 using a lentiviral shRNA infection at E9 . 5 . The authors investigated the skin phenotype of these animals and observed presenilin-2 mislocalization from the basal body of the epidermal cells , defects in Notch signaling , and polydactyly . Here , we used a more straightforward approach to assess the role of Arf4 in adult tissues by generating a floxed Arf4 mouse . To our surprise , Arf4 knockout from photoreceptors did not affect rhodopsin localization or photoreceptor morphology , and Arf4 knockout from kidney did not affect ciliogenesis or cause cystic disease . In contrast , the Arf4 knockout caused severe degeneration of the exocrine pancreas , consistent with Arf4 playing a critical role in endomembrane trafficking but not ciliary function .
To generate a conditional Arf4 knockout mouse ( Arf4flox ) , we floxed exons 2 and 3 that encode the functionally indispensable switch 1 and switch 2 regions of Arf4 . Cre recombination results in a frame shift after residue 22 and early termination 6 residues downstream from this residue ( Fig 1A ) . While this peptide cannot retain the function of a small GTPase , N-termini of Arf proteins have been shown to have inhibitory activities in vitro [25 , 26] . However , if there was such activity remaining in our mouse , it would be expected to manifest as a dominant-negative effect and we did not observe any evidence for this . Mice homozygous for the Arf4flox allele were viable and showed no detectable phenotypes . To understand the role of Arf4 in the postnatal mouse , we used the tamoxifen-inducible CagCreER driver to delete Arf4 . This Cre is broadly expressed but remains inactive until mice are treated with tamoxifen [27] . Postnatal day 2 ( P2 ) mice were treated with tamoxifen and genotyped on P10 . The viability of animals bearing the Arf4flox/CagCreER genotype at P10 was consistently lower than expected from Mendelian distribution , indicating an important role for Arf4 in the early postnatal period . The survivors were notably smaller than littermates ( Fig 1B ) and had a propensity to die at random times . Necropsy was unremarkable in most cases except for a notable reduction in the size of the pancreas and yellowish feces in the lower intestine . The hair on animals that survived past 6 weeks turned from black to grey ( Fig 1C ) . Since the first anagen phase of the hair cycle occurs at about P28 [28] the hair coming in at this time would have been formed after tamoxifen treatment , thus indicating a role for Arf4 in hair pigmentation . This phenotype is in line with the report by [18] that Arf4 plays an important role in skin cells . The major cellular phenotype observed in the germline Arf4 knockout mouse was a malformation of the microvilli on the visceral endoderm . Instead of being straight and having a uniform diameter , the mutant microvilli were curved and bulbous [24] . To determine if Arf4 is generally involved in microvilli formation or maintenance , we examined kidney proximal tubules and intestinal epithelia by transmission electron microscopy as these cell types have extensive brush borders . In both organs , the microvilli of the experimental animals appeared normal and had none of the structural defects observed in the visceral endoderm of the germline knockout ( S1 Fig ) . To determine the penetrance of the conditional Arf4 knockout in specific tissues and to address the cellular localization of Arf4 in wild type animals , we generated an antibody against a peptide corresponding to the residues 98–114 of mouse Arf4 . This sequence has the lowest homology to other Arf family members . This antibody recognized a 17 kDa band corresponding to the endogenous Arf4 in wild type mouse embryonic fibroblasts ( MEFs ) , which was absent in Arf4-/- MEFs ( Fig 2A ) . To further establish antibody specificity across the members of Arf protein family , we transiently transfected mIMCD3 cells with all six Arf proteins bearing a Flag-tag and found that only endogenous Arf4 and its Flag-tagged recombinant protein were recognized in Western blots of lysed cells ( Fig 2B ) . Immunofluorescence analysis of wild type MEFs indicated that Arf4 localizes to the Golgi complex and has no association with the cilium ( Fig 2C ) . Arf4-/- fibroblasts displayed a normal fraction of ciliated cells and normal ciliary length , indicating that Arf4 is not critical to ciliogenesis ( Fig 2D and 2E ) . To identify which compartment of the Golgi was labeled by Arf4 , cells were co-stained for Arf4 and either the cis-medial Golgi marker HPA or the trans-Golgi marker Golgin97 . Overlap between the cis-medial marker and Arf4 was extensive , while the trans-Golgi marker labeled structures next to the Arf4 staining with minimal overlap ( Fig 2F–2I ) . The Arf4 antibody also produced punctate labeling in the cytoplasm , but identical staining was observed in Arf4-/- cells suggesting that it represents nonspecific staining ( Fig 2C , 2F and 2H ) . These results indicate that the primary localization of Arf4 in ciliated cells is in the cis-medial Golgi rather than trans Golgi or cilium , a conclusion consistent with localization of GFP-tagged Arf4 [29] . Arf4 involvement in ciliary protein trafficking was first described for the visual pigment , rhodopsin . Therefore , Arf4 knockout was predicted to affect rhodopsin localization and likely photoreceptor health , as rhodopsin mislocalization is associated with retinal degeneration [13 , 15] . To test this prediction , we used inducible Arf4 conditional knockout mice ( Arf4flox/CagCreER ) to analyze rhodopsin localization and photoreceptor morphology . We used quantitative Western blotting to measure the relative amounts of Arf4 in the eyecups of experimental and control animals . Serial dilutions of eyecup extracts were analyzed on the same blot ( Fig 3A ) , the densities of Arf4 bands were measured , plotted as a function of total protein amount and fitted with straight lines . The extent of Arf4 reduction in experimental animals was calculated from the ratio between these fits ( Fig 3B ) . This analysis showed that Arf4 in Arf4flox/CagCreER eyecups was reduced to ~13% of control ( Fig 3C ) , indicating that the penetrance of tamoxifen induction of the knockout was highly efficient . Considering that activation of Cre recombinase in a given cell results in a complete and irreversible deletion of the Arf4 gene , this level of protein reduction indicates that Arf4 was completely lost from the majority of cells . Therefore , the small amount of remaining Arf4 originated from cells where Cre failed to delete the gene . Such a mosaic induction pattern is typical for this tamoxifen-inducible model [27] . Arf4 immunostaining of retinal cross-sections from control animals showed a complex staining pattern with strong signals found in all retinal layers ( Fig 3D , WGA used to label outer segments and retinal plexiform layers ) . A side-by-side staining of cross section from Arf4flox/CagCreER animals revealed a complete loss of Arf4 from the inner segments of photoreceptor cells , which are localized above photoreceptor nuclei and are partially co-stained with the Golgi marker , GM130 ( Fig 3D ) . On the other hand , the Arf4 staining at the base of the outer segment layer was preserved ( Fig 3D and 3E ) . Considering that outer segments start forming at least one-week post tamoxifen injection and are completely renewed every 12 days , we conclude that Arf4 staining at this location is non-specific . This is consistent with the anti-Arf4 antibody cross-reacting with several additional bands on Western blots obtained from eyecup extracts ( S2 Fig ) . Based on these observations , we conclude that the degree of Arf4 knockout in photoreceptors was essentially complete . An overall reduction of Arf4 staining was observed in the inner retina as well , although residual signal remained significant at various locations indicating either non-specific staining or resistance of certain cells to tamoxifen treatment ( Fig 3D ) . Regardless , these cells were not the subject of our investigation . We next investigated whether Arf4 knockout in photoreceptors affects rhodopsin localization by immunostaining retinal cross-sections with an anti-rhodopsin antibody . To our great surprise , rhodopsin was normally localized to rod outer segments of Arf4flox/CagCreER animals . No difference was observed between experimental and control animals even when the rhodopsin signal was grossly oversaturated ( Fig 3F ) . Consistent with normal rhodopsin localization , we did not observe any morphological abnormalities in the photoreceptor layer of Arf4flox/CagCreER mice , which was further documented using thin plastic retinal cross-sections ( Fig 3G ) . Given that our result contradicted the central paradigm in rhodopsin trafficking that Arf4 is an indispensable player in this process , we replicated this result by employing an alternative strategy to specifically knock out Arf4 from rod photoreceptor cells . We crossed Arf4flox mice with iCre75 mice , which express Cre recombinase under control of the rhodopsin promoter [30] . Because Arf4 is expressed in multiple cell types of the retina , Western blot of retinal extracts could not be used to assess the efficiency of Arf4 knockout in rods . Therefore , we resorted to the more sophisticated technique of serial sectioning with Western blotting [31] . We obtained individual 20 μm-thick tangential sections through the entire depth of a flat-mounted frozen retina and determined their Arf4 contents using an anti-Arf4 antibody . Specific sections were assigned to individual photoreceptor layers using two protein markers , peripherin localized in the outer segment [32] and phosducin localized throughout the entire photoreceptor cytoplasm [33] . As shown in Fig 4A , Arf4 was significantly reduced in sections representing the photoreceptor layer , whereas its content in the inner retina was well-preserved . Interestingly , the overall content of Arf4 in the inner retina of control animals was higher than in the photoreceptor layer . The knockout was further confirmed by a gross reduction of Arf4 immunostaining in photoreceptor inner segments ( Fig 4B ) . One difference from the Arf4flox/CagCreER mice was that the Arf4 signal was preserved in a small subset of photoreceptors , which is predicted because a small fraction of photoreceptors are cones not expressing Cre recombinase in this mouse . The outer segment layer staining with anti-Arf4 antibody was also present , reconfirming its non-specificity . Normal staining was observed with the Golgi marker , GM130 ( Fig 4C ) . The fact that rod-specific Arf4 knockout did not affect mouse health , allowed us to assess rhodopsin localization in older animals than in the Arf4flox/CagCreER line . No evidence of rhodopsin mislocalization was observed in mice up to 4 months of age ( Fig 4D ) . Similarly , the conditional Arf4 knockout did not affect the morphology of the photoreceptor layer ( Fig 4E ) . Taken together , the results from the two conditional Arf4 knockout lines provide compelling evidence that Arf4 is entirely dispensable for rhodopsin trafficking and its presence is not necessary for maintaining photoreceptor health . Ward et al . [17] showed that knockdown of Arf4 in cell culture reduces steady state levels of ciliary polycystin-1 , and we found a reduced rate of fibrocystin delivery to cilia when Arf4 was knocked down [24] . These results suggest that the lack of Arf4 should produce a cystic kidney phenotype . To test this , we analyzed kidneys from Arf4flox/CagCreER mice injected with tamoxifen at P2 ( Fig 5A–5F ) . In CagCreER mice , Cre is expressed in all segments of the tubule [27] and previous work has shown that deletion of cilia genes prior to ~P14 leads to rapid cyst formation [34 , 35] . To ensure that this protocol is capable of driving cyst formation , the same regiment was used with Pkd2flox/CagCreER mice and Ift20flox /CagCreER . Severe cystic disease developed by P14 in Pkd2flox/CagCreER mice and by P21 in Ift20flox /CagCreER mice ( Fig 5G–5L ) , indicating that this protocol is sufficient to induce cysts in susceptible mice . Arf4flox/CagCreER treated mice were collected at various ages between P10 and P91 ( see Fig 1B ) . Western blot analysis showed that Arf4 was reduced to about 15% of normal in the experimental kidneys ( Fig 5E and S3 Fig ) , indicating that the majority of kidney cells lacked Arf4 . Kidney to body weight was slightly larger in the experimental animals ( Fig 5F ) but no evidence of cystic enlargement in any portion of the tubule was observed ( Fig 5A and 5B ) . A small expansion of the renal pelvis , or hydronephrosis , was observed in the experimental animals ( Fig 5A ) . This may be enough to account for the slightly larger kidney to body weight . However , the experimental animals were undersized , which could also account for this difference if not all organs were equally growth-retarded . No defects were observed in the cilia of experimental mice ( Fig 5C ) and the fractions of ciliated proximal tubule cells were also equal in both groups ( Fig 5D ) . Since the failure to observe cyst formation in our Arf4flox/CagCreER animals was unexpected , we used HoxB7-Cre as a second method to delete Arf4 ( Fig 6 ) . HoxB7-Cre is active in kidney collecting ducts [36] . Previous work showed that HoxB7-Cre-driven deletion of cilia genes results in severe cystic kidney disease with kidneys becoming ~10 fold larger than normal by P21 [37 , 38] . Kidneys from Arf4flox/HoxB7-Cre experimental and control mice were collected at P21 , P175 and P365 ( Fig 6A–6E ) . Western blot analysis showed that Arf4 in experimental papilla was reduced to about 40% of normal ( Fig 6F and S3 Fig ) . Because the papilla is composed of about 50% collecting duct cells ( where Arf4 was deleted ) , with the rest being represented by cells from the thin loops of Henley , the interstitium and the vasa recta capillary bed , we conclude that the knockout penetrance was extremely deep . No evidence of cystic expansion of the collecting ducts was observed by histology in any experimental animals ( Fig 6A and 6B ) . Furthermore , the ratio of kidney to body weight was not significantly different between the control and experimental mice at any age ( Fig 6D ) . Normally , the kidney weighs about 1 . 5% of the total body weight , but the deletion of Ift20 by the same Cre driver increases this number to 16% at P21 [37] . In addition , no defects in cilia were observed ( Fig 6C ) and equal percentages of collecting duct cells were ciliated in both groups ( Fig 6E ) . To ensure that the HoxB7-Cre driver used for these experiments was active , we crossed the mTmG Cre reporter [39] into the line . This reporter expresses red fluorescent protein that is converted to green fluorescent protein upon Cre recombination . Control animals lacking Cre contained no GFP-positive collecting ducts while the vast majority of collecting duct cells were GFP positive in the experimental animals , which expressed HoxB7-Cre ( Fig 6G and 6H ) . Overall , our data indicate that Arf4 is dispensable for maintaining normal kidney structure and its loss does not lead to cystic kidney disease . During necropsy of Arf4flox/CagCreER animals we noted that their pancreas was abnormally small ( Fig 7A ) . In addition , experimental pancreases were opaque and had islet-sized spheres visible within the tissue , suggesting exocrine pancreas abnormality . The amount of Arf4 remaining in these pancreases was estimated to be 27% of normal at P10 and 45% at P28 ( Fig 7B and S4 Fig ) . The apparently smaller Arf4 reduction than in the kidney and Arf4’s increase with age can be explained by ongoing degeneration of cells in which Cre recombinase was induced . Accordingly , the fraction of non-induced cells in these pancreases increases with age . Histological examination revealed that islets in experimental pancreas appear normal but the surrounding exocrine tissue is reduced in volume , fragmented and partially replaced by what appears to be fibrotic material and adipocytes ( Fig 7C ) . Immunofluorescence with endocrine markers confirmed that the experimental islets were comparable to the control islets and both had similar distributions of alpha and beta cells ( Fig 7D ) . Trichrome blue , which labels collagen , normally highlights the tunicae surrounding the vasculature of the pancreas . This staining was observed in both experimental and control animals , but the experimental animals also showed extensive labeling within the field of secretory acini suggestive of developing fibrosis ( Fig 7E ) . H&E staining suggested that adipocytes may be interspersed within the experimental pancreas ( Fig 7C ) . The exocrine pancreas is associated with a fat pad but the adipocytes are normally only found around the edges of the pancreas and not infiltrated within the field of acini . Accordingly , perilipin-A ( a protein located on the surface of lipid droplets within adipocytes ) is essentially absent from the control exocrine pancreas . In contrast , perilipin-A-positive cells are found throughout the field of acini in the Arf4 deleted pancreas ( Fig 7F ) . Fatty pancreas or pancreatic steatosis can be caused by infiltration of adipocytes or transdifferentiation of exocrine cells into adipocytes [40 , 41] . Infiltration appears to be driving the fat formation within the pancreas when Arf4 is absent as we did not observe cells that were positive for both perilipin-A and exocrine markers . However , the degenerating Arf4-defective cells lose exocrine markers early in degeneration and so it is possible they are transdifferentiating but do not simultaneously express both zymogen granule markers and lipid droplet markers . To better understand the pancreatic pathology in Arf4 knockout animals , sections from the Arf4flox/CagCreER pancreases were examined by electron microscopy . Consistent with light microscopy observations , the islets of the endocrine pancreas appeared unaffected . The alpha and beta cells looked normal with abundant glucagon and insulin granules inside ( Fig 8A ) . In contrast , the acinar cells of the exocrine pancreas were variably affected in knockout mice and surrounded by adipocytes with large intracellular lipid droplets ( Fig 8B ) . In control animals , acinar cells are densely populated with spherical electron-dense zymogen granules , up to 1 micron in diameter , located near the apical end of the cell surrounding the central duct . In experimental animals , some acinar cells looked normal while others showed vacuolization of the zymogen granules ( Fig 8B ) . Electron dense granules in these cells were smaller than normal and associated with larger translucent 0 . 25 to 1 micron spheres .
In this study , we directly addressed whether Arf4 is required for rhodopsin trafficking by knocking it out from photoreceptors using two independent Cre-lox systems . We found that neither mouse line displayed even the slightest abnormality in subcellular rhodopsin localization , demonstrating that Arf4 is entirely dispensable for rhodopsin processing by the biosynthetic membranes , outer segment targeting and delivery . Consistent with normal rhodopsin trafficking , photoreceptors of knockout mice displayed no structural abnormality or signs of degeneration . One explanation for our negative result is that rhodopsin trafficking in photoreceptors follows an Arf4-independent route , such as intraflagellar transport ( reviewed in [9] ) . Of particular interest in this context is the role of Ift20 , a dynamic component of the IFT complex B , which localizes at both Golgi and at the ciliary base [43 , 44] . This positions Ift20 to recruit and guide rhodopsin transport vesicles from the biosynthetic membranes to the cilium . Accordingly , inducible knockout of Ift20 caused rhodopsin accumulation in the Golgi membranes supporting a role for Ift20 in sorting or transporting rhodopsin from Golgi to the outer segment [45 , 46] . Arf4 was predicted to be a cystic kidney disease gene based on polycystin-1 [17] and fibrocystin [24] trafficking defects in vitro . Ward and colleagues showed that Arf4 bound to a VxPx motif in the C-terminal tail of polycystin-1 that was required for targeting of the protein to cilia . They further showed that knockdown of Arf4 in cell culture prevented polycystin-1 from localizing to cilia . We identified a short motif in fibrocystin that can direct GFP to cilia and found that this peptide bound strongly to Arf4 even though it did not contain a VxPx motif . Knockdown of Arf4 delayed the delivery of a fibrocystin-GFP fusion protein to cilia but did not affect its steady state level [24 , 47] . In our current work we directly tested the role of Arf4 in cystogenesis by deleting the gene from the kidney using two different strategies . While these approaches yield cystic kidneys when other cilia genes are deleted neither yielded any evidence for cystogenesis when Arf4 was deleted . This indicates that if Arf4 is important in the kidney , other proteins can compensate in its absence . The VxPx motif , first identified in rhodopsin [48] , has been proposed as a predictor of ciliary localization [2] and VxPx motifs have been found in several ciliary proteins . The best studied example besides rhodopsin , is polycystin-2 , which is directed to cilia via the R6VxP motif in the N-terminal part of the protein [19] . A VxPx is also important in the outer segment targeting of RDH8 , a lipidated retinol dehydrogenase located in outer segment disc membranes [49] , as well as the delivery of CNGB1b to olfactory cilia [20] . Arf4 was reported to bind to a VxPx motif in the C-terminal tail of polycystin-1 that was thought to be the ciliary targeting sequence for this protein [17] . Subsequent work failed to reproduce the requirement for the VxPx motif in targeting polycystin-1 to cilia and found no effects on ciliary polycystin-1 when Arf4 was depleted [50] . The motif is also found in CRMP-2 and Nphp3 but mutational analysis indicates that the motif is not required for ciliary localization [51 , 52] . To further assess the value of the VxPx motif to predict ciliary localization , we analyzed the prevalence of this motif in the mouse proteome . Predicted frequency is 1 motif every 292 amino acids based on 6 . 6% V and 5 . 2% P in the average protein [53] . Similar to predictions , on average we found 1 motif every 237 amino acids and 58% of mouse proteins contained the motif . Since the best guesses of the ciliary complexity suggests that 5% or less of the proteins in the mouse are associated with cilia , finding a VxPx motif in a protein has no predictive value in terms of cellular localization . This is further compounded by the observations that ciliary targeting sequences in fibrocystin and multiple GPCRs do not contain VxPx motifs . While our work clearly demonstrated the absence of Arf4 does not cause retinal degeneration or cystic kidney disease , Arf4 is a critical protein in post-natal mouse development . In addition to augmented hair color , growth restriction and early death , the most striking phenotype observed in our study is severe degeneration of the exocrine pancreas . The acinar cells of the exocrine pancreas degenerated leaving the islets surrounded by adipocytes and fibrotic material rather than being embedded in exocrine tissue . Fatty pancreas , which goes by the names of pancreatic steatosis or pancreatic lipomatosis , is a significant human pathology [54] . In obesity-associated disease , the adipocytes are believed to infiltrate the pancreas [40] , although transdifferentiation of the exocrine cells into adipocytes is also possible upon the loss of cMyc [54] . In the cMyc-driven transdifferentiation , both lipid droplets and zymogen granules are observed in the same cells . In our case , we do not see lipid droplets and zymogen granules together in a cell . However , the exocrine cells appear to lose markers of zymogen granules early in the degeneration process , so it is possible that transdifferentiation is occurring . Mutations causing exocrine degeneration and fibrosis have been described in a number of mouse models , including other components of the secretory system such as Sec23 . Sec23 encodes a subunit of the COPII complex involved in transport between ER and Golgi . Similar to our observations in the Arf4 knockout , Sec23-defective mice have reduced zymogen granules [55] . However , these mice also showed very distended ER , which was not observed in our Arf4 knockout animals . Exocrine pancreas degeneration has also been described in several mouse models with ciliary defects including Ift88 and Pkd1 [56–58] . These mice share the phenotypes with Arf4 in that the exocrine pancreas is primarily affected without structural defects in the endocrine pancreas . However , the degeneration in the ciliary-related mutations starts with cyst formation in the exocrine ducts , which are not observed in the Arf4-defective pancreas . In summary , we have demonstrated that the loss of Arf4 from the mouse kidney and retina does not recapitulate phenotypes that would be expected if Arf4 was critical for sorting or transporting proteins to the cilium and the outer segment . However , the embryonic lethality that we observed with a germline mutation and the post-natal lethality observed in conditional allele supports critical role for Arf4 in transport through the endomembrane system in specific organs , most strikingly the exocrine pancreas .
Arf4flox mice were generated at the Duke Transgenic Mouse Facility . LoxP sites were inserted upstream and downstream of Arf4 exon 2 and 3 using standard BAC Recombineering techniques . Neo selection was used to select positive embryonic stem ( ES ) cell clones that were then validated by both Long Range PCR and Southern blot . Validated ES clones were then injected into the 8-cell embryo stage , using the VelociMouse method as described in [59] , to generate Arf4flox chimeras . Chimeras were breed to C57Bl6/J mice to produce Arf4+/flox mice . Mice were maintained as C57Bl6/J congenics . Genotyping was done with cArf4-KO-F gggaggattgggaagacaat , cArf4-KO-R1 caccacttgactgggaaggt and cArf4-KO-F2 agcagcctcattgtcctagc , which produces a 400 bp product in wild type and a 268 bp product in the Arf4flox allele . HoxB7-Cre [36] and CagCreER [27] were obtained from Jackson Laboratory . iCre75 [30] were a generous gift from Neena Haider , Schepens Eye Research Institute . For post-natal deletion of Arf4flox by CagCreER , 5 microliters of 20 mg/ml ( 0 . 1 mg ) tamoxifen dissolved in corn oil was administered by intraperitoneal injection on P2 . All mouse work was approved by institutional animal use committees at Duke University ( protocol A254-16-12 ) or University of Massachusetts Medical School ( protocol A-1174 ) as regulated by the National Institutes of Health Office of Laboratory Animal Welfare . EMBOSS patmatdb was run against the ENSEMBL GRCm38 protein collection . Of the 56999 proteins in this proteome , 33260 ( 58% ) had at least one VxPx motif and the average protein had 2 . 51 motifs . The average occurrence was 1 motif every 237 amino acids . Predicted frequency is 1 motif every 292 amino acids ( based on 6 . 6% V and 5 . 2% P in the average protein [53] ) . MEFs were generated from E14 embryos from pregnant mouse treated with 100 microliters of 10 mg/ml ( 1 mg ) tamoxifen ( by oral gavage ) 48 hr before harvest . Cells were immortalized with Large T antigen and cloned to generate control ( Arf4flox/flox/CagCreER ) and experimental ( Arf4-/-/CagCreER ) lines . MEFs were cultured in 90% DMEM ( 4 . 5 g/L glucose ) , 10% fetal bovine serum , 100 U/ml penicillin , and 100 μg/ml streptomycin ( all from Gibco-Invitrogen , Grand Island , NY ) . SV40 Large T immortalized cells were used for analysis . mIMCD3 [60] cells were cultured in DMEM/F12 with 10% fetal bovine serum , 100 U/ml penicillin , and 100 μg/ml streptomycin and transfected with Flag-tagged Arf constructs as described in [24] . Cells for immunofluorescence staining were grown on glass coverslips . Serum was reduced to 0 . 25% for 2 days before fixation to enhance ciliation [43 , 61] . Tissues from experimental and control mice were collected , homogenized and sonicated in 2% sodium dodecyl sulfate ( Sigma ) and 1× cOmplete protease inhibitor mixture ( Roche , Indianapolis , IN ) in phosphate-buffered saline ( PBS ) . Eyecup lysates were cleared at 14 , 000 g for 10 min at 22°C and total protein concentration was measured using the RC DC Protein Assay kit ( Bio-Rad , Hercules , CA ) . Kidney and pancreas lysates were cleared at 100 , 000 g for 20 min at 22°C and total protein concentration was measured using the BCA Protein Assay Kit ( Pierce ) . Serial dilutions of each lysate were subjected to SDS polyacrylamide gel electrophoresis using a 10–20% Tis-HCL criterion gel ( Bio-Rad ) and transferred to Immun-Blot LF PVDF ( Bio-Rad ) . Western blotting was performed by incubating in the appropriate primary antibody diluted in 50% / 50% of Odyssey blocking buffer / PBST overnight at 4°C . Blots were then rinsed 3 times with PBST before incubating in the corresponding secondary goat antibodies conjugated with Alexa Fluor 680 or 800 ( Invitrogen ) for 2 hr , at 22°C . Bands were visualized and quantified using the Odyssey infrared imaging system ( LiCor Bioscience , Lincoln , NE ) . | Primary cilia are sensory organelles found on most cells and contain specific receptors that detect extracellular stimuli . Defects in trafficking receptors to cilia cause a diverse set of diseases called ciliopathies , which include polycystic kidney disease , obesity , cerebral anomalies and retinal degeneration . Based mostly on in vitro studies , the small GTPase Arf4 was thought to be critically important for localizing rhodopsin to the outer segment of photoreceptor cells and cystoproteins to kidney cilia . Here we genetically remove Arf4 from mice in either a tissue specific or time dependent manner . To our surprise , the loss of Arf4 does not cause retinal degeneration or cystic kidney disease . Since ciliary dysfunction causes retinal degeneration and cystic disease , our findings indicate that Arf4 does not play a role in ciliary function . Instead , mice have zymogen granule defects and degeneration of the exocrine pancreas supporting roles for Arf4 in regulating endomembrane trafficking in specific cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"zymogens",
"enzymology",
"social",
"sciences",
"neuroscience",
"animal",
"models",
"model",
"organisms",
"experimental",
"organism",
"systems",
"sequence",
"motif",
"analysis",
"kidneys",
"cellular",
"structures",
"and",
"organe... | 2017 | Loss of Arf4 causes severe degeneration of the exocrine pancreas but not cystic kidney disease or retinal degeneration |
The interaction between an antibiotic and bacterium is not merely restricted to the drug and its direct target , rather antibiotic induced stress seems to resonate through the bacterium , creating selective pressures that drive the emergence of adaptive mutations not only in the direct target , but in genes involved in many different fundamental processes as well . Surprisingly , it has been shown that adaptive mutations do not necessarily have the same effect in all species , indicating that the genetic background influences how phenotypes are manifested . However , to what extent the genetic background affects the manner in which a bacterium experiences antibiotic stress , and how this stress is processed is unclear . Here we employ the genome-wide tool Tn-Seq to construct daptomycin-sensitivity profiles for two strains of the bacterial pathogen Streptococcus pneumoniae . Remarkably , over half of the genes that are important for dealing with antibiotic-induced stress in one strain are dispensable in another . By confirming over 100 genotype-phenotype relationships , probing potassium-loss , employing genetic interaction mapping as well as temporal gene-expression experiments we reveal genome-wide conditionally important/essential genes , we discover roles for genes with unknown function , and uncover parts of the antibiotic’s mode-of-action . Moreover , by mapping the underlying genomic network for two query genes we encounter little conservation in network connectivity between strains as well as profound differences in regulatory relationships . Our approach uniquely enables genome-wide fitness comparisons across strains , facilitating the discovery that antibiotic responses are complex events that can vary widely between strains , which suggests that in some cases the emergence of resistance could be strain specific and at least for species with a large pan-genome less predictable .
Bacteria evolve antibiotic resistance in response to selective pressures that emerge from the interaction between the antibiotic and the bacterium . Routes of escape often lead through ways that diminish the interaction with the direct target . For instance , escape from penicillin , whose direct target are the Penicillin Binding Proteins ( PBPs ) , can often be found in mutations in PBPs that decrease the affinity for the drug , or in functionally related genes that compensate for diminished function [1–4] . However , it has become clear that the relationship between a bacterium and an antibiotic reaches far beyond its direct target . Instead an antibiotic triggers a complex , multi-factorial process that may begin with the physical interaction between the drug and its target but quickly propagates into the involvement of a variety of processes that can include regulation , metabolism and/or energy generation [5–12] . These system-wide selective pressures could explain why clinical strains often contain multiple alterations that may contribute to resistance but are located in genes whose primary role is not resistance but rather are involved in fundamental bacterial processes [13–17] . The adaptive sequence space thus seems to lie well beyond the antibiotic’s direct target , which contributes to the complexity of determining and predicting how and where resistance evolves . Moreover , adaptive mutations can be species-specific . For instance , the direct target of fluoroquinolones in gram-negatives including Escherichia coli is DNA gyrase , while in gram positives such as Staphylococcus aureus it is topoisomerase IV , which may explain why mutations in gyrA such as S83L in E . coli can increase the MIC to fluoroquinolones while the equivalent mutation in gyrA in S . aureus does not necessarily have an effect on the MIC [18–23] . Additionally , some gyrA mutations are associated with a fitness cost in E . coli [24 , 25] , while a positive fitness effect can be observed in Campylobacter jejuni and Salmonella enterica [26 , 27] . These contrasting phenotypes suggest that the manner in which a bacterium experiences antibiotic-induced stress may differ depending on the genomic background and the underlying genomic network . While detailed insights into these factors could help in designing novel antimicrobial strategies , the importance of the genomic background and to which extent antibiotic-sensitivity and resistance depend on network architecture is currently unclear . In this study we use Streptococcus pneumoniae to explore the importance of the genomic background on antibiotic-sensitivity and the manner in which stress is experienced and processed . S . pneumoniae is a human nasopharyngeal commensal and respiratory pathogen . It triggers pneumococcal pneumonia , meningitis , and septicemia , which results in ~1 million deaths annually among children <5 years of age , and ~0 . 5 million among groups including the immunocompromised and the elderly ( >65 yrs . ) , making it one of the most important bacterial pathogens worldwide [28–30] . Although vaccination has been successful , we and others have shown that it does not result in complete protection , and that some groups , such as children with Sickle Cell Disease , remain especially vulnerable [31] . Antibiotics thus continue to be extremely important as a treatment option , especially in acute disease . However , as with almost any clinically important bacterial pathogen , the emergence of multidrug-resistant ( MDR ) strains is a global problem [32–37] and with 1 . 2 million drug-resistant pneumococcal infections annually in the US , and $96 million in excess medical costs , S . pneumoniae is a serious concern [30] . S . pneumoniae is one of several species for which the availability of complete bacterial genomes has demonstrated that a distinction can be made between its core-genome ( the pool of genes shared by all members of a species ) and pan-genome ( a species’ global gene repertoire ) [38–41] . On average two pneumococcal strains may differ by ~300 genes in their genomic content , i . e . the presence and absence of genes [42 , 43] , which highlights the genome’s plasticity to retain function in the presence of variation . Such plasticity is remarkable because no genomic element , gene , or pathway exists in a vacuum; rather they are connected through networks resulting in specific organismal properties [44 , 45] . A newly acquired element thus needs to be integrated thereby possibly affecting existing connections and creating new ones . Consequently , no two genomes may function in the same manner , potentially affecting phenotypes ranging from drug tolerance to virulence to evolutionary potential . By employing genome-wide approaches we , and others , have shown that it is possible to determine , upon exposure to an environmental perturbation , where stress in the bacterial genome is experienced [31 , 46–55] . Here we apply Tn-Seq , a tool for systems-level analysis of microorganisms , which combines transposon mutagenesis with massively parallel sequencing to determine genome-wide fitness in a single experiment . We develop daptomycin-sensitivity profiles for two strains of the bacterial pathogen S . pneumoniae . Although the exact mechanism of action of daptomycin is not completely clear it seems to insert itself into the membrane for which the presence of phosphatidylglycerol in the membrane is required . Following insertion , membrane structure and curvature may be distorted leading to cells with altered cell shapes . These distortions in the membrane at the site of daptomycin insertion may lead to leakage of ions and loss of membrane potential and local dysregulation of cell division and/or cell wall-biosynthesis [56–61] . The daptomycin-sensitivity profiles generated in this study illustrate how the antibiotic’s effects ripple through the organism and how the bacterium deals with this stress with a diverse set of genes from different functional categories and organizational levels including: cell-wall organization , membrane integrity and transport , control and regulation of fundamental processes , and metabolism . Surprisingly , the sensitivity profiles turn out to be highly strain-specific highlighted by over 50% of sensitivity-profile genes that increase antibiotic sensitivity in one strain but have no effect , or even decrease sensitivity in the other strain . We show that these differences are partially the result of a network architecture that is not well conserved , exemplified by strain-specific differences in Potassium ( K+ ) -release and ClpP functionality , as well as differences in regulatory relationships between genes from different organizational levels . Importantly , we present a generally applicable , and highly sensitive approach that enables comparisons of environment-induced fitness effects on a genome-wide scale and species-wide level .
A major goal of this study is to determine whether bacterial strains from the same species that differ in their genomic content , respond in an identical manner to antibiotic stress . On a species level the genome of S . pneumoniae can be divided up in a core genome consisting of ~1600 genes , and a pan genome of ~4000 genes , while a genome on average has approximately 2000 genes . We selected two strains , TIGR4 ( T4 ) and Taiwan-19F ( 19F ) ( S1 Fig ) , which can both cause invasive disease: T4 is a serotype 4 strain that was originally isolated from a patient from Norway with Invasive Pneumococcal Disease ( IPD ) [62 , 63] , while 19F is a multi-drug resistant ( MDR ) strain isolated from a patient with IPD from Taiwan [64 , 65] . With respect to genomic content the strains share 1711 genes , while T4 has 324 genes that are absent in 19F , and 19F has 204 genes that are absent in T4 . On average two pneumococcal strains may differ by ~15% in their genomic content , and thus the amount of variation between these two strains is representative for differences observed between strains within the species [42 , 43] . All genes were split into 17 functional categories and except for the number of genes with unknown function , both strains share a similar distribution over these categories ( S2 Fig; S1 Table ) . Both strains are differentially susceptible to different antibiotics , for instance 19F is approximately 25-fold less susceptible to penicillin than T4 , while they are equally sensitive to daptomycin ( Fig 1 ) . Because equal sensitivity creates the simplest opportunity to test whether strains use the same genes in dealing with antibiotic induced stress , daptomycin is used here . To identify in detail which genes in the genome are involved in dealing with daptomycin-stress we employed transposon insertion sequencing ( Tn-Seq ) , which enables high-throughput and accurate calculations of the growth rate for each possible gene-knockout in the genome [31 , 46 , 48 , 49] . Six independent transposon libraries , each consisting of ~10 , 000 mutants were created in T4 and were grown in the absence and presence of daptomycin at a concentration of 25 μg/ml , which moderately slows the growth rate by ~15% ( Fig 1 ) . For each condition reproducibility was determined by comparing fitness between different libraries , which in each case was high ( R2 = 0 . 78–0 . 89 ) . Fitness values for each insertion in each gene were averaged and genes with a significant antibiotic-specific response were visualized in a network with Cytoscape [66] and grouped according to their functional category ( Fig 2A; S2 Table ) . Previously we showed how such a visual network approach provides a detailed overview of how an environmental disturbance can affect a bacterium on multiple different levels [31 , 48] . The same is true for this network , demonstrating how a large variety of genes from different functional categories become important for the survival of T4 in the presence of daptomycin ( Fig 2A; S2 Table ) , which highlights several aspects of daptomycin’s modus operandi: 1 ) Our results show that any gene that affects peptidoglycan ( PG ) biosynthesis , stability , or regulation can , upon its removal , make the bacterium more susceptible to daptomycin . Even a decrease in PG acetylation ( mediated by SP1479 ) or a decrease in the speed and efficiency of Penicillin Binding Protein ( PBP ) folding , which has been shown to be mediated by PrsA in B . subtilis [67] ( prsA/SP0981 ) , increases susceptibility to daptomycin . Associations of daptomycin with the cell wall have also been shown in other bacteria: in B . subtilis , and S . aureus , daptomycin induces a cell wall stress response [59 , 68] , and mutations that increase cell wall thickening have been associated with resistance in S . aureus . [69 , 70]; 2 ) Lipo- and membrane proteins that provide structural support to the membrane become important in the presence of daptomycin , suggesting that the interaction of daptomycin with the membrane has a de-stabilizing effect on membrane integrity . The importance of the membrane anchored protease FtsH ( SP0013 ) , whose function includes proteolysis of aberrant membrane proteins and thereby influences membrane turnover [71 , 72] , further suggests that daptomycin negatively affects membrane-protein stability and thus membrane integrity . In B . subtilis , daptomycin preferentially interacts with regions of the membrane enriched in phosphatidylglycerol ( PhG ) [59] , it has been physically associated with sites of membrane distortion [56] and resistance is linked to the overall PhG content [73] , in S . aureus mutations in mprF increase daptomycin resistance by changing membrane lipid composition and charge [74–76] , while in Enterococci changes in cardiolipin synthesis can increase daptomycin resistance [77 , 78]; 3 ) A Trk-system ( SP0479-0480 ) , which mediates K+-uptake [79] , becomes important in T4 in the presence of daptomycin , which suggests that T4 suffers from daptomycin-induced K+-loss . Indeed , it is assumed that daptomycin triggers potassium loss through its interaction with the membrane [57 , 60 , 61] . Besides K+ , other ions may be leaking out as well [61] , or at least other ions become more important and may compensate for K+-loss , which is highlighted by the importance of several ion transport systems in our network including SP1623 ( annotated as a cation-transporter ) , which we previously associated with pH-homeostasis [48]; 4 ) The sensitivity profile highlights the importance of a diverse set of cell division , RNA and protein turnover , signaling and regulation , and metabolism genes , indicating that the antibiotic’s effects resonate throughout some of the most embedded systems in the bacterium . Importantly , these profiles can also suggest roles for genes with unknown or unclear functions in at least two ways: 1 ) A gene with unknown function adjacent to a gene with a defined function and a similar sensitivity suggests that the two genes are involved in the same process . For instance SP1730 and SP1731 are hypothetical genes and have a similar fitness as their regulatory neighbors SP1732 ( stkP ) and SP1733 ( phpP; the cognate phosphatase of stkP ) , which sense intracellular peptidoglycan and have regulatory control over cell-division [80] . 2 ) Genes with domains that suggest a function or association with a specific functional category are more likely to be correct if they fit within the sensitivity profile . For instance BLAST searches and protein domain predictions predict that both SP1505 and SP1720 are membrane proteins . These predictions fit well with the sensitivity profile where membrane proteins make up one of the most important categories . Moreover , with GFP-fusions we confirmed the localization of both proteins in the membrane . We expected to uncover highly similar sensitivity profiles due to the strains’ equivalent susceptibility to daptomycin . However , less than 50% of the responsive genes have a conserved phenotype between strains ( Fig 2B; S3 Table ) , and based on a Jaccard similarity index , the networks are significantly different ( J = 0 . 24 , p<0 . 05 ) [81–83] . Additionally , the overall distribution of functional gene categories is significantly different between strains ( two proportion exact test; Z = 5 . 83 , N1 = 52 N2 = 57 , p<0 . 01 ) . This means that both on the individual gene-level as well as the overall functional level there is little conservation between strains in the distribution of the type of genes that are important in dealing with daptomycin stress . However , genes and pathways interact with each other and responses could be more conserved on a global scale . Therefore we grouped functional categories to determine whether we could analytically track how an antibiotic interacts with a bacterium and thereby identify how a bacterium in first instance perceives a ( extracellular ) threat and subsequently how this threat is processed . To enable this , functional categories were combined into four hierarchical groups , or layers . The first layer combines categories that make up the first physical layer an antibiotic could interact with , which is the capsule and the cell wall represented by peptidoglycan genes . The second physical layer of interaction is represented by the membrane and consists of the categories membrane , lipoprotein and transporter genes . The third layer combines genes that control and orchestrate fundamental processes: cell division , DNA turnover , RNA turnover , protein turnover , transcription and translation and regulation . The fourth and last layer combines all metabolism genes including nucleotide , carbohydrate and amino acid metabolism . These four layers thus combine the physical location of the gene-product with its molecular function ( Fig 2 ) . For both strains the first layer is indeed the first point of interaction ( Fig 2A and 2B ) after which the membrane becomes the next obstacle . This second layer includes 21 responsive genes in T4 and 20 in 19F and even though only 25% of the transporter genes in this layer are conserved , ~70% of the membrane and lipoprotein genes are conserved between strains . Thus , at least for the part of the network that is important for membrane integrity the two strains seem to experience and process daptomycin stress in a similar fashion . Moreover , when a global analysis is performed , in which the gene-categories are first collapsed into the described four layers , and then the four layers are compared between the two strains , we no longer observe a dissimilar response ( two proportion exact test , Z = 2 . 85 , N1 = 52 N2 = 57 , p = 0 . 16 ) , indicating that although on the individual gene level there is little conservation , the global response is more similar . To confirm that the wide variety of genes involved in dealing with antibiotic stress , as well as the lack of phenotypic conservation is not limited to daptomycin we further performed Tn-Seq with an aminoglycoside , a glycopeptide and a fluoroquinolone , and in both T4 and 19F , which shows that also these three classes of antibiotics trigger stress that is processed with genes from a wide variety of categories ( Fig 3 ) . In addition , conservation of phenotypes , i . e . the genes that either strain uses to deal with antibiotic stress , shows , similar to daptomycin , a limited signature of conservation between ~40–50% ( Fig 3 ) . To validate the sensitivity profiles , and exclude that the lack of conservation comes from low confidence Tn-Seq data , we compared Tn-Seq fitness ( WTn-Seq ) to fitness obtained from individual growth curves and/or from 1x1 competition assays ( W1x1 ) , in which a deletion-mutant is competed against the wild type . Note that in all three of these cases fitness ( W ) is calculated as the growth rate thereby enabling direct comparisons . In total 34 deletion mutations in T4 and 19F were constructed and sixty-five genotype-phenotype relationships were validated in the presence and absence of daptomycin ( Fig 4A and 4B; Table 1 ) . This resulted in a strong correlation ( R2 = 0 . 87 ) , which is similar to correlations we achieved previously [48] and confirms high-confidence Tn-Seq fitness data . Therefore , even though the strains have the same susceptibility to daptomycin , belong to the same species and share ~85% of their genomic content , this suggests that the underlying genomic networks must be different , which makes the strains respond in a different manner to the same stress . To develop a better understanding of how the underlying networks differ between strains we first set out to determine the role of a Trk-K+-uptake system which the Tn-Seq data indicates is important in the presence of daptomycin in T4 ( WTrk1 = 0 . 82 ) , while it is dispensable in 19F ( WTrk1 = 0 . 98 ) . Ion homeostasis is an essential part of life and transport systems are mandatory for ion uptake and extrusion . Although in general only traces of potassium are available in the environment it is generally the most abundant cation in bacteria and plays an essential role in for instance the maintenance of internal pH , in membrane potential adjustment , it acts as second messenger for stress signaling and it is a regulatory element for transcription control [79] [84] . Tn-Seq data for T4 indicates that 8 transporters become important in the presence of daptomycin , of which two encode a single Trk-K+-uptake system ( Trk1: SP0479-SP0480; Fig 5A ) suggesting that T4 suffers K+-loss upon exposure to daptomycin and that this two-gene system is important in countering that loss . The importance of Trk1 in the presence of daptomycin was validated ( Fig 5B , Table 1 ) and is concentration dependent indicated by a further drop in fitness upon an increase in daptomycin ( Fig 5C ) . By adding additional K+ both the wt and the Trk1 mutant can be ( partially ) compensated , which shows that the system becomes less important when K+ is more abundant ( Fig 5D ) . Moreover , the mutant is also sensitive to valinomycin ( an ionophore that releases K+ ) confirming that the importance of Trk1 is indeed due to daptomycin induced K+-loss ( Fig 5E ) . Interestingly , S . pneumoniae T4 has an additional Trk-uptake system ( Trk2: SP0078-SP0079; Fig 5A ) that was not indicated by Tn-Seq ( there was no loss of fitness ) , which we indeed verified ( Fig 5F , Table 1 ) . To further investigate the role of the two Trk-systems in K+-homeostasis as well as their importance in dealing with daptomycin stress , internal K+-concentrations under different conditions were determined for the wild type and the two Trk-system mutants . Although under standard growth conditions there is no clear difference in fitness between the wt , and the two Trk-system deletion mutants , internal K+-concentrations do vary between strains: ΔTrk2 contains on average 7-fold less K+ , while ΔTrk1 contains on average 14-fold less K+ relative to the wt ( Fig 6A ) . Interestingly , ΔTrk1 becomes very unstable under standard growth conditions and after it reaches peak OD it settles at the bottom of the culture vial , indicating that at high OD it is highly susceptible to lysis . Temporal gene expression analysis ( with/without daptomycin ) did not reveal any differences between the different Trk-systems . However , upon exposure to daptomycin clear differences in the loss of K+ were observed between all three strains: K+ -loss for the wt is the greatest ( ~4 . 5 fold ) , it is intermediate for ΔTrk2 ( ~2 . 5 fold ) , while for ΔTrk1 the already low K+ concentration drops by another 1 . 5-fold ( Fig 6B ) . K+ -loss thus partially seems dependent on the K+ concentration in the cell , i . e . the higher the concentration the bigger the loss . However , even though ΔTrk1 has the smallest relative loss , its final concentration is still ~7 . 5 fold lower than the wild type ( Fig 6A ) , which is enough to drastically hamper its growth ( Figs 1 and 5B–5C ) . In contrast , ΔTrk2 only has a 2-fold lower K+-concentration than the wt , which is apparently not large enough to affect growth . These results show that there is a clear hierarchy in Trk-system importance; Trk1 being the most important K+-uptake system to control K+-homeostasis both during normal growth conditions as well as during exposure to daptomycin . The lack of phenotypic conservation for Trk1 with respect to daptomycin-sensitivity between T4 and 19F could be the result of a third K+-uptake system in 19F that creates redundancy in K+-control . Indeed , we found a strain-specific third system in 19F encoded by SPT1006 , which is annotated as a K+-ion channel ( Fig 7A ) . To determine whether there is any hierarchy among these three systems , single knockouts were created , as well as double knockouts for all three possible combinations . However , none of these six mutants made 19F more susceptible to daptomycin . Even the double knockout consisting of the strain-specific 19F K+-system ( SPT1006 ) and Trk1 did not affect sensitivity of 19F to daptomycin nor valinomycin ( Fig 7B–7H ) . Measuring internal K+-concentrations confirmed that daptomycin-induced release of K+ was similar for 19F-wt and the three single mutants ( S3 Fig ) . The double mutants showed more variation but none of them was as dramatic as the knockout for Trk1 in T4 ( Fig 6A–6D ) . This suggests that daptomycin-induced release of K+ is counteracted by all three 19F K+-uptake systems and that , in contrast to T4 , there is no hierarchy amongst the systems . Additionally , K+-concentrations are approximately 5-fold higher in wt-T4 compared to wt-19F and daptomycin may have less of an impact on K+-loss in 19F , while instead other ions may be leaking out as suggested by the importance of the manganese transporter in 19F ( Fig 2B; SP1649-1650 ) . Importantly , these data confirm the Tn-Seq data and the lack of conservation between T4 and 19F indicating that the two strains handle K+-stress differently , possibly due to differences in their underlying networks . We have previously shown that we can ( partially ) reconstruct underlying networks by generating a genetic interaction map ( GIM ) , which is accomplished by creating a transposon library in the background of a query strain ( i . e . a strain that carries a knockout of the gene of interest ) [46 , 48] . For T4 , ΔTrk1 was used as a query strain , while for 19F the double mutant ΔTrk1-ΔSPT1006 was used to compensate for the presence of the additional K+-ion channel . Significant response genes ( i . e . interactions that deviate from the multiplicative model; see methods section for definition ) were visualized in a network and grouped according to their functional category ( S4A and S4B Fig; S4 and S5 Tables ) . There seems to be little conservation in genetic interactions: the networks are significantly different ( J = 0 . 08 , p<0 . 001 ) , the overall distribution of functional gene categories is significantly different ( two proportion exact test , Z = 8 . 11 , N1 = 49 N2 = 34 , p<0 . 01 ) , as well as the global response ( two proportion exact test , Z = 4 . 38 , N1 = 49 N2 = 34 p<0 . 01 ) . One of the most obvious differences between the strains is found in the second layer , which includes 18 transporters in T4 , of which Trk1 and 2 are indicated as synthetically lethal , which we confirmed by our inability to construct this double mutant , five transporters have an aggravating interaction ( a lower fitness than expected from the multiplicative model ) , while 12 interactions are alleviating ( a higher fitness than expected from the multiplicative model ) . Moreover , a further 21 additional alleviating interactions are present between Trk1 and genes from all four layers . In contrast , there are only 5 overall conserved interactions between strains , including a single alleviating transporter interaction . Moreover , of the seven aggravating transporter interactions in 19F , six are annotated as involved in metal-ion transport , indicating that the manner in which 19F deals with daptomycin-induced loss of membrane-potential may be decentralized and spread across multiple different transporters . To confirm the genetic interactions and lack of conservation between strains we validated 21 genotype-phenotype relationships of which one ( ΔTrk1-ΔSP2195 ) had a significantly higher fitness than expected ( Fig 8A and 8B; Table 2 ) . The GIMs are thus robust , and give hints as to the type of relationships between interacting genes . For the metal-ion transporters in either map it seems relatively easy to explain why they interact with the query genes: the removal of the K+-homeostasis system ( s ) makes both strains sensitive to a further disturbance in the bacterium’s ion-mediated potential , and thus a loss of any transporter that is involved in retaining what is left of that potential will have a negative effect on fitness . In contrast , the alleviating interactions in T4 show that the removal of these genes has a positive effect on fitness . This effect could also be accomplished by transcriptionally repressing these genes , and thus this suggests that these genes are dysregulated in a T4-ΔTrk1 background . To test this hypothesis we picked 7 genes from the T4-ΔTrk1 network: 4 regulators and 3 transporters . Expression of these genes was followed in wt-T4 , wt-19F and the two query strains for 5 different time-points and in three independent experiments . In the query strain T4-ΔTrk1 the expression of six out of seven genes changed abruptly by approximately 4-fold between 30 and 45 minutes after addition of daptomycin ( Fig 9A and 9B ) , while T4-wt expression did not change for any of these genes , or changes were gradual over time ( Fig 9A and 9B ) . In 19F-wt and its query strain , expression changes for all genes were comparable ( Fig 9C and 9D ) , and fluctuations over time , except for the response regulator ciaR ( SP0798/TCS05 ) were within 2-fold . Even though the small changes in 19F could still be affecting the response , these results further show that the stress T4 and 19F experience is also processed in a different manner , seemingly with dysregulation in a select set of genes , including stress regulators such as ciaR and ctsR ( SP2195 ) , as a result . One could argue that the lack of conservation in the GIMs could be unique for the K+-transporters , which have a relatively straightforward function and may not be that deeply integrated into the organismal network . Therefore , we set out to construct GIMs for ClpP , a protease that plays a crucial role in the regulation of various cellular responses by controlling proteolysis . For instance ClpP has been shown to repress competence in B . subtilis and activate stress proteins by targeted degradation of the repressor CtsR , additionally it has been associated with cell division , sporulation and cell wall biosynthesis [72 , 85] . Thus , by definition , this conserved protease is , in its role as a protein turnover gene , deeply integrated into the organismal network . Tn-Seq analysis indicates , and individual growth curves confirm , that the role of ClpP in basic growth as well as its sensitivity to daptomycin is strain dependent: ΔclpP ( SP0746 ) in T4 only slightly affects growth and , surprisingly , decreases sensitivity to daptomycin ( Fig 10A and 10B; Table 1 ) , while in 19F ΔclpP substantially lowers the growth rate ( Fig 10C; Table 1 ) . The T4-specifc decrease in daptomycin-sensitivity seems specific for this antibiotic , since it has the opposite effect on gentamicin sensitivity , a protein synthesis inhibitor ( Fig 10D ) . Importantly , sensitivity to valinomycin is not affected in ΔclpP ( Fig 10E and 10F ) , and thus the change in daptomycin-sensitivity does not seem to be related to intracellular K+-concentrations , which further confirm that the effects of daptomycin reach beyond K+ . GIMs were constructed to determine ClpP connectivity within each strain as was done for the K+-systems ( S5A and S5B Fig; S6 and S7 Tables ) . 19 genotype-phenotype relationships were validated; two of those relationships ( SP0798-ΔclpP; SP0047-ΔclpP ) have a significantly different fitness compared to Tn-Seq data , but the phenotype is stronger than initially measured , which thus confirms the interaction ( Fig 8A and 8B; Table 2 ) . Also here , there is little conservation between strains , the GIMs are significantly different ( J = 0 . 06 , p<0 . 001 ) , the overall distribution of functional gene categories is significantly different ( two proportion exact test , Z = 10 . 49 , N1 = 23 N2 = 21 , p<0 . 01 ) , as well as the global response ( two proportion exact test , Z = 5 . 06 , N1 = 23 N2 = 21 , p<0 . 05 ) . For instance in T4 , ClpP interacts in an aggravating manner with genes that are located adjacent to ClpP on the genome as well as a large number of nuclear metabolism genes . This latter relationship indicates that ClpP has control over nucleotide metabolism in T4 , which has indeed been suggested for S . pneumoniae [86] as well as B . subtilis [87] , but here that relationship seems to become important during daptomycin stress . In 19F , we confirmed that this strong relationship is not present ( Table 2 ) , instead , the clearest pattern in 19F emerges from ClpP interacting with a set of genes of which several play a role during competence including two component system-12 ( TCS12 ) , the regulators stkP ( SP1732 ) and ciaR ( SP0798 ) , and comM ( SP1945 ) , a membrane protein that protects against lysins and fratricide [88–90] ( Fig 11A , S5B Fig ) . The importance of TCS12 suggests that ClpP has a repressive regulatory effect on this system; i . e . since the comD/E system that makes up TCS12 becomes important in the absence of ClpP this suggests that it is activated ( Fig 11A ) . To verify this , we determined expression of the two TCS12 genes , comC ( which is in an operon with TCS12 ) and comM which is located ~300 genes downstream of TCS12 . As predicted comD , E and C , were highly upregulated ( between 20–60 fold ) in 19F in the absence of ClpP ( Fig 11B ) while comM was upregulated approximately 12-fold . Because of the importance and upregulation of the ‘anti-lysis’ gene comM , we expected that Pmp23 ( SP1026 ) , a membrane protein that is associated with lysis through its possible role in peptidoglycan turnover [91–93] , and SP0650 , a membrane protein with possible hydrolase activity ( which are both present in the network; S5B Fig ) , would also be upregulated , and that ComM was possibly protecting against their actions . However , this hypothesis had to be rejected since the expression of pmp23 and SP0650 hardly changed in the absence of ClpP , which does not exclude that ClpP still has control over these genes through its protease activity . In contrast , relative expression of all 6 genes was unchanged in T4-wt and T4-ΔclpP ( Fig 11B ) , confirming that under the tested conditions there are no relevant interactions between these genes in the T4 background . These data thus show that even for a highly conserved gene , genetic interactions are not necessarily conserved , which can lead to responses that are largely strain dependent . It has become clear that the interaction between an antibiotic and bacterial cell is a complex , multi-factorial process that resonates through the organism requiring the involvement of a diverse set of fundamental processes to overcome the antibiotic-induced stress [5–8] . The selective pressures invoked by an antibiotic are thus not only felt by the direct target but are dispersed across many different layers . This distribution of stress thus expands the adaptive sequence space , which may explain why multiple genetic perturbations across different layers can combine to confer elevated levels of resistance [13–17] . Here , we develop daptomycin-sensitivity profiles showing in detail the genes that are important in coping with daptomycin-induced stress . By creating hierarchical layers , that partially represent physical barriers the antibiotic interacts with as well as fundamental processes that regulate and ensure all aspects of the bacterium’s life cycle , we identify where the antibiotic has its biggest impact . We believe that these types of analyses can be used to uncover the bacterium’s weakest-links in the presence of an antibiotic and thus identify novel targets that could work synergistically with existing drugs , while it also indicates where in the genome the bacterium may adapt to decrease its sensitivity to the stress . For instance , mutations in daptomycin adapted strains of S . aureus , B . subtilis and Enterococci have been observed in ClpP and other proteases , different regulatory genes and TCSs , capsule genes , transporters , nucleotide metabolism genes , peptidoglycan genes , lipoproteins and membrane genes [56 , 69 , 70 , 73–78 , 94–98] . Although many of these mutations have not ( yet ) been directly linked to higher resistance , we show here that they may indeed contribute to drug-sensitivity . Importantly , it turns out that the sensitivity profiles strongly depend on the genomic background , and that even within a species responses can be strain specific . We show that it is possible to at least partially dissect the underlying network of the response through constructing a GIM . By removing the query-gene , on which these maps are based , it is as if a protective layer is removed from the organism , thereby further exposing parts that become important in the presence of the stress , and at the same time revealing the type of dependencies that exist between genes , including regulatory relationships . However , by diving deeper into the response by means of these GIMs , we uncover more complexity and even less conservation across strains . Our approach thus reveals that an important part of antibiotic-induced stress is experienced and processed by S . pneumoniae in a strain dependent manner . Take this one step further and it implies that adaptation to an antibiotic will , at least partially , be strain dependent . And thus , this could be one of the reasons why it remains so difficult to predict the emergence of antibiotic resistance . This study provides , a clear approach as well as important arguments to not only construct antibiotic-sensitivity profiles for different antibiotics but also perform this across different bacterial species and strains . Such profiles in combination with in vitro and in vivo adaptation experiments could provide an important improvement in our ability to predict where in the genome mutations may arise that decrease susceptibility to an antibiotic and put the organism on the road towards full clinical-resistance .
Experiments were performed with S . pneumoniae strains TIGR4 ( NCBI Reference Sequence: NC_003028 . 3 ) and Taiwan-19F ( NC_012469 . 1 ) . TIGR4 is a serotype 4 strain that was originally isolated from a patient from Norway with Invasive Pneumococcal Disease ( IPD ) [62 , 63] , while 19F is a multi-drug resistant strain isolated from a patient with IPD from Taiwan [64 , 65] . All gene numbers in the tables and figures are according to the TIGR4 genome , except when it concerns a strain-specific gene , these are preceded by SP or SPT referring to a T4 or 19F gene respectively . PATRIC [99] and BLAST were used to compile S8 Table , which matches gene numbers between T4 and 19F and lists strain-specific genes for each genome . A gene is considered strain-specific if 70% of the sequence has less than 70% similarity with the other genome [42] . Single gene knockouts were constructed by replacing the coding sequence with a chloramphenicol and/or spectinomycin resistance cassette as described previously [46 , 48 , 100] . S . pneumoniae was grown on sheep’s blood agar plates or statically in semi-defined minimal media ( SDMM ) at pH 7 . 3 , which contains 70 μg/ml calcium to ensure activity of daptomycin , 20 mM glucose and 5 μl/ml Oxyrase ( Oxyrase , Inc ) , at 37°C in a 5% CO2 atmosphere [48] . Where appropriate , cultures and blood plates contained 4 μg/ml chloramphenicol ( Cm ) and/or 200 μg/ml Spectinomycin ( Spec ) . Library construction was performed as described with transposon Magellan6 , which lacks transcriptional terminators , therefore allowing for read-through transcription , and it diminishes polar effects [46 , 48 , 101 , 102] . Additionally , the mini-transposon contains stop codons in all three frames in either orientation when inserted into a coding sequence . Six independent transposon libraries were constructed in wt-T4 and wt-19F and in four query strains ( T4: ΔTrk1 , ΔClpP; 19F: ΔTrk1-SPT1006 , ΔClpP ) , and selection experiments were conducted in SDMM in the presence or absence of 25 μg/ml daptomycin , which in this environment moderately slows growth for both strains by ~15% ( Fig 1 ) . Sample preparation , Illumina sequencing and fitness calculations were done as described [31 , 46–49 , 101–103] . In short , for each insertion , fitness Wi , is calculated by comparing the fold expansion of the mutant relative to the rest of the population by using an equation that we specifically developed to have fitness represent the growth rate of a mutant [46 , 48 , 103] . All of the insertions in a specified region or gene are then used to calculate the average fitness and standard deviation of the gene knockout in question . The advantage of using this approach is that Wi now represents the actual growth rate per generation , which makes fitness independent of time and enables comparisons between conditions and strains . To determine whether fitness effects are significantly different between conditions or strains three requirements have to be fulfilled: 1 ) Wi is calculated from at least three data points , 2 ) the difference in fitness between conditions has to be larger than 10% ( thus Wi—Wj = < -0 . 10 or > 0 . 10 ) , and 3 ) the difference in fitness has to be significantly different in a one sample t-test with Bonferroni correction for multiple testing [46 , 48] . All significant fitness values were visualized in a network with Cytoscape [66] . Importantly , here , fitness ( Wi ) represents the actual growth rate per generation , which makes fitness independent of time and enables comparisons between conditions and strains . To determine whether the observed distributions in the antibiotic sensitivity profiles that are based on the functional categories or layers are different it is enough to show that , if the classes are grouped into 2 macro classes , the resulting distributions are different . To compare two 2-class distributions we use a two proportion exact test and we reject the equality hypothesis at a p-value ≤ 0 . 05 . We build these two macro classes such that the differences are as large as possible . Hence , when a test cannot distinguish between these two reduced distributions it indicates that the original , non-reduced , distributions are also similar . Genetic interactions are defined as a deviation from the multiplicative model , which states that if a strain deleted for gene i has a fitness Wi and a strain deleted for gene j has a fitness Wj , then the double mutant strain Wij is expected to have the fitness Wi x Wj [46 , 48] . Genetic interactions were determined for the four query strains and has generally more experimental noise [46 , 48] , therefore to minimize false positives , we set more stringent cut offs: 1 ) fitness needs to be composed of at least five data points; 2 ) expected and observed fitness have to deviate by at least 17 . 5% , and 3 ) significant interactions have to pass a student’s t-test with Bonferroni correction for multiple testing . An exponentially growing culture was washed and resuspended in TA buffer to an OD600 of ~0 . 3 , and a small amount of culture was plated on blood agar for enumeration . The external background potassium concentration was measured every 3 seconds for one minute at room temperature using the MI-442 K+-ion microelectrode and the MI-409 dip-type reference microelectrode ( Microelectrodes , Inc . , Bedford , NH ) . Note that: 1 ) longer measurements proved unnecessary as readings stabilized after several seconds , and 2 ) for every set of measurements the electrode was first calibrated with known concentrations of KCl to ensure a linear regression ( Vmeas = mlog10[K+] + z ) , where Vmeas is the average mV of 20 data points measured over one minute . In samples for which the effect of daptomycin on K+-loss was determined cells were exposed to daptomycin for 20 minutes after which they were washed and resuspended in TA-buffer . For each sample , the internal K+-concentration was determined in a second measurement after lysing all cells through boiling . External and internal K+ concentrations were calculated by converting Log10 [K+] into molar concentrations of K+ as described previously [104] . For 1x1 competitions two strains were mixed in a 1:1 ratio and grown for approximately 8 generations to late exponential growth phase . Fitness , representing the growth rate , was calculated through the same approach as Tn-Seq data above by determining the expansion of the competition over the experiment and by determining the ratios of the competing strains at the start and at the end of the competition by plating appropriate dilutions on blood agar plates with selective antibiotics [46 , 48] . Mutants were always competed against their background strain: strains with a single gene knockout were competed against the wild type strain , while double mutants were competed against the query strain . Each competition was performed no less than four times , while single strain growth was performed no less than three times in 96-well plates by taking OD600 measurements every half hour using a Tecan Infiniti Pro plate reader ( Tecan ) . Additionally , competition assays and single strain growth were performed in the absence and presence of varying concentrations of daptomycin to determine whether growth rates changed with increasing concentration according to expectations ( Fig 1 ) , which was always the case . Lastly , figures throughout the manuscript depict a typical growth curve for the specific condition or mutant , while Tables 1 and 2 list growth rates calculated over all experiments . RNA was isolated from cultures at different times using the Qiagen RNAeasy kit ( Qiagen ) . RNA was treated with the TURBO-DNAfree kit ( Ambion ) , after which cDNA was generated from 1 μg RNA with iScript complete kit ( BioRad ) and random hexamers . Quantitative PCR was performed using a BioRad MyiQ . Each sample was measured in both technical and biological triplicates , and samples were normalized against the 50S ribosomal gene SP2204 . Tn-Seq sequencing data is deposited at the Sequence Read Archive under BioProject PRJNA318012 . | While antibiotic resistant bacterial pathogens cause millions of deaths each year it remains largely unclear how a bacterium deals with antibiotic-induced stress and how this leads to the emergence of resistance . Moreover , many bacterial species are composed of strains whose genomes vary considerably , and while this variation may significantly affect phenotypes such as antibiotic sensitivity , its importance is unknown . Here we apply the method Tn-Seq , showing it is feasible to develop a detailed view of how a bacterium experiences antibiotic stress , while simultaneously determining the influence of the genomic-background . We show for two strains of the bacterial pathogen Streptococcus pneumoniae that , even though they experience the same stress triggered by daptomycin , they use a majority of different genes to withstand this stress , including genes important for integrity of the membrane , potassium uptake and protein turnover . Additionally , by untangling underlying genomic networks we unexpectedly expose large differences in genetic-interactions as well as transcriptional regulation . Our study provides not only a sensitive approach to untangle the influence of the genomic-background on phenotypes such as antibiotic sensitivity , but also highlights that this knowledge is instrumental in understanding how bacteria respond to environmental stress , which in turn influences the manner in which they evolve . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Methods"
] | [
"bacteriology",
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"genetic",
"networks",
"pathology",
"and",
"laboratory",
"medicine",
"pneumococcus",
"gene",
"regulation",
"pathogens",
"drugs",
"microbiology",
"antibiotic",
"resistance",
"regulator",
"genes",
"a... | 2016 | Strain Dependent Genetic Networks for Antibiotic-Sensitivity in a Bacterial Pathogen with a Large Pan-Genome |
A major bottleneck in understanding zoonotic pathogens has been the analysis of pathogen co-infection dynamics . We have addressed this challenge using a novel direct sequencing approach for pathogen quantification in mixed infections . The major zoonotic food-borne pathogen Campylobacter jejuni , with an important reservoir in the gastrointestinal ( GI ) tract of chickens , was used as a model . We investigated the co-colonisation dynamics of seven C . jejuni strains in a chicken GI infection trial . The seven strains were isolated from an epidemiological study showing multiple strain infections at the farm level . We analysed time-series data , following the Campylobacter colonisation , as well as the dominant background flora of chickens . Data were collected from the infection at day 16 until the last sampling point at day 36 . Chickens with two different background floras were studied , mature ( treated with Broilact , which is a product consisting of bacteria from the intestinal flora of healthy hens ) and spontaneous . The two treatments resulted in completely different background floras , yet similar Campylobacter colonisation patterns were detected in both groups . This suggests that it is the chicken host and not the background flora that is important in determining the Campylobacter colonisation pattern . Our results showed that mainly two of the seven C . jejuni strains dominated the Campylobacter flora in the chickens , with a shift of the dominating strain during the infection period . We propose a model in which multiple C . jejuni strains can colonise a single host , with the dominant strains being replaced as a consequence of strain-specific immune responses . This model represents a new understanding of C . jejuni epidemiology , with future implications for the development of novel intervention strategies .
Understanding the ecology of zoonotic pathogens in the animal host is crucial for controlling infections in humans [1 , 2] . Our knowledge is limited , however , with respect to within-host dynamics of pathogens . One reason for this is the lack of experimental models addressing the effect of co-infections on pathogen colonisation . Here , we present the application of a novel approach in which we can quantify mixed populations directly in infected material using direct sequencing and statistical analysis , without prior cultivation of bacterial isolates [3] . The aim of our work was to determine strain dependence and dynamics in a Campylobacter jejuni co-infection model in two different background floras , mature ( treated with Broilact , a product consisting of bacteria from the intestinal flora of healthy hens ) and spontaneous , using the direct sequencing approach . C . jejuni is a leading cause of diarrhoeal disease and food-borne gastroenteritis in humans . This bacterium is zoonotic and poultry is considered a major reservoir for transmission to humans [4] . C . jejuni is able to colonise the GI tract of chickens without causing any disease in the host [5 , 6] . The principal localisation of C . jejuni is the lower gastrointestinal tract , especially the caecum [7] . Multiple C . jejuni genotypes have been found in the GI tracts of individual chickens and within commercial broiler flocks [8–13] . It has also previously been shown , using antibiotic-resistant strains , that there can be interference in colonisation between pairs of C . jejuni strains [14] . The effects of co-infection dynamics and multiple strain infections , however , have not yet been described . This knowledge is important for our understanding of the epidemiology of Campylobacter , and for the development of intervention strategies that can prevent C . jejuni from entering the food chain . The seven isolates selected in our study were isolated from an epidemiological field experiment [15] . We have both field data and experimental infection data for the strains used , and we present evidence for a relatively rapid shift in the dominating C . jejuni strain between the ages of 27 and 30 days in our infection trials . We also show that this shift is relatively unaffected by the dominating microflora . The relevance of our findings is discussed in the context of the epidemiology and control of C . jejuni .
We found C . jejuni–positive chickens in three out of four farms using real-time PCR quantification . The positive flocks , at farms A , C , and D , were the same as those found in the study by Johnsen et al . [15] . The selected flocks at farms A and D became infected at 2 wk of age , while the flock at farm C became infected at 3 wk of age [15] . The number of C . jejuni–positive samples measured with real-time PCR was 53% at farm A , 38% at farm C , and 18% at farm D . The colonisation level relative to the total flora for the C . jejuni–positive chickens from farms A , C , and D were −3 . 89 log , −4 . 11 log , and −3 . 92 log ( with standard deviations of 0 . 85 , 0 . 85 , and 0 . 67 ) , respectively ( Figure S1 ) . Six of the C . jejuni–positive caecum samples were chosen for studying the diversity of C . jejuni isolates in the chicken caecum ( three chickens from each of the farms A and D ) . From both farms , samples were selected to represent high , medium , and low C . jejuni colonisation levels . Products from the amplification of the housekeeping genes gltA ( citrate synthetase; n = 62 ) , glnA ( glutamine synthetase; n = 80 ) , glyA ( serine hydroxymethyltransferase; n = 3 ) , and tkt ( transketolase; n = 1 ) were cloned and sequenced . Most polymorphic sites were found in the genes of gltA and glnA . The alignments showed that there were multiple genotypes present at the same time in all six chickens , and this was particularly the case for farm A , the chicken from which contained up to nine different genotypes of both gltA and glnA genes . A total of 127 C . jejuni strains from the field study [15] were screened with regard to two of the partial multilocus sequencing ( MLST ) housekeeping genes , gltA and glnA . We found a larger variety of genotypes of the gltA gene than of the glnA gene . However , comparison of the partial MLST with the amplified fragment length polymorphism ( AFLP ) typing by Johnsen et al . [15] showed good agreement with both the gltA and the glnA genotypes ( Figure S2 ) . We therefore chose the gltA gene for further analyses , as it had the highest degree of variation . The gene genealogies were estimated using the TCS software v1 . 21 based on statistical parsimony [16 , 17] . Sequences representing each of the genotypes detected in the cloned samples and from the strains from the field study were used in the analysis ( Figure 1 ) . Strain G110 ( not used in the infection model ) came out as the haplotype with the highest out-group probability . It is particularly interesting to note that we can see a separate grouping of sequences from the cloned samples from farm A . A BLAST search showed an approximately 90% similarity to the gltA gene of C . jejuni , which was the first hit on the result table . These unique sequences were not found in any of the strains from the field study . Seven C . jejuni strains were selected for use in the infection model; these strains were G10 , G12 , G98 , G109 , G114 , G125 , and G147 ( information about these strains are given in Table 1 , and marked in Figure 1 ) . The selection was based on specific mutations in the gltA gene , differences in the AFLP pattern , and the fact that strains from all farms were represented in the infection model . All chickens were infected with a dose of approximately 8 . 7 log10 colony-forming units ( cfu ) of the C . jejuni mixture . This is within the range expected when chickens eat infected faecal material since intestinal contents often harbour 5–9 log10 cfu per gram [18] . Colonised chickens appeared healthy and showed no signs of disease . Plate counts showed a caecal colonisation level of approximately 8 log10 cfu g−1 caecum during the infection period ( results not shown ) . The C . jejuni–specific quantitative real-time PCR amplification [19] gave an average colonisation level between −2 and −4 log values relative to the total flora in the caecum for the infected chickens ( Figure 2 ) . Analysis of variance ( ANOVA ) was performed on data from the real-time quantification using Minitab v14 . 2 . The factor variables treatment and day ( day 16 to 36 ) were tested against the colonisation level as response variable . The results showed a significant effect of treatment on the colonisation levels ( p-value < 0 . 001 ) . The mean colonisation levels of C . jejuni in Broilact-treated chickens were lower than the mean colonisation levels of C . jejuni in chickens with spontaneous background flora . This difference in colonisation level of C . jejuni between the two different groups is most evident at day 27 ( Figure 2 ) . We found that mainly two of the seven C . jejuni isolates colonised the caecum , isolates G109 and G125 . Other strains only colonised sporadically ( Figure 3 ) , but at levels above the detection limit ( p < 0 . 05 ) . In order to identify time trends and effects of Broilact treatment on G109 and G125 relative abundances , ANOVA was carried out . Since our response variables were proportions , the logit transform ( i . e . , log-odds ) was applied to the estimated relative abundances prior to modelling . Day of sampling was found to have a significant effect on both G109 and G125 abundances ( p < 0 . 001 in both cases ) , indicating a pronounced time trend . The analysis did not detect any significant effect of Broilact treatment , except as an interaction term with day of sampling ( p-values of 0 . 006 and 0 . 003 for G109 and G125 , respectively ) , indicating a significant influence of Broilact treatment on the time trend . Further inspection of the models showed that the interaction terms were significant only for day 27 in both cases ( p-values of 0 . 003 and 0 . 049 for G109 and G125 , respectively ) . Figure 3 shows an evident shift in abundances of the two isolates in Broilact-treated chickens sampled on day 27 , with a precipitous decline in G109 proportions and a corresponding upsurge of G125 . The shift persists until the end of the sampling period . The same shift can be observed for non-treated chickens , but not until day 30 . To further investigate the observed time trends , the logit transformed proportions were modelled as continuous functions of time . Both for G109 and G125 , models were fitted separately for Broilact- and non-treated chickens ( Figure S3 ) . For G109 , significant negative time trends were found within both the treated and non-treated groups ( p < 0 . 001 in both cases ) . For G125 , equally significant positive trends were found . This suggests constant change rates in relative abundance through time , but in opposite directions , for the two isolates . When treatment was included in the model as an interaction term , we did not find significant differences in constant change rates in Broilact- and non-treated chickens for either C . jejuni isolate . The direct sequencing method was also applied to duodenum and jejunum samples . The colonisation levels in these parts of the intestinal tract are low . The typing information obtained , however , showed the same colonisation trend as that of the caecum samples ( results not shown ) . The minimum spanning network in Figure 1 shows that the main colonisers , G109 and G125 , are both connected near the haplotype with the highest out-group probability , G110 . G109 is a sister group to dominating sequences from cloned samples , and this isolate belongs to a group of dominating sequences . This indicates that the isolate could be quite common in the environment of chickens . The other isolates used in the infection model were more distinct from the proposed out-group . A direct sequencing approach of a universally conserved region of the 16S rRNA genes was used for the classification of the total caecal flora [3] . Principal component analysis ( PCA ) was performed on the mixed spectra . The two first principal components ( PCs ) explained 83% of the total variance in the dataset ( Figure 4 ) . Adding more factors did not markedly increase the percentage explained of variation . The first PC , explaining 81% of the total variance , clearly separated samples collected from Broilact-treated and untreated chickens as determined by multiple linear regression ( MLR ) ( p < 0 . 001 ) . The signature sequence 5′-CAG ACG GCC TTT TAA GTC ANC TCT GAA AGT TTG CGG GTC AAC CGT AAA ATT-3′ ( corresponding to Escherichia coli position 581 to 631 ) deduced form the positive loading for PC1 ( Figure S4 ) showed that the Broilact-treated chickens were associated with bacteria belonging to Bacteroidetes , while the negative loading signature sequence 5′-TAN ACG GGA NAA GCN NGN CTG GAN TGA AA ACC CNG GGC TCA ACC CCG GGA CGT GCT TGT G-3′ ( E . coli position 581 to 637 ) showed that the untreated chickens were associated with Clostridia . The second PC , explaining 2% of the total variance in the data , correlated well with the time of sampling ( p < 0 . 001 by MLR ) . The untreated chickens had a better separation on the second PC than the Broilact-treated chickens . The loading plot for this component showed that Clostridia was associated with early colonisation ( signature 5′-TAG AGT GCN GGN GNG GTA NGA GGT T-3′ , E . coli position 652 to 675 ) , while Gammaproteobacteria ( signature 5′-AGA AGA GTA AAA NNC AAN AT-3′ , E . coli position 661 to 675 ) were associated with later time points .
In this study we demonstrated a shift in the dominating C . jejuni strains colonising chickens during the course of an infection . This shift was found to be independent of individual variation among chickens . Furthermore , we detected up to nine genotypes in a single chicken at the farm level . We have also shown that there was no major effect of the dominating microflora on the C . jejuni colonisation pattern . Plasmodium is , to our knowledge , the only pathogen that has been investigated in detail with respect to mixed strain/species infections [20] . A model proposed for Plasmodium is that a specific density-dependent adaptive host immune response prevents the outgrowth and suppresses the dominant strain/species . This response also suppresses the non-dominant strains/species in a non-specific manner . When the negative-selected dominant Plasmodium population drops below a certain threshold , then the density-dependent reaction is turned off , allowing the outgrowth of a new population that is not controlled by a specific immune response [21] . The colonisation pattern for C . jejuni determined by our confection trials can be explained using the Plasmodium model . It is also well known that C . jejuni has a variable surface antigen structure [22] , analogous to that of Plasmodium [23] . Variable surface antigen structures are an adaptation to avoid the host adaptive immune response . We propose that infections with multiple C . jejuni strains in nature allow for stable infections in a host with an adaptive immune response , such as chickens . Multiple strain infections could in fact be a general mechanism among pathogens to maintain stable infections . C . jejuni is a prokaryote pathogen with a main reservoir in the gut , whereas Plasmodium is a eukaryote blood parasite . Thus , it is likely that other pathogens in other reservoirs have developed similar mechanisms . The Broilact treatment had a major effect on the total microflora , while there was only a minor effect on C . jejuni colonisation . The Broilact effect on C . jejuni may not be directly linked to the microflora itself , but could rather be an indirect effect of the dominating microflora on the chicken host immune system . It is likely that the intestinal microflora derived from Broilact represents a lower burden on the host than the spontaneous microflora . This is supported by the findings that the spontaneous microflora was dominated by Clostridia , whereas the microflora derived from Broilact-treated chickens was dominated by Bacteroidetes . It is well known that Clostridia contain bacteria that are a burden to the host , while most bacteria belonging to Bacteroidetes are beneficial [24] . The immune systems in the Broilact group of chickens may therefore be more responsive than those in spontaneous group . This would explain both the lower colonisation levels and the advanced occurrence of shifts in the dominating C . jejuni strain . The reason why we do not favour an explanation related to a direct effect of the dominating microflora is the fact that C . jejuni infection courses are similar under completely different background flora regimes . It could of course be argued that it is the mucosal microflora and not the dominant microflora that is important in determining the C . jejuni colonisation pattern . This explanation , however , would require a rapid shift in the mucosal microflora corresponding to the C . jejuni shift . We find this unlikely . Thus , the most parsimonious explanation is that the chicken host is the most important factor in determining the C . jejuni colonisation pattern , and not the mucosal nor the dominating luminal microflora . There have been numerous trials using Broilact or other competitive exclusion ( CE ) approaches to combat C . jejuni [25–27] . The general conclusion from these trials is that it is possible to reduce , but not eliminate , C . jejuni by CE . The issue of CE , however , has not yet been properly addressed with respect to the host immune system , making it difficult to separate the potential CE effect from the host immune response . In particular , it would be interesting to know if CE approaches targeting mucosal surfaces also trigger the adaptive and/or innate immune system , and if the host responses are confounded with the proposed CE effects [13 , 26 , 28] . Until now technology has limited pathogen infection models to single or double-strain infections . Our seven-strain infection model points towards a novel understanding of the epidemiology of C . jejuni , which again could lead to a new way of thinking with respect to the development of intervention strategies . Certainly , investigating multiple strain infections will also be important for understanding the epidemiology of other bacterial pathogens , and for learning how to combat them .
Johnsen et al . [15] investigated four Norwegian broiler farms ( farms A–D ) for genetic diversity of Campylobacter in broilers and in the environment of broiler farms . These farms had a history of producing Campylobacter-positive broiler flocks , and samples were taken from 11 May to 14 September 2004 . The farms were visited nine times , three times prior to the broiler flock sampling , weekly during the 4-wk growing period of the selected flocks , and twice thereafter . During the growing period of the broiler flocks , broiler caecal material and five different sites inside the broiler house were sampled . A total of 144 Campylobacter spp . strains obtained from the sampling were typed using AFLP [15] . Of the 144 Campylobacter spp . strains , 127 were identified as C . jejuni . Seven of the C . jejuni strains were selected for use in the infection model ( described in “Experimental infections , ” below ) . Thirty whole caeca were sampled from each flock ( farms A–D ) at slaughter , and samples were frozen at −80 °C for further quantification and typing of colonising C . jejuni . The experimental infections were carried out at Foulum Research Centre ( Tjele , Denmark ) following Danish legislation for animal welfare and use of experimental animals . The chickens used for the experimental infections were conventional broiler chickens ( Ross 308 ) of mixed sex , purchased at 1-d-old from a local hatchery ( DanHatch ) . The C . jejuni–free animals were transferred directly from the hatchery to the experimental unit , where they were housed in isolators ( Montair Andersen B . V . HM 1500 ) . The four experimental groups were kept in separate isolators , and group size was 23–24 chickens , each 1 d of age . Two of the groups were treated with Broilact ( Orion Oyj ) at day 1 , while the two other experimental groups did not receive any microbial treatment . All four experimental groups were inoculated with a mixture of seven different C . jejuni strains at day 14 . The chickens were inoculated individually by crop instillation with 500 μl of the bacterial suspension ( approximately 9 log10 cfu/ml ) , using a 1-ml syringe with an attached flexible tube . Bacterial inoculum was prepared from cultures on blood agar base plates ( Oxoid ) supplemented with 5% ( v/v ) calf blood ( BA ) and incubated at 42 °C for 48 h under microaerobic conditions . Bacterial suspensions were prepared by shaking of bacterial material in 0 . 9% saline at 4 °C . For each strain , the bacterial suspension was adjusted to an optical density of approximately OD620 = 0 . 6 . According to the measured OD620 , the strains were mixed in equal concentrations . Cfu for the bacterial suspensions of each strain were determined ( Table 1 ) , and total cfu of the mixture were determined before and after inoculation , and the minimal colonisation dose was calculated as the mean of these counts . During the experiments , three chickens were removed from each group ( a total of six chickens per treatment ) twice a week after the inoculation until day 36 . The chickens were killed by decapitation , and each chicken was sampled and examined individually . Contents from caecum , duodenum , and jejunum were collected separately in tubes and stored at −80 °C . The separation of the small intestine into jejunum and ileum is often done at the Meckel's diverticulum , which is the site where the yolk sack is attached . This definition has been used in this experiment . C . jejuni counts were determined as cfu per gram of chicken caecum . The caecal contents were weighed and diluted in buffered peptone water . Ten-fold dilution series were made and streaked onto modified charcoal cefoperazone deoxycholate agar ( mCCDA ) plates ( CM 739 , Oxoid ) with selective supplement ( SR 155 , Oxoid ) . The plates were incubated microaerobically at 42 °C for 48 h . DNA isolation and purification of contents from caecum , duodenum , and jejunum were performed using an automated procedure as described earlier by Skånseng et al . [19] . Quantification of C . jejuni was performed relative to the total flora [19] . Universal 16S rDNA primers and probe [29] were used for quantification of the total flora . C . jejuni–specific real-time PCR was performed using the primer and probe set described by Nogva et al . [30] . Using AmpliTaq Gold DNA polymerase ( Applied Biosystems ) in the real-time PCR reaction , the mixture contained 1× TaqMan Buffer A ( Applied Biosystems ) , 5 mM MgCl2 , and 200 μM dNTP mix . Universal 16S rDNA PCR reactions contained 0 . 2 μM of each primer , 0 . 1 μM probe , 1 . 25 U AmpliTaq Gold DNA polymerase , and 1 μl DNA in a 25-μl PCR reaction . C . jejuni–specific PCR reactions contained 0 . 3 μM of each primer , 0 . 02 μM probe , 2 . 5 U AmpliTaq Gold DNA polymerase , and 4 μl DNA in a 50-μl reaction . With the use of DyNAzyme II Hot Start DNA Polymerase ( Finnzymes Oy ) in the real-time PCR , the reaction mixture contained 1× Hot Start Buffer ( Finnzymes ) , 0 . 5 μM ROX reference dye ( Invitrogen ) , and 200 μM dNTP mix . Universal 16S rDNA real-time PCR contained 0 . 2 μM of each primer , 0 . 1 μM probe , 1 U DyNAzyme II Hot Start DNA Polymerase , and 0 . 5 μl DNA in a 25-μl PCR reaction . C . jejuni–specific real-time PCR reactions contained 0 . 3 μM of each primer , 0 . 02 μM probe , 1 U DyNAzyme II Hot Start DNA Polymerase , and 2 μl DNA in a 25-μl reaction . The amplification profile was 40 cycles of 95 °C for 30 s and 60 °C for 1 min , with an initial heating step of 95 °C ( AmpliTaq Gold ) or 94 °C ( DyNAzyme II Hot Start ) for 10 min . The reactions were performed in an ABI PRISM 7900 HT Sequence Detection System ( Applied Biosystems ) and the data were analysed using the SDS 2 . 2 Software ( Applied Biosystems ) . Amplification of the C . jejuni housekeeping genes , glnA , gltA , glyA , and tkt was performed using primers [31] listed in Table S1 . The PCR amplification reactions contained 1× PCR Buffer II ( Applied Biosystems ) , 5 mM MgCl2 , 200 μM dNTP mix , 0 . 2% BSA ( New England Biolabs ) , 2 . 5 U AmpliTaq Gold DNA polymerase ( Applied Biosystems ) , 0 . 2 μM of each primer , and 5 μl DNA in a 25-μl reaction . The amplification profile was an initial step of 95 °C for 10 min , then 40 cycles of 95 °C for 30 s , 50 °C for 2 min , and 72 °C for 30 s , and a final extension at 72 °C for 7 min . PCR products were cloned into a plasmid vector using a TOPO TA Cloning Kit ( Invitrogen ) as previously described by Rudi et al . [32] . Cells with insertions were amplified using primers HU ( 5′-CGC CAG GGT TTT CCC AGT CAC GAC G-3′ ) and HR ( 5′-GCT TCC GGC TCG TAT GTT GTG TGG-3′ ) . The PCR products were purified before sequencing . This was done by adding 10 U of Exonuclease I and 2 U of Shrimp Alkaline Phosphatase ( USB Corporation ) to 8 μl of PCR product . The thermal profile was 37 °C for 15 min and 80 °C for 15 min . The sequencing reaction contained 0 . 75× BigDye v1 . 1/3 . 1 Sequencing Buffer , 1 μl BigDye Terminator v3 . 1 Cycle Sequencing Kit , 0 . 32 μM of primer M13 , and 3 μl of purified PCR product in a 10-μl reaction . The sequencing reaction was carried out in 25 cycles of 96 °C for 15 s , 50 °C for 10 s , and 60 °C for 4 min . Purification of the sequence products was performed with the Montage SEQ96 Sequencing Reaction Cleanup Kit ( Millipore ) using a Biomek 2000 Workstation ( Beckman Coulter ) . Ten microliters of Hi-Di Formamide ( Applied Biosystems ) was added to the purified sequence products . Sequencing was performed on an ABI PRISM 3100 Genetic Analyzer ( Applied Biosystems ) . Typing of C . jejuni isolates for use in the infection model . Amplification of the partial C . jejuni housekeeping genes glnA and gltA ( Table S1 ) was performed on isolates from the study of Johnsen et al . [15] . The PCR amplification reactions contained 1× PCR Buffer II ( Applied Biosystems ) , 5 mM MgCl2 , 200 μM dNTP mix , 2 . 5 U AmpliTaq Gold DNA polymerase ( Applied Biosystems ) , 0 . 2 μM of each primer , and 1 μl DNA in a 25-μl reaction . The amplification profile was an initial step of 95 °C for 10 min , then 30–40 cycles of 95 °C for 30 s , 50 °C for 2 min , and 72 °C for 30 s , and a final extension at 72 °C for 7 min . Sequencing was performed using 0 . 5 μM of primers gln1F or glt1F . Further sequencing was performed as described in the section “Multilocus sequence typing ( MLST ) ” . Typing of C . jejuni from the infection model . Amplification of the C . jejuni gltA genes in the caecum samples from the experimental infection was performed using glt1F and glt1R [31] ( Table 1 ) . The PCR amplification reactions contained 1× Hot Start Buffer ( Finnzymes ) , 200 μM dNTP mix , 1U DyNAzyme II Hot Start DNA Polymerase ( Finnzymes ) , 0 . 2 μM of each primer , and 1 μl DNA in a 25-μl reaction . The amplification profile was an initial step of 95 °C for 10 min , then 35 cycles of 95 °C for 30 s , 50 °C for 2 min , and 72 °C for 30 s , and a final extension at 72 °C for 7 min . The PCR products were purified before sequencing ( as described in “Typing of C . jejuni isolates for use in the infection model” ) . The sequencing reaction contained 0 . 75× BigDye v1 . 1/3 . 1 Sequencing Buffer , 2 μl BigDye Terminator v3 . 1 Cycle Sequencing Kit , 0 . 25 μM of primer glt1F , and 3 μl of purified PCR product in a 20-μl reaction . The sequencing reactions were carried out in 25 cycles of 96 °C for 15 s , 50 °C for 10 s , and 60 °C for 4 min . Precipitation of the sequencing products was performed using an ethanol/EDTA- procedure ( Applied Biosystems ) . Five microliters of 125 mM EDTA and 60 μl of 96% ethanol were added to the sequencing products and the reactions were mixed by inverting the 96-well plate four times . The products were incubated at room temperature for 15 min before centrifuging at 4 , 500 rpm for 45 min at 4 °C . The supernatants were then removed , and the inverted plate was spun up to 1 , 000 rpm . Sixty microliters of 70% ethanol were added to the products and the mixtures were centrifuged at 4 , 500 rpm for 30 min at 4 °C . The supernatants were removed by inverting the plate , and the plate was spun at 1 , 000 rpm for 1 min . The pellets were resuspended in 14 μl of Hi-Di Formamide . Sequencing was performed on an ABI PRISM 3100 Genetic Analyzer . Quantification of C . jejuni isolates colonising chickens . For estimating relative strain abundances in the caecal samples , we used MLR analysis of mixed sequence electropherograms according to the linear mixture model [3] . Briefly , this entails modelling the DNA sequence spectrum from a mixture of homologous gene fragments as a linear combination of the pure sequence spectra constituting the mixture . In our case , there were seven strains of C . jejuni , giving us the following model: yj = βj1x1 + βj2x2 + βj3x3 + βj4x4 + βj5x5 + βj6x6 + βj7x7 + ɛj , where yj is a mixed spectrum , x1 , . . . , x7 are the pure strain spectra , β1 , . . . , β7 are regression coefficients , and ɛj is an error term . According to the linear mixture model , the regression coefficients may be interpreted as relative amounts of the strain corresponding to spectrum xi ( i = 1 , . . . , 7 ) [33] , if the system is additive . Additivity was tested by applying the model to a test set , and no serious deviations from a linear relationship between response and covariates were found . For the analysis we used spectral data from 14 SNPs in the gltA gene , extracting spectral readings at the point of base calling , as well as three flanking readings on each side . The resulting ( 98×4 ) spectral matrices were re-scaled by multiplying all values within blocks of ( 7×4 ) ( i . e . , emission readings for one polymorphic site ) with the ratio between the total spectral mean and the block mean . The matrices were subsequently unfolded and mean normalised [3] . DNA isolated from caecum was amplified with universal 16S rDNA primers [29] . The PCR mixture contained 0 . 2 μM of each primer , 1 U DyNAzyme II Hot Start DNA Polymerase , 1× Hot Start Buffer , 200 μM dNTP mix , and 1 . 0 μl DNA in a 25-μl PCR reaction . The amplification profile was an initial step of 94 °C for 10 min , then 30 cycles of 94 °C for 30 s , 60 °C for 30 s , and 72 °C for 30 s , and a final extension at 72 °C for 7 min . The PCR products were purified before sequencing , using 0 . 4 μl of ExoSap-IT ( USB Corporation ) to 5 μl of PCR product . Thermal profile was 37 °C for 30 min and 80 °C for 15 min . The sequencing was performed using a universally conserved primer U515F [34] with C-tail extension ( U515Fc30 ) , consisting of 30 bases on the 5′-end . The sequencing reaction contained 0 . 75× BigDye v1 . 1/3 . 1 Sequencing Buffer , 1 μl BigDye Terminator v1 . 1 Cycle Sequencing Kit , 0 . 32 μM of primer U515Fc30 , and 0 . 5 μl of purified PCR product in a 10-μl reaction . The sequence reaction was performed by 25 cycles of 96 °C for 15 s and 60 °C for 4 min . Precipitation of the sequence products was performed using BigDye XTerminator Purification Kit ( Applied Biosystems ) , according to instructions supplied by the manufacturer . Sequencing was performed on an ABI PRISM 3100 Genetic Analyzer ( Applied Biosystems ) . The most appropriate spectral region in 16S rRNA gene for data analysis was found to be between the conserved start sequence 5′-ATTTANTGGGT-3′ and end sequence 5′-GAATTCNNNGTGTA-3′ , covering a region corresponding to nucleotide positions 565 to 677 in the E . coli 16S rRNA gene . We found this region to comprise enough DNA sequence dissimilarity to distinguish main groups of bacterial DNA sequences found in the mixtures . The trimmed DNA sequence spectra were imported into Unscrambler software v9 . 6 ( CAMO Software ) and analysed using PCA [35] . PCA is used to separate essential information from noise in data with many variables and thus allows viewing the analysed data as easily interpretable plots . The scores separate the samples analysed , while the loading plot shows which parts of the sequence spectra are important for the separation of the samples . The correlation analyses were done using MLR based on the orthogonal PCA scores . The significance of the correlations was determined using ANOVA . These analyses were done using the Unscrambler software . We used empirically determined signature sequences identified in the loading plot to identify bacterial that differ in the samples analysed . The signature sequences were subsequently assigned to a hierarchical taxonomy using Probe Match in the Ribosomal Database Project II ( http://rdp . cme . msu . edu/ ) .
The housekeeping genes gltA , glnA , glyA , and tkt from the six caecum samples studied from the field study were sequenced and deposited in GenBank ( http://www . ncbi . nlm . nih . gov/Genbank/index . html ) under accession numbers EF546072 , EF546074 , EF546076–EF546138 , EF546140 , EF546151–EF546155 , EF546157–EF546169 , EF546171–EF546185 , EF546187–EF546230 , EF546232 , EF546233 , and EF546235 . | Pathogenic bacteria that can be transferred from animals to humans represent a highly potent human health hazard . Understanding the ecology of these pathogens in the animal host is of fundamental importance . A major analytical challenge , however , is the fact that individual animal hosts can be colonised by multiple strains of a given pathogen . We have addressed this challenge by developing a novel high-throughput approach for analyses of mixed strain infections . We chose Campylobacter jejuni colonisation of the chicken gastrointestinal ( GI ) tract as a model . C . jejuni is a major cause of food-borne disease in humans , and chickens are considered a main reservoir from which this bacterium may enter the food chain . We analysed the co-colonisation of seven C . jejuni strains in two groups of chickens with very different background GI microfloras . We found that mainly two of the C . jejuni strains colonised the chickens , with a shift in the dominant coloniser during the infection period . The C . jejuni colonisation pattern , however , was little affected by the dominating GI microflora . We propose a model where the chicken immune response is the important determinant for C . jejuni colonisation , and suggest that multiple strain colonisation could be a way of maintaining stable infections in the animal host . This new knowledge is very important for future development of novel intervention strategies to prevent C . jejuni from entering the human food chain . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"chicken",
"ecology",
"public",
"health",
"and",
"epidemiology"
] | 2007 | Co-Infection Dynamics of a Major Food-Borne Zoonotic Pathogen in Chicken |
Autocrine priming of cells by small quantities of constitutively produced type I interferon ( IFN ) is a well-known phenomenon . In the absence of type I IFN priming , cells display attenuated responses to other cytokines , such as anti-viral protection in response to IFNγ . This phenomenon was proposed to be because IFNα/β receptor1 ( IFNAR1 ) is a component of the IFNγ receptor ( IFNGR ) , but our new data are more consistent with a previously proposed model indicating that regulated expression of STAT1 may also play a critical role in the priming process . Initially , we noticed that DNA binding activity of STAT1 was attenuated in c-Jun−/− fibroblasts because they expressed lower levels of STAT1 than wild-type cells . However , expression of STAT1 was rescued by culturing c-Jun−/− fibroblasts in media conditioned by wild-type fibroblasts suggesting they secreted a STAT1-inducing factor . The STAT1-inducing factor in fibroblast-conditioned media was IFNβ , as it was inhibited by antibodies to IFNAR1 , or when IFNβ expression was knocked down in wild-type cells . IFNAR1−/− fibroblasts , which cannot respond to this priming , also expressed reduced levels of STAT1 , which correlated with their poor responses to IFNγ . The lack of priming in IFNAR1−/− fibroblasts was compensated by over-expression of STAT1 , which rescued molecular responses to IFNγ and restored the ability of IFNγ to induce protective anti-viral immunity . This study provides a comprehensive description of the molecular events involved in priming by type I IFN . Adding to the previous working model that proposed an interaction between type I and II IFN receptors , our work and that of others demonstrates that type I IFN primes IFNγ-mediated immune responses by regulating expression of STAT1 . This may also explain how type I IFN can additionally prime cells to respond to a range of other cytokines that use STAT1 ( e . g . , IL-6 , M-CSF , IL-10 ) and suggests a potential mechanism for the changing levels of STAT1 expression observed during viral infection .
Although type I and type II interferons ( IFNs ) have distinct roles in immune responses , there is substantial overlap between the genes and cellular responses they regulate . It has been known for some time that many cells secrete small priming quantities of type I IFNs that facilitate more potent responses to subsequent stimuli [1]–[3] . Moreover , cellular responses to CSF-1 or IFNγ can be affected by neutralizing type I IFN antibodies or knockout of type I IFN-Receptors ( IFNAR ) [2] , [4] , [5] . Notably , the protective anti-viral effects of IFNγ were much less potent in IFNAR1−/− than wild-type fibroblasts which appeared to be caused by a lack of type I IFN priming [4] , [5] . The molecular events that underpin these priming events have not been fully characterized , although it has been proposed that type I and II IFNs shared receptor components [5] . However , as the majority of responses to type I and II IFNs require the expression of the STAT1 transcription factor [6] , this is also a possible point of crosstalk between them . STAT1 is a key mediator of cytokine-induced gene expression as it is activated either as homo- or heterodimer with other STATs by many cytokines including type I and type II IFNs , interleukin ( IL ) -6 and IL-10 . STAT1 activity is of particular importance to the IFN system as STAT1−/− mice display many similar phenotypes to mice lacking IFNAR1 or the IFN Receptor ( IFNGR ) 1 . In particular , anti-viral , anti-mycobacterial , and anti-tumor responses are compromised [6]–[9] . Induction of STAT1 expression is a potential explanation for the priming activity of type I IFN because it is an IFN-stimulated gene ( ISG ) itself [10]–[12] and its 5′ promoter region contains an IRF/gamma activated sequence ( GAS ) element bound by IFN-stimulated transcription factors [13] . Inducing the expression of STAT1 would increase the pool of this factor available for activation by IFNγ . Consistent with such a hypothesis , low expression of STAT1 correlated with IFN-resistance in melanoma samples when compared to surrounding normal tissue [14] . In unstimulated cells , STAT1 resides in the cytoplasm as a latent factor that is activated by a series of post-translational modifications initiated when it is recruited to cytokine receptors following receptor ligation [15] . At the receptor , STAT1 is phosphorylated on tyrosine 701 , by Janus family kinase ( JAK ) s , which facilitates its dimerization either with other STAT1 molecules or other STAT proteins depending on the cytokine receptor . In addition , STAT1 proteins are phosphorylated on serine 727 prior to nuclear translocation which is essential for their full transcriptional activity [16] . Conversely , STAT1 activity is negatively regulated by phosphatases , SOCS proteins , and the SUMO ligase Protein Inhibitor of Activated STAT ( PIAS ) 1 [15] . Recently , in the course of our studies on IFNγ-activated AP-1 DNA binding , we noticed that IFNγ-induced GAS DNA binding was suppressed in c-Jun−/− cells compared to wild-type cells [17] and this correlated reduced levels of STAT1 in c-Jun−/− cells . The level of STAT1 expression in c-Jun−/− murine embryonic fibroblasts ( MEFs ) were restored to wild-type levels following culture in media conditioned by wild-type fibroblasts suggesting that c-Jun deficiency caused the disruption of an autocrine/paracrine loop that regulated STAT1 expression . The STAT1-inducing component of media conditioned by wild-type fibroblasts was IFNβ , because the activity could be blocked by neutralizing antibodies directed against type I IFN and antibodies used were raised against IFNAR and attenuated by targeted knockdown of IFNβ by RNA interference ( RNAi ) . While c-Jun has been demonstrated to co-operate with ATF-2 , IRF-3 , and NFκB for virus-induced production of IFNβ [18] , to our knowledge our studies are the first to demonstrate that c-Jun is necessary for basal expression of low-level IFNβ . Fibroblasts in which this autocrine/paracrine loop was disrupted by the loss of components of type I IFN receptors also express lower levels of STAT1 . As many biological functions of IFN require STAT1 [6] , [7] , this suggested that previous observations of attenuated responses to IFN in IFNAR1−/− cells may be related to the reduced STAT1 expression that has been observed [19] . Consistent with this hypothesis , restoring STAT1 expression in IFNAR1−/− fibroblasts rescued IFNγ-induced gene transcription and anti-viral properties . In summary , this study provides evidence of an autocrine/paracrine stimulatory loop that requires the expression of c-Jun , IFNβ , and IFNAR to regulate the expression of STAT1 . Importantly , this basal IFNβ production occurs via a mechanism distinct from the pathogen-stimulated IFNβ production mediated by IRF and NFκB pathways [18] . One model to explain crosstalk between type I and II IFNs states that type I and II IFN-R physically interact in a ligand-dependant manner , such that the presence of type I IFNs is essential for a fully competent IFNγ response [5] . Herein , we demonstrated that attenuated IFNγ-mediated gene induction and an associated defective anti-viral response to IFNγ that is observed in IFNAR1-deficient cells can be rescued by re-expressing STAT1 and is therefore independent of IFNAR1 . We propose that an alternative model to explain the functional synergy between type I and II IFNs is based on the regulated expression of STAT1 via c-Jun-mediated production of basal levels of IFNβ .
In the course of our studies of IFN-induced signaling and gene expression , we performed elecrophoretic mobility shift assays ( EMSAs ) assessing GAS binding species in nuclear extracts from IFNγ-stimulated wild-type and c-Jun−/− MEFs . A GAS binding complex was detected in both wild-type and matched c-Jun−/− MEFs following 15–30 min of exposure to IFNγ , however in the absence of c-Jun , IFNγ-induced GAS binding activity was markedly attenuated ( Figure 1A ) . The decrease in GAS binding activity in c-Jun−/− MEFs was a consequence of reduced expression of STAT1 . Both STAT1 mRNA and protein were ∼10-fold lower in c-Jun−/− MEFs compared to wild-type cells ( Figure 1B and C ) . However , expression of STATs was not globally affected , as expression of STAT3 , another GAS-binding transcription factor , remained unchanged ( Figure 1C ) . Reduced STAT1 expression was not a clone-specific phenomenon as similar results were obtained using an independently derived matched pair of wild-type and c-Jun−/− MEFs ( Figure S1 ) . To determine if c-Jun could regulate STAT1 levels by inducing the secretion of a soluble factor that acted in autocrine/paracrine fashion to induce STAT1 expression , conditioned media from wild-type or c-Jun−/− MEFs were cultured in ( i ) fresh media , ( ii ) media conditioned by c-Jun−/− MEFs , or ( iii ) media conditioned by wild-type MEFs . Cells were harvested after 16 h of culture in conditioned media and STAT1 mRNA and protein expression was assessed . Expression of STAT1 mRNA and protein was unaltered in wild-type MEFs cultured in fresh media or conditioned media from wild-type or c-Jun−/− MEFs ( Figure 2A and B ) . In c-Jun−/− MEFs , basal expression of STAT1 was much lower than in wild-type cells and was not increased when the cells were cultured in either fresh media or conditioned media from c-Jun−/− MEFs ( Figure 2A and B ) . In contrast , when c-Jun−/− MEFs were cultured in media conditioned by wild-type MEFs , STAT1 mRNA and protein expression was induced almost to the levels observed in wild-type cells ( Figure 2A and B ) . These data confirmed that fibroblasts secrete a c-Jun-dependent soluble factor that induces STAT1 expression through an autocrine/paracrine feedback loop . Type I IFN is constitutively secreted from unstimulated fibroblasts and can induce STAT1 expression [10] . To determine if type I IFN was the STAT1-inducing active component of fibroblast conditioned media , c-Jun−/− MEFs were cultured in either fresh or conditioned media from wild-type cells in the presence of a type I IFN blocking antibody [20] . STAT1 expression was increased in c-Jun−/− MEFs cultured in conditioned media from wild-type cells in the presence of control antibodies ( Figure 3A ) and this enhanced expression was entirely blocked by the presence of type I IFN neutralizing antibodies used at concentrations capable of neutralizing ∼5 IU/mL IFNβ . Additional studies ( Figure 3B and C ) revealed that the STAT1-inducing activity of wild-type-conditioned media was almost ablated by a blocking mAb raised against IFNAR1 [21] . Together , these data demonstrate that type I IFN is a component of conditioned media from wild-type cells that is necessary for the rescue of STAT1 expression in c-Jun−/− cells . It has been reported that STAT1 levels are diminished in IFNβ−/− cells [22] indicating that IFNβ could be the key component of the conditioned media from wild-type cells shown to induce expression of STAT1 in c-Jun−/− MEFs . Treatment of c-Jun−/− MEFs with doses as low as 1 IU/mL IFNβ induced STAT1 mRNA and doses between 5 and 10 IU/mL were sufficient to restore STAT1 mRNA and protein expression to levels seen in wild-type cells ( Figure S2A and B ) . STAT1 mRNA levels were slightly increased in wild-type MEFs treated with IFNβ ( Figure S2C ) , which is consistent with studies demonstrating that STAT1 expression is induced in fibosarcoma cell lines treated with IFNα or β [12] and in splenic leukocytes where STAT1 levels were increased following virus infection in a type I IFN-dependent manner [11] . Comparison of the levels of expression of IFNβ mRNA in wild-type and c-Jun−/− cells revealed that c-Jun−/− MEFs expressed ∼50% of the wild-type levels of IFNβ mRNA ( Figure 4A ) . AP-1 sites are known to be important for inducible expression of IFNβ [23] , but little is known of what regulates constitutive production of type I IFN in unstimulated cultured fibroblasts . Chromatin immunoprecipitation ( ChIP ) assays on unstimulated wild-type and c-Jun−/− MEFs demonstrated a >2-fold increase in c-Jun bound to the murine IFNβ promoter when compared to Ig control samples ( Figure 4B ) . Together , these data imply that expression of c-Jun and subsequent occupation of the IFNβ promoter by c-Jun is required for basal secretion of IFNβ . To determine if IFNβ was the type I IFN necessary to maintain STAT1 expression , we used RNAi to knock down IFNβ in wild-type MEFs ( Figure 4C ) and assessed the ability of conditioned media from these cells to induce the expression of STAT1 mRNA in c-Jun−/− MEFs . As expected , STAT1 mRNA levels were greater when c-Jun−/− MEFs were cultured in conditioned media from wild-type MEFs or from MEFs expressing a control knockdown vector than if these cells were cultured in fresh media ( Figure 4D ) . In contrast , the ability of conditioned media from wild-type cells with RNAi-mediated knockdown of IFNβ to induce STAT1 expression in c-Jun−/− MEFs was significantly reduced ( Figure 4D ) . These data confirm that IFNβ is expressed by unstimulated wild-type fibroblasts and is necessary for the maintenance of STAT1 expression . As disruption of autocrine/paracrine stimulation by IFNβ affected the level of STAT1 expression in c-Jun−/− MEFs , we predicted that cells lacking either chain of the type I IFN receptor would also express less STAT1 than wild-type cells . Primary MEFs ( Figure 5A ) and splenocytes ( Figure 5B ) from either IFNAR1−/− or IFNAR2−/− ( unpublished data ) mice expressed significantly lower levels of STAT1 than wild-type cells . We extended these studies to compare the expression of STAT1 across multiple tissues in wild-type versus IFNAR1−/− mice . As shown in Figure 5C , the levels of STAT1 were consistently reduced in all tissues from IFNAR1−/− mice compared to their wild-type counterparts , suggesting this defect may have broad physiological importance . Interestingly expression of STAT2 was also reduced in IFNAR11−/− MEFs while the levels of STAT3 were unaffected by knockout of the type I IFN receptor ( Figure S3A ) . Our model predicted that , unlike c-Jun deficiency that affected production of an autocrine stimulus , IFNAR1 deficiency affects responses to the autocrine stimulus . In support of this model , wild-type-conditioned media was able to rescue the expression of STAT1 in c-Jun−/− MEFs , but in IFNAR1−/− MEFs STAT1 expression was unaffected by culture in wild-type-conditioned media ( Figure S3B ) . These data support the existence of an autocrine loop involving IFNβ that regulates basal STAT1 expression levels and suggest that defects in any part of this loop are likely to affect the expression of STAT1 . STAT1 is important for not only IFNα/β signaling but also the signaling of several other cytokines , including IFNγ [15] . The expression of approximately two-thirds of IFNγ-induced genes is dependent upon STAT1 expression , however not all IFNγ-mediated biological responses are entirely dependent on STAT1 expression [17] , [24] . It has previously been reported that IFNAR1−/−cells are refractory to IFNγ treatment due to the proposed interaction between IFNAR1 and IFNGR [5] . To determine if decreased expression of STAT1 may confer the observed decrease of IFNγ-mediated responses in IFNAR1−/− cells , STAT1 levels were restored in these cells by retroviral transduction ( Figure 6A ) . GAS binding activity was assessed by EMSA using nuclear extracts from IFNγ-treated wild-type MEFs , IFNAR1−/− MEFs , and IFNAR1−/− MEFs reconstituted with empty vector ( IFNAR1−/− MSCV ) or STAT1 ( IFNAR1−/− HA-STAT1 ) . Consistent with previous studies [5] , [25] , IFNγ induced less GAS binding in IFNAR1−/− cells than wild-type cells ( Figure 6B ) . This low level of GAS binding was also observed in cells transduced with empty vector but was rescued in cells reconstituted with HA-STAT1α . These data demonstrated that the reduced GAS binding observed in IFNAR1−/− cells was caused by reduced STAT1 expression rather than being a direct consequence of IFNAR1 deficiency . Previous studies demonstrated that IFNγ-induced gene expression was attenuated in IFNAR1−/− cells [5] . We therefore assessed the impact of re-expression of STAT1α in IFNAR1−/− cells upon the IFNγ-induced expression of genes such as β-2-microglobulin and SOCS3 that require STAT1 expression [26] . Both genes were induced in response to IFNγ in wild-type cells , although with differing kinetic profiles , but induction was weak or absent in IFNAR1−/− cells . IFNγ-induced expression of both β-2-microglobulin and SOCS3 was restored in cells that re-expressed HA-STAT1α , but not in cells transduced with an empty vector ( Figure 6C , D ) . Similar results were observed when other IFNγ-responsive genes were tested ( Figure S4 ) . In order to determine whether the reduced levels of STAT1 in IFNAR1−/− cells could affect biological responses to IFNγ , we investigated whether re-expression of STAT1 in IFNAR1−/− cells impacted upon the ability of IFNγ to protect them against infection by the cytopathic virus murine encephalomyocarditis virus ( EMCV ) . Wild-type , IFNAR1−/− , IFNAR1−/− MSCV , and IFNAR1−/− HA-STAT1 MEFs were infected with a dose of virus sufficient to induce 100% lysis of wild-type MEFs in the presence or absence of various doses of IFNγ , and the cytopathic effects were determined by assessing cell viability after 24 h . As was shown previously [5] , the ability of IFNγ to protect cells from EMCV-mediated lysis was significantly reduced in IFNAR1−/− MEFs when compared to wild-type MEFs at most doses of IFNγ and the concentration of IFNγ ( 500 IU/ml ) required to provide 80% protection from the virus for IFNAR1−/− cells was much greater than that required to provide a similar level of protection for wild-type cells ( 10 IU/ml ) . The response of IFNAR1−/− MEFs transduced with empty vector to IFNγ was not significantly different from the untransduced IFNAR1−/− MEFs at any dose of IFNγ and the concentration of IFNγ required to provide 80% protection ( 450 IU/ml ) was of a similar order of magnitude ( Figure 7 ) . In contrast , protection from virus-induced lysis was significantly enhanced in IFNAR1−/− HA-STAT1 MEFs at most doses of IFNγ . These data provide direct evidence that the attenuated protective anti-viral responses to IFNγ observed in IFNAR1−/− cells is a consequence of reduced STAT1 expression .
Herein we demonstrate that c-Jun is essential for the constitutive production of small quantities of IFNβ that initiates autocrine or paracrine feedback loops required to maintain the expression of STAT1 ( Figure 8 ) . This system was disrupted either by c-Jun deficiency , which prevents production of IFNβ , or by IFNAR deficiency , which affects the ability of cells to respond to the autocrine stimulus . Consistent with our data , others found that cells lacking IFNβ also express much lower levels of STAT1 [22] and virus-mediated induction of STAT1 is dependent on type I IFN signaling [11] . As IFNγ signaling is attenuated when the autocrine stimulus is blocked ( Figure 8 ) but restored by adding back STAT1 , it appears the level of STAT1 expressed by the cell determines the response of the cell to other cytokines . These results suggested the ability of IFNγ to induce a protective anti-viral state was due to the type I IFN-mediated maintenance of STAT1 expression rather than the recruitment of IFNAR1 into the IFNγR complex as has been previously proposed [5] . These findings define a novel mechanism through which STAT1-mediated signals can be regulated and highlight the importance of crosstalk between type I and II IFNs for anti-viral immunity . It has been known for some time that , as well as being produced in large quantities following viral infections , cells can secrete low levels of type I IFN constitutively [1] , [2] , [27] . Virus-induced activation of the IFNβ enhanceasome is one of the best-characterized transcriptional modules [18] , [23] . Viral activation of the IFNβ promoter involves the binding of NFκB , IRF3 , and ATF2/c-Jun complexes to a series of DNA elements termed PRD I-IV [23] . In this setting , c-Jun binds to PRD IV of the promoter and facilitates co-operative binding of the other factors . Removing PRD IV from the promoter , or even reversing its orientation , has a major impact on the transcriptional activity of the promoter [23] , suggesting the role of c-Jun is critical in the context of viral infection . In contrast , little is known of the molecular mechanisms of constitutive type I IFN production . Our study indicates that PRD IV of the IFNβ promoter is occupied by c-Jun even in “resting” cultured cells ( Figure 4B ) . This requirement for c-Jun explains why we found that constitutive IFNβ production and hence the expression of STAT1 was attenuated in c-Jun−/− cells . In addition to regulating basal expression of IFNβ , we have recently demonstrated that c-Jun is activated following IFNγ treatment and may also play a direct role in regulating the expression of a subset of IFNγ-responsive genes ( ISGs ) [17] . Indeed we identified ISGs that were dependent on c-Jun for induction by IFNγ , others that required STAT1 , and others that required both c-Jun and STAT1 for increased expression following treatment with IFNγ [17] . These results , coupled with the functional data provided herein , highlight the complex molecular interplay between c-Jun and canonical mediators of type I and II IFN signaling such as STAT1 in regulating a comprehensive response to IFN treatment . Takaoka and colleagues previously demonstrated the importance of IFNβ in the production of an IFNγ-mediated anti-viral response [5] . In that paper the authors showed that IFNβ−/− MEFs were defective in mounting an IFNγ-induced antiviral response . These data mirror what we have demonstrated herein where we show that IFNAR1−/− MEFs show a similar defect in mounting an IFNγ-induced antiviral response . However , we showed that restoring STAT1 expression in IFNAR1−/− cells significantly rescued the ability of IFNγ to protect cells against EMCV , suggesting that regulating the levels of STAT1 expression through the autocrine loop may play an important role in responses to this challenge . The ability of type I and II IFNs to co-operate , for example , in treatment of melanoma tissue [28] or priming of macrophage cytotoxicity [29] has long been recognized . Interestingly , at a cellular level , IFNAR1−/− cells were known to have an anomalously poor response to IFNγ with respect to induction of GAS DNA binding , induction of gene expression , and protection against the cytopathic effects of EMCV [4] , [5] , [25] . IFNγ function is not entirely compromised in IFNAR1−/− animals because IFNGR1−/− mice have distinct phenotypic differences from IFNAR1−/− mice [30] . Inhibiting autocrine priming by type I IFN does not only affect signaling by IFNγ . Therefore its is not surprising that IL-6 signaling [31] and CSF-1 signaling are affected by inhibiting priming by type I IFN [2] and that signals induced by IL-10 can be affected by priming with IFNs [32] . It was proposed that the ligand-bound IFNAR1 chain acts as a component of the IFNGR and promotes recruitment of STAT1 to the IFNGR because IFN receptors are clustered within caveolar membrane fractions to facilitate their association [5] . Such a hypothesis is inconsistent with mapping of the docking site of STAT1 to the IFNAR2 chain of the type I IFN-R rather than the IFNAR1 as specified by the shared receptor model [33] . We demonstrated herein that IFNAR1−/− cells express lower basal levels of STAT1 relative to wild-type controls ( Figure 5 ) , and as STAT1 is a critical mediator of IFN signaling , this is an alternative reason why these cells may lack sensitivity to IFNγ . Our model not only explains the inability of IFNγ to prime IFNAR1−/− cells for an anti-viral response and the rescue of IFNγ function in IFNAR1−/− cells by STAT1α expression but also the attenuated responses to other cytokines , such as IL-6 and CSF-1 , observed in IFNAR1−/− cells [2] , [31] and predicts they may also be rescued by expression of STAT1 . As IFNγ function was not entirely recovered following re-expression of STAT1 in IFNAR−/− cells , we cannot exclude that the shared receptor mechanism makes a contribution , but there are other reasons why reconstitution of STAT1 may not have fully rescued IFNγ function . These include the absence of other as yet unidentified signal transducing proteins from cells of this genotype . The level of STAT1 expression in cells can have functional consequences with respect to immune responses . In response to viral infection , Ag-specific CD8+ T cells express peak levels of STAT1 for a shorter period of time than CD4+ cells [34] . This decreased sensitivity to IFN-induced growth inhibition allows expansion of Ag-specific CD8+ cells while the proliferation of cells with higher STAT1 is inhibited [34] . The relative amounts of different STATs can also affect the biological responses to cytokines . For example STAT1:3 and STAT1:4 ratios have been shown to alter cellular responses , and thus regulating the levels of these transcription factors will affect the outcome of immune responses [11] , [35] . Our previous studies revealed that loss of IFN signaling abrogated the immune-mediated neo-natal lethality of SOCS1−/− deficiency [36] , and more recently we discovered that deleting IFNAR1 also rescued this pathology [37] to a level equivalent to SOCS1−/− IFN+/− . Although SOCS1 directly regulated type I IFN signaling , another reason why IFNAR1 deficiency can protect SOCS1−/− animals may be the similarities and crosstalk between type I and II IFN signaling pathways . These data highlight the patho-physiological importance and mechanism of crosstalk between type I and II IFN that are important considerations in understanding the contributions of individual cytokines to host defense and in their therapeutic targeting .
c-Jun−/− , IFNAR1−/− [38] , IFNAR2−/− [39] , and wild-type matched MEFs were derived from embryos and either used as early passage primary MEFs or immortalized by the “3T3” method . IFNAR1−/− MEFs were transduced with Murine stem cell leukemia virus supernatants encoding GFP alone or cDNA encoding HA-tagged STAT1 ( generous gift from Thomas Decker ) . Supernatants were produced by transient transfection of PhoenixE cells with MSCV vector by calcium phosphate precipitation using standard methods . c-Jun−/− MEFs were transduced with lentiviral supernatants encoding miR sequences targeting IFNβ ( Open Biosystems , Huntsville , AL , USA; product numbers: RMM4431-98755134 , RHS4346 ) . Supernatants were produced by transient transfection of 293T cells with pGIPZ vector using Lentiphos HT kit ( Clontech , Mountain View , CA , USA ) according to the manufacturer's instructions . Cells were cultured in DMEM supplemented with 5% foetal bovine serum ( JRH Biosciences , Lenexa , KS , USA ) and 2 mM L-Glutamine ( JRH Biosciences , Lenexa , KS , USA ) . All tissue culture reagents were certified sterile and free of Mycoplasma and pyrogens . Antibodies for the following targets were used: STAT1 ( BD Biosciences Franklin Lakes , NJ , USA ) , STAT3 ( Santa Cruz Biotech , Santa Cruz , CA , USA ) , HA ( Cell Signaling Technology , Beverly , MA , USA ) , c-Jun ( Santa Cruz Biotech , Santa Cruz , CA , USA ) , α-tubulin ( Sigma Chemical Co . , St . Louis , MO , USA ) , and hsp70 ( Clone N6 was a kind gift of Dr . Robin Anderson; Peter MacCallum Cancer Centre , Melbourne , Australia ) . Neutralizing anti type I IFN [20] and anti-IFNAR1 were described previously [40] . HRP-conjugated secondary antibodies were purchased from Dako ( Glostrup , Denmark ) . 3×106 MEFs were cultured in 175 cm2 tissue culture flasks in 20 mL media for 3 d . Supernatant was collected , cell debris removed by centrifugation ( 670 g , 4 min ) , sterilized using a 0 . 22 µM filter , and stored at 4°C . Western blotting was performed as previously described [17] . Briefly , cells were washed , resuspended in whole cell lysis buffer ( 50 mM Tris-HCl pH 8 , 0 . 1% Triton X-100 , 150 mM NaCl , 0 . 1 mM EDTA , 0 . 1 mM EGTA , 10% glycerol , 1 µg/mL aprotinin , 0 . 5 µg/mL leupeptin , and 0 . 2 mM PMSF ) , and after ( 4°C 10 min ) lysates were cleared by centrifugation . Proteins were separated by SDS-PAGE , transferred to immobilon P membranes ( Millipore ) , and probed with specific antibodies . Secondary antibodies were conjugated to horseradish peroxidase and images were visualized by chemiluminescence ( ECL , GE Healthcare , Bucks , UK ) . Nuclear extractions and EMSAs were performed as previously described [17] . Briefly , cells were resuspended in hypotonic lysis buffer ( 10 mM HEPES , 1 . 5 mM MgCl2 , 10 mM KCl , and protease inhibitors ) ( 4°C , 5 min ) , NP-40 was added to a final concentration of 0 . 25% , and the nuclei isolated by centrifugation ( 2 , 000 g , 10 min ) . Nuclei were resuspended in hypertonic lysis buffer ( 5 mM HEPES pH 8 , 1 . 5 mM MgCl2 , 0 . 2 mM EDTA , 0 . 5 M NaCl , 25% glycerol , and protease inhibitors ) ( 4°C , 1 h ) . For binding reactions 5–10 µg of nuclear lysate was incubated ( 4°C , 30 min ) with 5×104 cpm T4 PNK-32P-labeled oligonucleotides in binding buffer ( 20 mM Tris/HCl pH 8 , 6 mM KCl , 2 mM MgCl2 , 12% Glycerol , 5 µM DTT , 2 . 5 µg polydI . dC . polydI . dC , and 0 . 05% NP-40 ) . Complexes were separated by 5% native PAGE , and gels were dried and visualized by autoradiography on X-ray film ( Kodak ) . The sequence of the GAS oligonucleotide 5′-TAGGGATTTACGGGAAATTGATGAAGCTGATC-3′ was derived from the FcγRI promoter; the AP-1 oligonucleotides were described previously [17] . RNA was extracted using Trizol ( Invitrogen , Carlsbad , CA , USA ) according to the manufacturer's instructions . cDNA was synthesized from 2 µg RNA using superscript III ( Invitrogen , Carlsbad , CA , USA ) as per the manufacturer's instructions . The abundance of specific genes in the samples was quantitated using the SYBR Green dye detection method ( Applied Biosystems , Foster City , CA , USA ) . Primers to murine GBP-1 ( 5′-TGTGGTTGCTGGATGAGCAGAGTA-3′; 5′-AAGGAAACACAGTAGGCTGGAGCA-3′ ) , SOCS3 ( 5′-CCTTCAGCTCCAAAAGCGAG-3′; 5′-GCTCTCCTGCAGCTTGCG-3′ ) , and IFNβ ( 5′-AGCTCCAAGAAAGGACGAACAT-3′; 5′-GCCCTGTAGGTGAGGTTGATCT-3′ ) were designed using Primer Express 2 software ( Applied Biosystems , Foster City , CA , USA ) . Primers to murine STAT1 gene ( 5′-CGCGCATGCAAGTGGCATATAACT-3′; 5′-AAGCTCGAACCACTGTGACATCCT-3′ ) were designed using PrimerQuest software ( Integrated DNA Technologies ) . Primers to ribosomal L32 ( 5′-TTCCTGGTCCACAACGTCAAG-3′; 5′-TGTGAGCGATCTCGGCAC-3′ ) were as previously described [17] . Threshold cycle numbers ( Ct ) were measured in the exponential phase for all samples . Relative abundance of sample genes was calculated using the ΔΔCt method relative to the L32 control gene [17] . mRNA abundance was normalized to the untreated samples of each genotype . ChIP assays were performed as described previously [41] using 5 µg of anti c-Jun or rabbit IgG control antibodies . The abundance of specific sequences in ChIP samples was quantitated using the SYBR Green dye detection method ( Applied Biosystems , Warrington , UK ) . Primers used for PCR reactions were mIFN PRDIV ( 5′-ATTCCTCTGAGGCAGAAAGGACCA; 5′-GCAAGATGAGGCAAAGGCTGTCAA ) and were designed using Primer Express 2 software . Threshold cycle values ( Ct ) were measured in the exponential phase , and promoter occupancy was calculated using the formula 2 ( Ct Ig − Ct c-Jun ) . Statistical significance was tested using one-way ANOVA testing with OriginLab 7 . 5 software ( Northampton , MA , USA ) or Prism Software Graphpad ( La Jolla , CA , USA ) . 103 cells of each genotype were plated in duplicate wells in a 96 well plate and allowed to adhere . Media was replaced with fresh media containing murine EMCV ( M . O . I of 0 . 1 ) and various concentrations of IFNγ ( 0–1 , 000 IU/mL ) and cultured for 16 h . As controls , cells were cultured in fresh media alone ( 100% survival ) or with EMCV alone ( 0% survival ) . Cells were washed in PBS , formalin fixed ( 10 min at RT ) , washed ( twice with PBS ) , and stained in 0 . 5% Crystal Violet/20% methanol . Stained cells were extensively washed , crystal violet was solubilized in 10% acetic acid , and OD550 nm was recorded . Viability was calculated by comparison against a standard curve . | Cells of the immune system release interferons ( IFNs ) in response to pathogens or tumor cells; these proteins signal to other immune cells to initiate the body's defense mechanisms . The two classes of IFNs—types I and II—have different receptors and distinct effects on the cells; however , there is “crosstalk” between them . In particular , small quantities of type I IFN can “prime” cells to produce a robust response to type II IFN . In this paper , we provide evidence to explain the molecular basis of this crosstalk . We show that continuous expression of the transcriptional activator c-Jun is responsible for producing basal , priming levels of a type I IFN; this signals to immune cells with the type I IFN receptor ( IFNAR1 ) to maintain expression of STAT1 inside these cells . STAT1 is a key factor for immune cell responses to type II IFN . Thus , signaling by low levels of type I IFN primes the cells with sufficient STAT1 to respond robustly to a subsequent type II IFN signal . This work provides an alternative explanation of the priming phenomenon to a previous proposal that the ligand-bound type I receptor , IFNAR1 , acts as a component of the type II IFN receptor . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology/cell",
"signaling",
"biochemistry/chemical",
"biology",
"of",
"the",
"cell",
"immunology/immune",
"response",
"infectious",
"diseases/viral",
"infections",
"immunology/immunity",
"to",
"infections",
"cell",
"biology/gene",
"expression"
] | 2010 | Functional Crosstalk between Type I and II Interferon through the Regulated Expression of STAT1 |
Clonorchiasis , caused by the infection of Clonorchis sinensis ( C . sinensis ) , is a kind of neglected tropical disease , but it is highly related to cholangiocarcinoma and hepatocellular carcinoma ( HCC ) . It has been well known that the excretory/secretory products of C . sinensis ( CsESPs ) play key roles in clonorchiasis associated carcinoma . From genome and transcriptome of C . sinensis , we identified one component of CsESPs , severin ( Csseverin ) , which had three putative gelsolin domains . Its homologues are supposed to play a vital role in apoptosis resistance of tumour cell . There was significant similarity in tertiary structures between human gelsolin and Csseverin by bioinformatics analysis . We identified that Csseverin expressed at life stage of adult worm , metacercaria and egg by the method of quantitative real-time PCR and western blotting . Csseverin distributed in vitellarium and intrauterine eggs of adult worm and tegument of metacercaria by immunofluorence assay . We obtained recombinant Csseverin ( rCsseverin ) and confirmed that rCsseverin could bind with calciumion in circular dichroism spectrum analysis . It was demonstrated that rCsseverin was of the capability of actin binding by gel overlay assay and immunocytochemistry . Both Annexin V/PI assay and mitochondrial membrane potential assay of human hepatocarcinoma cell line PLC showed apoptosis resistance after incubation with different concentrations of rCsseverin . Morphological analysis , apoptosis-associated changes of mitochondrial membrane potential and Annexin V/PI apoptosis assay showed that co-incubation of PLC cells with rCsseverin in vitro led to an inhibition of apoptosis induced by serum-starved for 24 h . Collectively , the molecular properties of Csseverin , a molecule of CsESPs , were characterized in our study . rCsseverin could cause obvious apoptotic inhibition in human HCC cell line . Csseverin might exacerbate the process of HCC patients combined with C . sinensis infection .
Clonorchis sinensis ( C . sinensis ) has been proven to be the causative agent of clonorchiasis , which is endemic in China , Korea and Vietnam [1] , [2] , [3] . As an important food-borne parasite , C . sinensis has afflicted more than 35 million people in world and approximately 15 million in China , creating a socio-economic burden in epidemic regions [4] . Most clonorchiasis cases are due to the consumption of raw freshwater fish containing infective C . sinensis metacercariae , which excyst in the duodenum until they grow into juvenile C . sinensis and then migrate into the bile ducts of their host [5] , [6] . Both experimental and epidemiological evidence have implied that long-term infections with liver flukes lead to chronic pathological changes , including hepatomegaly , hepatic fibrosis , cholangitis , cholecystitis , adenomatous hyperplasia , and cholangiocarcinoma ( CCA ) [7] , [8] , [9] . Furthermore , C . sinensis was recently classified along as a Group I biological carcinogen by the World Health Organization [10] , [11] . In endemic area of China , 16 . 44% of HCC patients were infected with C . sinensis , while 2 . 40% were infected in non-tumor patients . The OR value and 95% CI in HCC group were 8 . 00 and 4 . 34–14 . 92 [12] , [13] , [14] , so that we should pay high attention to the relationship between primary hepatocellular carcinoma and the infection of C . sinensis . It has been well known that the excretory/secretory products of C . sinensis ( CsESPs ) can cause histopathological changes such as bile duct dilatation , inflammation and fibrosis , and adenomatous proliferation of the biliary epithelium [15] . In the present studies , from the published genome [16] and transcriptome [17] , [18] of C . sinensis , we identified one component of CsESPs , Csseverin , which has three putative gelsolin domains . The gelsolin superfamily is conserved in mammalian as well as in non-mammalian organisms and takes the leading role in controlling actin organization or actin filament remodeling . The family has some specific and apparently non-overlapping particular roles in several cellular processes , including cell motility , control of apoptosis and regulation of phagocytosis [19] . Initial evidence of anti-apoptotic effect of gelsolin was provided by the observation that a point mutation in mouse gelsolin confers on this protein tumor-suppressor activity against H-ras oncogene transformed NIH-3t3 cells [20] , [21] . Direct evidence of the inhibitory role of gelsolin was provided by Ohtsu et al . , who generated Jurkat transfectants expressing up to threefold gelsolin than wild-type cells . These transfectants exhibited a phenotype more resistant to apoptosis induced by several stimuli [22] . Moreover , it has been reported that human cytoplasmic gelsolin can prevent apoptotic mitochondrial changes such as mitochondrial membrane potential loss by binding to mitochondrial voltage-dependent anion channel ( VDAC ) [23] . Large-scale gene sequencing efforts have revealed gelsolin homologues in the majority of parasitic phyla [24] , [25] , [26] , [27] , [28] . In the current study , we presented for the first time the molecular characteristics of Csseverin . We described the detection of recombinant Csseverin ( rCsseverin ) binding to cytoskeletal actin filaments of human hepatocarcinoma PLC cells and investigated its potential anti-apoptotic role on PLC cells as an ingredient of CsESPs in vitro . The present study is a cornerstone for researches on biological characterization of Csseverin in the future . In addition , our work will provide an exploratory sight view of mechanism involved in progress of carcinoma associated with the infection of C . sinensis .
C . sinensis flukes were isolated from naturally infected cats ( Guangdong Province , China ) for sample preparation . Animals in experiments were all purchased from animal center of Sun Yat-sen University and raised carefully in accordance with National Institutes of Health on animal care and the ethical guidelines . All experimental procedures were approved by the animal care and use committee of Sun Yat-sen University ( Permit Numbers: SCXK ( Guangdong ) 2009-0011 ) . PLC and human normal hepatocyte L-02 cells were a gift from Dr . Wang Shutong and Dr . Xie wenxuan ( the first affiliated hospital of Sun Yat-Sen University ) and routinely cultured in high glucose DMEM medium ( Gibco , USA ) supplemented with 10% fetal bovine serum ( Gibco , USA ) and penicillin–streptomycin ( 100 units/ml ) in 5% CO2 at 37°C . Serum-starved PLC were prepared by incubating the cells in high glucose DMEM medium at 37°C and 5% CO2 with fetal bovine serum deprivation for at least 24 h . The gene ( GenBank accession No . GAA30384 . 2 ) predicted encoding homologue of severin was screened from C . sinensis genome by blastx and Open Reading Frame ( ORF ) Finder program at NCBI ( http://www . ncbi . nlm . nih . gov ) . The alignment of its deduced amino acid sequences with homologues from other species were analyzed and shown with Vector NTI . Proteomics bioinformatics tools such as Motif-Scan , InterPro-Scan and Swiss-Model were used to analyze the protein characteristics including physicochemical parameters , conserved domains and spatial structure . The phylogenetic tree was constructed online ( http://www . ebi . ac . uk/Tools/clustalw/index . html ) . The ORF of severin was amplified using the following primers: sense: 5′- ATAGGATCCATGCCGGAGTACT -3′ ( underlined , BamHI ) and antisense: 5′- CGCAAGCTTTCATTCGAGAACC-3′ ( underlined , Hind III ) . The PCR was carried out for 32 cycles at 94°C for 45 s , 51°C for 45 s , and 72°C for 45 s , and extension for 10 min at 72°C after the last cycle in a DNA-Thermal Cycler ( Biometra , Germany ) . PCR products were purified and digested with BamHI and Hind III , and then subcloned into prokaryotic expression vector 6×His tag pET28a ( + ) ( Novagen , Germany ) . After digestion with BamHI and Hind III , the recombinant plasmid was confirmed by DNA sequencing and then transformed into E . coli , BL21 ( Promega , USA ) . The expression of rCsseverin was induced by 1 mM isopropyl-β-D-thiogalactopyranoside ( IPTG ) for 5 h at 37°C . After induction , the bacteria were harvested by centrifuging at 4°C for 15 min at 8 , 000×g and suspended in lysis buffer ( 0 . 5 M NaCl , 20 mM Tris–HCl , 5 mM imidazole , pH 8 . 0 ) , sonicated on ice , and centrifuged at 10 , 000×g for 15 min at 4°C . The fusion protein was batch-purified using His Bind Purification kit ( Novagen , USA ) and the eluted fractions containing rCsseverin were pooled and dialyzed with phosphate-buffered saline ( 10 mM phosphate buffer , 27 mM KCl , 137 mM NaCl , pH 7 . 4 ) . Protein samples were subjected to 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis ( SDS-PAGE ) and visualized by Coomassie brilliant blue G-250 , the concentration was measured by a Bicinchoninic acid assay ( BCA , Novagen , USA ) according to manufacturer's instructions . Then , 100/50 µg of rCsseverin were mixed with an equal volume of incomplete Freund's adjuvant and injected subcutaneously to six-week-old male Sprague-Dawley ( SD ) rats ( purchased for experiments under the Guide for the Care and Use of Laboratory Animals ) . Boost injections were given at 2 and 5 weeks after first injection . Anti-serum was collected at 1 week after the second booster , then aliquoted and stored in −80°C . Sera from naïve rats were also collected for using as control . CsESPs and sera from CsESPs immunized rat were obtained by referring to previous study [29] . 10 µg of rCsseverin or CsESPs were subjected to 12% SDS-PAGE and transferred to polyvinylidene fluoride ( PVDF ) membranes . Successively , the membranes were blocked with 1% bovine serum albumin in phosphate-buffered saline ( PBS ) overnight at 4°C , washed five times with PBS-0 . 05% Tween 20 ( PBS-T , pH 7 . 4 ) , and incubated with His-tag monoclonal antibody , sera from naïve rats , rCsseverin immunized rats , C . sinensis-infected rats or CsESPs immunized rats ( 1∶100 dilutions ) followed by HRP-conjugated goat anti-mouse/rat IgG ( Proteintech; dilution of 1∶2 , 000 ) at 37°C for 2 h . After adequately washing with PBS-T , the membrane was incubated with horseradish peroxidase ( HRP ) -conjugated goat anti-rat IgG in 1∶2000 dilutions ( Proteintech , USA ) at 37°C for 1 h . Detection was then carried out by enhanced chemiluminescence ( ECL ) method . Intact living adult worms were collected from biliary tracts of infected cats and washed extensively and gently in physiological saline to remove any contamination from hosts . Eggs and metacercariae were also collected as described previously [30] , [31] . They were stored in sample protector ( Takara ) at −80°C for RNA/DNA extraction or 4% formaldehyde for immunofluorescence assay . Total RNA was extracted from each sample using TRIZOL reagent ( Invitrogen , USA ) according to manufacturer's instructions , and total RNA was treated with DNase ( Promega , USA ) to remove any contaminated DNA . Their total cDNA were obtained by the method of reverse transcription PCR by using Reverse Transcriptase XL ( TaKaRa ) and Oligo18 primer referred to the manuals . Severin RNA was detected with SYBR Premix Ex Taq Kit ( TaKaRa , Japan ) according to the manufacturer's protocol . Real-time PCR was conducted in the BIO-RADiQ5 instrument ( BioRad , USA ) using specific primers ( sense: 5′-TACAGCACCGTGAAGTAGATGG-3′; antisense: 5′- CAGACCGTGACAGAGTAGCAGA-3′ ) . β-actin from C . sinensis ( GenBank accession No . EU109284 ) was used as an internal control [32] , which was amplified with the primers ( forward primer: 5′-ACCGTGAGAAGATGACGCAGA-3′ , reverse primer: 5′-GCCAAGTCCAAACGAAGAATT-3′ ) designed by primer premier 5 . 0 . The transcripts of Csseverin were detected using SYBR Premix Ex Taq Kit ( TaKaRa , Japan ) according to the manufacturer's protocol . PCR was carried out in a total volume of 20 µl , consisting of 2 µl cDNA , 10 µl SYBR Premix Ex Taq ( 2× ) , 0 . 4 µl Severin forward and reverse primer ( 10 µM ) , and 7 . 2 µl RNase-free distilled H2O . The real-time PCR program consisted of an initial denaturation step at 95°C for 30 s , 45 cycles of 95°C for 5 s , and 60°C for 20 s . The real-time PCR amplification was conducted in the BIO-RADiQ5 instrument ( BioRad , USA ) . To complete the protocol , a melting curve was constructed using the following program: 95°C for 30 s , 65°C for 15 s , followed by increase to 95°C while continuously collecting fluorescence signal . Semiquantitative analysis as performed by the comparative 2−ΔΔCt method [33] . The total proteins of adult worms , metacercariae , and eggs were respectively homogenized in RIPA lysis buffer ( containing 1 mM proteinase inhibitor PMSF , Biotech , USA ) followed by centrifugation at 10 , 000×g for 15 min . 20 µg of total proteins from each life cycle stage were separated on SDS-PAGE ( 12% gel ) and electro-transferred onto PVDF membrane . The membrane was blocked with 1% bovine serum albumin in PBS overnight at 4°C , washed with PBS-T , and incubated with anti-Csseverin rat serum ( 1∶100 dilutions ) or pre-immune rat serum ( 1∶100 dilutions ) at 37°C for 2 h . After extensively washing with PBS-T , the membrane was incubated with HRP-conjugated goat anti-rat IgG in 1∶2000 dilutions ( Proteintech , USA ) at 37°C for 1 h . Detection was then carried out by ECL . Fresh adult worms and metacercariae of C . sinensis were fixed with 4% formaldehyde , embedded with paraffin wax , and sliced into 4-µm-thick sections . After dewaxing and dehydration , slides were blocked with goat serum overnight at 4°C , and incubated with anti-rCsseverin sera ( 1∶100 in 0 . 1% PBS-T ) at room temperature for 2 h . Sera from naïve rats were used as a negative control . The slides were washed twice and incubated with goat anti-rat IgG labeled with red fluorescent Cyanine dye 3 ( Cy3 , Proteintech; 1∶400 in 0 . 1% PBS-T ) . Fluorescence microscopy was used in visualization of antibody staining . As the protein contains a potential Ca2+-binding domain , Ca2+-binding will change its conformation of secondary structure which can be detected by CD [34] , [35] , [36] . CD measurements were carried out on a J-810 Circular Dichroism Spectrometer ( Jasco , Japan ) with the Jasco Spectra Manager software at room temperature . Three samples were assayed: purified rCsseverin in PBS , purified rCsseverin in PBS containing 1 µM CaCl2 , and purified rCsseverin in PBS containing 1 µM EDTA to remove combined Ca2+ during expression of rCsseverin in bacteria and purification in solutions . Secondary structure was analyzed using Jasco Spectra Manager Secondary Structure Analysis program . Far-UV CD spectrum was acquired using a 0 . 2 mm path length cell at 0 . 2 nm intervals over the wavelength range from 190 to 250 nm . Three scaning values were averaged for each sample and were corrected by subtracting buffer contribution from parallel spectra in the absence of Csseverin . The concentration of Csseverin was kept at 1 µM in 10 mM sodium phosphate buffer pH 7 . 4 and then the CD data were converted to molar units . Gel overlay assay and immunocytochemistry were employed to investigate the actin binding activity of rCsseverin . F-actin ( from rabbit muscle , 99% similar to human F-actin , Sigma-Aldrich ) and its fragments digested with 0 . 25% trypsin ( Sigma-Aldrich , USA ) at 37°C for 1 h , were separated on 12% SDS-PAGE and electrophoretically transferred onto PVDF membranes . Membranes then were blocked with TBS-T ( 25 mM Tris-HCl , pH 7 . 2 , 50 mM NaCl , 0 . 5% Tween-20 ) containing 5% BSA overnight at 4°C and washed ( 3 times , for 15 min each ) in TBS-T . Then , membranes were incubated with 0 . 1 mg/ml rCsseverin in TBS-T for 1 h at room temperature . After washing extensively , membranes were incubated with anti-Csseverin rat serum ( 1∶100 dilutions ) in TBS-T for 1 h at room temperature . The membranes were incubated with 1∶2000 HRP-conjugated secondary antibodies against rat IgG in TBS-T for 1 h at room temperature after washing . Following extensive washing in TBS-T , the membranes were at last incubated with diaminobenzidine substrate solution to develop color after washing again [37] . In immunocytochemistry assay , the PLC cells were seeded into sterile Petri dish ( Nest , diameter of 15 mm ) which is special for the detection of laser scan confocal microscopy , at a density of 2×104 cells per well and then cultured for 24 h . The PLC cells were washed four times with PBS and then fixed with 2 ml of 4% paraformaldehyde solution in PBS at room temperature for 30 min , then treated with 50 mM NH4Cl for 10 min , to reduce aldehyde groups . The cells were permeabilized for 4 min at 4°C with 0 . 3% Triton X-100 in PBS . At the next step , cells were incubated in PBS buffer containing 3% of BSA for 1 h , followed by coated with rCsseverin overnight at 4°C . To visualize cytoskeleton , cells were incubated overnight at 4°C with mouse anti human F-Actin monoclonal antibody ( AbD Serotec , UK ) diluted 1∶1000 , then subsequently incubated overnight at 4°C with rat anti-rCsseverin serum ( 1∶100 ) for 12 h at 4°C . The incubation with secondary antibodies was carried out at RT for 2 h , using fluorescein isothiocyanate ( FITC ) -conjugated goat anti-mouse IgG ( Proteintech , USA ) diluted 1∶200 and Cyanine dye 3 ( Cy3 ) -conjugated goat anti-rat IgG ( Proteintech , USA ) diluted 1∶400 at the same time . All antibodies were diluted with 1% BSA in PBS buffer and all steps described above were preceded by intensive washes in PBS . After finally washing with water , cover dishes were mounted on slides with Hoechst 33258 ( Sigma , USA ) . By contrast , to visualize whether rCsseverin could bind with cytoskeletal actin filaments in vitro , PLC cells were serum-starved overnight after incubating 24 h in standard conditions , and coated with rCsseverin in DMEM with 2% FBS for 48 h before fixed with 4% paraformaldehyde solution . The following steps were similar with that mentioned above previously . Images were finally obtained with the LSM 710 laser scanning confocal microscope ( Zeiss ) . After being induced spontaneous apoptosis by serum-starved for 24 h and treated with rCsseverin at different concentrations of 10 , 20 , 40 , 80 µg/ml and PBS for 48 h , 1–5×105 PLC cells were collected by centrifugation , and then incubated with Annexin V/propidium iodide ( PI ) , provided by the Apoptosis Detection Kit ( Lankebio , China ) . The cells were washed twice in PBS and resuspended in 500 µl of 1×Binding Buffer before being incubated with 5 µl of Annexin V and 10 µl of PI . The cells were then analyzed by using flow cytometry after incubation for 5–10 min in dark . Early apoptotic cells were stained with AnnexinV alone whereas necrotic and late apoptotic cells were stained with both Annexin V and PI . PLC cells ( 5×104 cells per well ) were seeded into a 6-well culture plate and cultured as described above . After treatment with Apoptosis Inducers ( Beyotime , Chain ) , the cells were washed twice with PBS , permeabilized with 0 . 3% Triton in PBS , and stained with Hoechst 33258 for 5 min in dark . Morphologic changes in apoptotic nuclei were observed and photographed under the inverted fluorescence microscope ( Leica DMI4000B , Germany ) with emission wavelength at 460 nm and excitation wavelength at 350 nm . MMP assay kit ( Beyotime , China ) with JC-1 probe was used to measure MMP in PLC cells . Briefly , cells were seeded in six-well plates overnight and serum-starved for 24 h , then treated with various concentration of rCsseverin for 48 h . The cells were then washed with ice-cold PBS and incubated in a 5% CO2 humidified incubator at 37°C for 20 min after adding 1 ml of JC-1 working solution . The supernatant was then discarded and the cells were washed twice with JC-1 staining buffer . Next , 2 ml medium was added to each well and MMP was monitored using an inverted fluorescence microscope ( Leica DMI4000B , Germany ) and laser scanning confocal microscope ( Zeiss LSM 710 , Germany ) . The red JC-1 fluorescence was observed at 525 nm excitation ( Ex ) /590 nm emission ( Em ) and the green cytoplasmic JC-1 fluorescence was observed at 485 nm Ex/530 nm Em . Quantitative changes of MMP at the early stage of cell apoptosis were measured by flow cytometry with JC-1 probe . After being incubated with 10 , 20 , 40 and 80 µg/ml of rCsseverin for 48 h , 1–5×105 cells were harvested and resuspended with ice-cold PBS ( 1 , 500 rpm×5 min ) . Then , the cell suspensions were incubated with 0 . 5 ml JC-1 working solution in 0 . 5 ml DMEM for 20 min at 37°C . The staining solution was removed by centrifugation . The cells were washed with JC-1 ( 1× ) washing buffer twice , then resuspended in 500 µl JC-1 ( 1× ) staining buffer and detected by flow cytometer ( Bechman Coulter Gallios , USA ) . All of the experiments were repeated at least three times . Experimental values were obtained from three independent experiments with a similar pattern and expressed as means ± standard deviation ( SD ) . Statistical analyses were performed using SPSS software package 17 . 0 . Data were analyzed by one-way analysis of variance ( ANOVA ) followed by least significant difference ( LSD ) for comparison between control and treatment groups . Significance was set at p value<0 . 05 .
The ORF of Csseverin contained 1077 base pairs ( bp ) encoding a protein of 358amino acids ( predicted MW 40 . 88 kDa , pI 5 . 24 ) . Blastx analysis showed that the deduced amino acid sequence was homologous to gelsolin of Schistosoma mansoni , Schistosoma japonicum , Suberites domuncula , Echinococcus granulosus , Strongylocentrotus purpuratus and Hydra magnipapillata with 54% , 65% , 50% , 65% , 48% , 47% identities respectively . The amino acid sequence had no N-terminal signal peptide or transmembrane domain . According to MotifScan and InterproScan prediction , there were three gelsolin domains ( aa51–133 , aa171–247 , aa278–354 ) indicating that Csseverin might have similar role with gelsolin superfamily . Furthermore , we inferred that the location of putative actin binding surface of Csseverin was from 50 to 150 amino acids by Gene Ontology analysis ( http://www . geneontology . org/ ) . The nuclear magnetic resonance ( NMR ) derived structure of human ( Homo sapiens ) gelsolin ( PRF: 225304 ) was used as the template to build a molecular model of Csseverin . The two proteins shared 36% identity among their gelsolin core domains and there was significant similarity between their tertiary structures ( Figure S1 ) . Csseverin grouped very closely with Schistosoma japonicum ( Figure S2 ) , a parasite that increases the risk of HCC incident when associated with positive hepatitis B surface antigen [38] . The Csseverin was also closely relative to severin/gelsolin from Echinococcus granulosus , followed by Dictyostelium discoideum , but far from those of H . sapiens and M . musculus . The soluble rCsseverin was expressed with 6×His-tag in E . coli BL21 after induced by 1 mM IPTG at 37°C for 5 h . The purified recombinant protein showed a single band around 45 kDa ( including His-tag sequence ) in 12% SDS-PAGE , consistent with the predicted molecular mass ( Figure S3 , lane 7 ) . The final protein concentration was 0 . 8 mg/L . The anti-rCsseverin serum was collected from immunized rat . Purified rCsseverin could be recognized by rat anti-rCsseverin serum , anti-His tag monoclonal antibody , serum from C . sinensis-infected rat and serum from CsESPs-immunized rat at 45 kDa , while not blotted with serum from naïve rat . The CsESPs was probed by rat anti-rCsseverin serum at about 45 kDa . However , no band was detected by serum from naïve rat ( Figure 1 lanes 1–6 ) . Csseverin were detected to express at life stage of metacercaria , egg and adult worm of C . sinensis , but at different levels . Statistically significant differences of transcripts were detected among metacercaria , egg and adult worm when normalized by β-actin . The transcription level of Csseverin in egg was about 60 times higher than that in adult worm ( Figure 2A ) . The expression level of Csseverin was consistant with the transcriptional level . Egg has the highest expression level of Csseverin protein , followed by adult worm and metacercaria ( Figure 2B ) . The analysis of immunofluorescence localization by using rat anti-rCsseverin serum showed that in C . sinensis adult intensive fluorescences were observed in vitellarium while scattered fluorescences were detected in tegument . In metacercaria , specific fluorescences were only deposited in tegument . In addition , intensive fluorescences were presented in intrauterine eggs of adult worm ( Figure 3D , F and J ) . By comparison , no specific fluorescence was detected either in adult worm or in metacercaria when treated with serum from naïve rat ( Figure 3B , H ) . According to the profile of CD spectrum , the secondary structure of rCsseverin changed from the presence of Ca2+ shifted to the absence of Ca2+ ( presence of EDTA ) ( Figure 4 ) . With Ca2+ , the secondary structure of rCsseverin contained 23 . 6% α-helix , 56 . 6% β-sheet , and 19 . 8% random loop . While with equivalent EDTA , it changed to 21 . 5% α-helix , 41 . 2% β-sheet , and 37 . 3% random loop . The conformation of the purified rCsseverin was between the two conditions with 24 . 6% α-helix , 49 . 9% β-sheet , 25 . 5% random loop . Ca2+-binding altered the conformation of EF-hand domain from α-helix to β-sheet . The purified rCsseverin partially combined Ca2+ during the processes of expression and purification . We showed that rCsseverin was easily to precipitate when calciumion was added into the solution , and can be resolved by adding EDTA . The binding of rCsseverin to F-actin and its fragments were examined using gel overlay assay as described above . After incubation with rCsseverin , F-actin and its fragments were blotted by anti-rCsseverin serum ( Figure 5A , pane b , lane 1–2 and pane c , lane 1 ) . While incubation with BSA or without rCsseverin ( Figure 5A , pane b , lane 2–3 ) , F-actin couldn't be probed by anti-rCsseverin serum . Whether PLC cells were incubated with rCsseverin before or after fixation and permeabilization , both the green fluorescence ( FITC–conjugated affinipure goat anti-mouse IgG reacted with anti-F-actin monoclonal antibody ) and the red fluorescence ( Cy3–conjugated affinipure goat anti-rat IgG reacted with anti-rCsseverin serum ) were observed . The locations of green fluorescence were mostly coincident with those of the red fluorescence ( Figure 5B , pane a and b ) . There was no red fluorescence or green fluorescence in negative control group ( Figure 5B , pane c and d ) . Thus , we suspected that rCsseverin might enter into PLC cells and bind to actin . To identify the effect of rCsseverin on PLC cells , we tested the total percentage of Annexin V+/PI− and Annexin V+/PI+ cells by flow cytometry . As shown in Figure 6A , incubation of PLC cells with different dosages of rCsseverin ( 10 , 20 , 40 , and 80 µg/ml ) for 48 h after induced spontaneous apoptosis by serum-starved for 24 h decreased the percentage of Annexin V+/PI− and Annexin V+/PI+ cells in a dose-dependent manner ( 30 . 63 , 26 . 98 , 14 . 36 , and 9 . 68% , respectively ) , as compared to the PBS-treated controls , which showed 40 . 74% Annexin V+/PI− and Annexin V+/PI+ cells . The results showed that rCsseverin exhibited potent anti-apoptosis activity on PLC cells in concentration-dependent manner . We also tested the effect of rCsseverin on human normal hepatocyte L-02 cells . No significant decrease of Annexin V+/PI− and Annexin V+/PI+ cells was observed ( Figure 6B ) . We also compared the morphology of PLC cells in the presence of 80 µg/ml rCsseverin to that of PBS-treated cells under the inverted phase-contrast microscopy . Hoechst staining of PBS-treated cells after induced spontaneous apoptosis by serum-starved for 24 h revealed marked morphological changes , such as cell shrinkage , vesicular degeneration , threadlike morphology , nuclear condensation , and nuclear fragmentation , which are typical features of apoptotic cell death . While morphological changes of the PLC cells in presence of 80 µg/ml rCsseverin after treatment with serum-starved for 24 h were not significant ( Figure 6C ) . To further investigate the molecule events triggered by rCsseverin inhibition , we measured MMP in the PLC cells by using flow cytometry and JC-1 staining in situ . The decline of MMP is considered as a symbolic event of early cellular apoptosis . Changes in MMP can be assessed by monitoring JC-1 , which accumulates in mitochondria forming red fluorescent aggregates at high membrane potential and exits mainly in cytosol forming a green fluorescent monomer , presenting a collapse of the membrane [39] . In our study , rCsseverin-treated cells showed reduction of green fluorescence and production of an obvious red fluorescence . The treatment of rCsseverin recovered the MMP in a concentration-dependent manner ( Figure 7 , A and B ) , as indicated by an increase of red ( JC-1 aggregates ) /green ( JC-1 monomers ) ratio . At 48 h , the percentage of 80 µg/ml rCsseverin and PBS treated PLC cells which emitted green fluorescence was 15 . 42 and 9 . 63% , respectively , indicating the recovery of mitochondrial membrane depolarization . The PLC cells that treated with apoptosis introducers exhibited mitochondrial green fluorescence with little red fluorescence , suggesting the cells in depolarization state . The red fluorescence in PLC cells increased , as monitored by in situ JC-1 staining , after the treatment of 10 , 20 , 40 , 80 µg/ml rCsseverin as compared with the PBS group ( Figure 7C ) .
In the present study , we identified that Csseverin , which expressed at life stage of egg , metacercaria and adult worm was a component of CsESPs . We also demonstrated its ability of binding with calciumion and actin filaments . Furthermore , co-incubation of PLC cells with rCsseverin in vitro led to an inhibition of apoptosis induced by serum-starved for 24 h , by using morphological analysis of PLC , detection of the apoptosis-associated change of mitochondrial membrane potential as well as Annexin V/PI apoptosis assay . We inferred that rCsseverin may play an intracellular protective role via preventing apoptotic mitochondrial changes ( the loss of mitochondrial membrane potential ) , just like endogenous human gelsolin did [40] . Gelsolin family is found in a diverse range of organisms including bacteria , invertebrates , plants , primates , rodents and vertebrates . The superfamily in mammals consists of seven different proteins: gelsolin , adseverin , villin , capG , advillin , supervillin and flightless I . All of them contain three or six homologous repeats of a domain named gelsolin-like ( G ) domain [41] . Bioinformatics analysis showed that Csseverin comprised three gelsolin homology domains , calciumion and actin binding motifs . The amino acid sequence of Csseverin shared 36% identity with that of human gelsolin , but there was significant similarity between their tertiary structures . Our phylogenetic analysis suggested that a majority of gelsolin proteins do not form clades based on taxonomic groupings but rather group according to protein functions . The individual gelsolin domains from human gelsolin form distinct clades with homologues from other species , supporting the notion that these proteins have evolved to perform distinct functions in different organisms . Increased Ca2+ influx through voltage-dependent Ca2+ channels is the major determinant of cell injury following excitotoxicity [42] , [43] . The activity of these channels is modulated by dynamic changes in the actin cytoskeleton [44] , [45] , which may occur , in part , through the actions of gelsolin [46] . We obtained soluble and stable rCsseverin . CD measurements actually showed that rCsseverin could bind to calciumion . It has been documented that gelsolin family is of actin-regulatory function [47] . Cytoskeletal actin filaments are dynamic structures that form membranous networks interacting with cell surface receptors and intracellular effectors [48] , [49] . Gel overlay and immunocytochemistry assay indicated the binding activity of rCsseverin . Gelsolin expression in certain tumors correlates with poor prognosis and therapy-resistance . In vitro , human gelsolin has anti-apoptotic and pro-migratory functions and is critical for invasion of some types of tumor cells [50] , [51] , [52] , [53] . We found that gelsolin was highly expressed at tumor borders infiltrating into adjacent liver tissues [54] . In Jurkat lymphoblastoid T-cell line , gelsolin has been shown to inhibit apoptosis , and the overexpression of gelsolin inhibits the loss of mitochondrial membrane potential and cytochrome c release from mitochondria [55] . Additionally , in several models of neuronal cell death , endogenous gelsolin has been demonstrated that has an anti-apoptotic property which correlates to its dynamic actions on the cytoskeleton mediated by inhibition of mitochondrial permeability transition [56] . Here we also showed that rCsseverin could cause obvious apoptotic inhibition in the human HCC cell line . Flow cytometry was used to evaluate rCsseverin-inhibited apoptosis after dual staining of cells with AnnexinV and PI . Due to that Annexin V binding is based on the transposition of phosphatidyl serine from the inner to the outer face of the cell membrane during the early stages of apoptosis [57] . This method has been widely used to discriminate between normal cells ( AnnexinV−/PI− ) , early apoptotic cells AnnexinV+/PI− ) , late apoptotic cells ( AnnexinV+/PI+ ) , and necrotic cells ( AnnexinV−/PI+ ) . Compared with PBS-treated group ( negative control ) , there were less typical apoptotic changes in rCsseverin-treated PLC cells after induced spontaneous apoptosis by serum-starved for 24 h in morphology analysis . We also measured the changes in mitochondrial membrane potential ( MMP ) using a JC-1 probe that gives a red fluorescence when MMP is high and green fluorescence when MMP is low that occurs in early apoptosis cells . We found that interact directly with rCsseverin led to the recovery of mitochondrial membrane potential in PLC cells . Moreover , rCsseverin could be probed by sera from rat infected with C . sinensis besides anti-CsESPs serum that confirmed Csseverin was a molecular of CsESPs . Although it is still unclear about the mechanism of uptake or internalization of CsESPs by host cells , internalized CsESPs could play roles in the interaction between the host and parasite . These data demonstrated that Csseverin , as an anti-apoptotic molecule to carcinoma cell , might be a pathogenic factor in CsESPs , contributing to the development of a pro-tumorigenic environment that was conductive to HCC . Tissue-specific distribution of Csseverin in muscular locations such as teguments of adult worm and metacercaria , as well as its actin binding activity , we inferred that Csseverin might involve in regulating the contraction of smooth muscle and movement of worm body [58] , [59] , [60] . What was more , relative high transcript/protein level of Csseverin at egg stage was consist with its intensive immunolocalization in the intrauterine eggs of adult worm . As a food-borne parasite , C . Sinensis adult lives in the bile ducts of the host and the worm releases a mass of eggs and ESPs , so that Csseverin exists in parasitism circumstance sustainedly and takes a part in the interaction between the host and parasite . Overall , we presented the molecular characteristics of Csseverin , a molecule of CsESPs . Recombinant Csseverin ( rCsseverin ) could bind to Ca2+ and cytoskeletal actin filaments and cause obvious apoptotic inhibition in human HCC cell line . By promoting apoptosis inhibition , Csseverin might exacerbate the process of HCC patients combined with C . sinensis infection . More experiments should be further conducted . The current study may provide a novel insight in understanding the pathogenesis of carcinoma associated with the infection of C . sinensis , which was an inducing factor that cannot be ignored in the process of the development of primary hepatic carcinoma . Since gelsolin has actin-regulatory functions , modulation of the actin network might be responsible for the inhibition of apoptosis , the actin cytoskeleton may be a target to protect from apoptosis [61] . The anti-apoptotic mechanism of Csseverin are worthy of studying in the future . | Clonorchis sinensis ( C . sinensis ) has afflicted more than 35 million people in world and approximately 15 million in China , creating a socio-economic burden in epidemic regions . The infection of C . sinensis is highly related to cholangiocarcinoma and hepatocellular carcinoma ( HCC ) . It has been documented that excretory/secretory products of C . sinensis ( CsESPs ) involved in the pathogenesis of HCC . Csseverin , expressed at life stage of egg , metacercaria and adult worm , was a component of CsESPs . In the current study , we characterized the properties of Csseverin such as sequence signature , actin and calciumion binding activity . In addition , we demonstrated that Csseverin could cause apoptotic inhibition in spontaneously apoptotic human HCC cell line PLC cells by using morphological analysis , detection of the apoptosis-associated change of mitochondrial membrane potential ( MMP ) as well as Annexin V/PI apoptosis assay . Our study provided an exploratory sight view of mechanism involved in progress of carcinoma associated with the infection of C . sinensis and Csseverin might exacerbate the process of C . sinensis infected HCC patients . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2013 | Molecular Characterization of Severin from Clonorchis sinensis Excretory/Secretory Products and Its Potential Anti-apoptotic Role in Hepatocarcinoma PLC Cells |
Despite the wealth of knowledge regarding the mechanisms of action and the mechanisms of resistance to azole antifungals , very little is known about how the azoles are imported into pathogenic fungal cells . Here the in-vitro accumulation and import of Fluconazole ( FLC ) was examined in the pathogenic fungus , Candida albicans . In energized cells , FLC accumulation correlates inversely with expression of ATP-dependent efflux pumps . In de-energized cells , all strains accumulate FLC , suggesting that FLC import is not ATP-dependent . The kinetics of import in de-energized cells displays saturation kinetics with a Km of 0 . 64 uM and Vmax of 0 . 0056 pmol/min/108 cells , demonstrating that FLC import proceeds via facilitated diffusion through a transporter rather than passive diffusion . Other azoles inhibit FLC import on a mole/mole basis , suggesting that all azoles utilize the same facilitated diffusion mechanism . An analysis of related compounds indicates that competition for azole import depends on an aromatic ring and an imidazole or triazole ring together in one molecule . Import of FLC by facilitated diffusion is observed in other fungi , including Cryptococcus neoformans , Saccharomyces cerevisiae , and Candida krusei , indicating that the mechanism of transport is conserved among fungal species . FLC import was shown to vary among Candida albicans resistant clinical isolates , suggesting that altered facilitated diffusion may be a previously uncharacterized mechanism of resistance to azole drugs .
The incidence of invasive fungal disease has increased over 200% in the US in the last 25 years [1] , likely the result of a parallel increase in the immunocompromised patient population . Candida species are the most common invasive fungal pathogens , with Candida albicans accounting for more than 50% of all infections [2] . C . albicans causes oral , vaginal and systemic disease , with the highest morbidity rate ( 30%–50% ) occurring with systemic Candida infections in neutropenic transplant patients [3]–[5] . One of the first lines of defense for treating pathogenic fungal infections are the azole drugs , including FLC , the most commonly used azole . FLC and other azoles affect the biosynthesis of ergosterol ( the major sterol in the fungal plasma membrane ) by inhibiting 14α lanosterol demethylase , the product of the ERG11 gene . The significant increase in invasive fungal infections and the prolonged and repeated treatment of AIDS patients has resulted in a marked increase in the emergence of FLC-resistant C . albicans isolates [6]–[8] . In C . albicans , several mechanisms of resistance have been well characterized ( reviewed in [6]–[8] ) . However , clinical isolates have not been investigated for altered azole import as a mechanism of resistance . Several studies have investigated the accumulation of drugs in resistant clinical isolates of C . albicans [7] , [9]–[12] . These studies show reduced intracellular FLC in the isolates , which is energy dependent and the result of overexpression of the major facilitator pump gene MDR1 , and the ABC transporter efflux pump genes CDR1 and CDR2 [7] , [8] . Both Cdr1p and Cdr2p are ATP-dependent efflux pumps , whereas Mdr1p utilizes proton motive force at the membrane to transport drugs and other compounds . Surprisingly , the mechanism ( s ) by which FLC enters the C . albicans cell remain unstudied . Defects in drug import are a common mechanism of drug resistance in other pathogenic organisms , but to date , there have been no reports that C . albicans utilizes altered import as a resistance mechanism . Azoles are widely assumed to enter the fungal cell via passive diffusion [13]–[15] , but there is little evidence to support this . Some evidence for facilitated diffusion of azoles has been reported , but these experiments were performed in energized cells in which drug efflux was active , and therefore failed to uncouple import and export [9] , [16] . This study biochemically characterized the mechanism by which FLC is taken into C . albicans cells . The results suggest that the FLC enters the cell by energy-independent facilitated diffusion in C . albicans and other pathogenic fungi . In addition , import levels vary among resistant clinical isolates , suggesting that import is a previously uncharacterized mechanism of resistance to azole drugs in C . albicans .
The accumulation of [3H]-FLC was analyzed in energized cells in the presence of glucose ( Fig . 1A ) , in which energy-driven importers and efflux pumps would have an effect on drug accumulation . Four C . albicans strains were tested: a wild type strain ( SC5314 ) , a susceptible clinical isolate ( #1 ) , a matched resistant clinical isolate that overexpresses CDR1 , CDR2 and MDR1 ( #17 , ref [17] , [18] ) and a genetic construct strain DSY1050 that is deleted for CDR1 , CDR2 and MDR1 [19] . Cells were incubated with [3H]-FLC and uptake was quenched by rapid dilution and filtration as described in Experimental Procedures . FLC accumulation was observed to be linear over 3 h , with maximum accumulation observed after 24 h ( Fig . 1 ) , as accumulation did not increase after 24 h ( data not shown ) . In the presence of glucose , the expression of efflux pumps does have a minor effect on [3H]-FLC accumulation: strain #17 , which overexpresses CDR1 , CDR2 and MDR1 , shows the lowest accumulation , and strain DSY1040 , which is deleted for the three pumps , shows the highest level of accumulation . Since FLC accumulation in the presence of glucose will be the sum of both FLC import and export with ATP-dependent efflux pumps , the strains were tested after glucose starvation for 2 h ( Fig . 1B ) . If import is energy-independent , the depletion of ATP should inactivate ATP-dependent efflux pumps and accumulation should largely be due to import . Under these conditions , there is less variation between strains in the accumulation of [3H]-FLC and accumulation still occurred at a similar rate ( Fig . 1B ) . The addition of the glycolysis inhibitor 2-deoxy-D-glucose during the 2 h preincubation starvation did not alter [3H]-FLC accumulation levels , suggesting that the cells are indeed de-energized in the absence of glucose ( data not shown ) . This suggests that accumulation does occur in de-energized cells and that efflux pumps do not have an effect on FLC accumulation in the absence of glucose . Cells were inactivated to determine if import depends on living cells . [3H]-FLC did not accumulate when SC5314 cells were killed by heat ( 70°C for 45 m; Fig . 1 and Table 1 , row 1 ) or methanol treatment ( 95% methanol for 45 m , Table 1 , row 1 ) . The heat-killed and methanol-killed cells appeared intact when observed under light microscopy . However , live cells were impermeable to propidium iodide ( <1 . 0% of cells were stained ) while heat-killed or methanol-killed cells were permeable to propidium iodide ( 98 . 8% and 99 . 5% cells stained respectively ) , suggesting that the lack of accumulation in killed cells is due to permeable cell walls and/or membranes . Fig . 1 showed that all strains accumulate FLC in the absence of glucose . A variety of conditions that might have an affect on azole import were assayed for altered uptake of [3H]-FLC ( Table 1 ) . [3H]-FLC accumulation is temperature dependent , with minimal accumulation at 4°C and maximum accumulation at 30°C , with a slight reduction at 37°C ( Table 1 , row 2 ) . This is inconsistent with passive diffusion of FLC , which should increase with an increase in temperature . Import per cell was also studied during various growth phases ( Table 1 , row 3 ) with maximum import occurring in cells growing in mid-log phase ( OD600 0 . 4 ) . Germ tubes ( hyphae ) exhibited a 4 fold increase in [3H]-FLC accumulation when compared to yeast cells ( Table 1 , row 4 ) , although cell numbers are difficult to determine for hyphal cells . Accumulation appears to be unaffected by changes in pH ( Table 1 , row 5 ) . Surprisingly , the highest change in accumulation occurs under micro-aerophilic conditions , where cells accumulate over 4 fold more than cells growth with normal aeration ( Table 1 , row 6 ) . The changes in accumulation associated with growth and oxygen levels , and the reduced import at higher temperatures confirms that accumulation is not simply passive diffusion . The high level accumulation of [3H]-FLC in de-energized cells ( Fig . 1B ) and the changes in accumulation with changes in the environment ( Table 1 ) suggest FLC import is carried out by facilitated diffusion . Therefore , the kinetics of FLC import were assayed in detail for de-energized SC5314 cells , studying the initial rate of import across a spectrum of FLC concentrations ( Fig . 2 ) . FLC import displayed saturation kinetics with a Km of 0 . 64 uM and Vmax of 0 . 0056 pmol/min/108 cells . The saturation kinetics strongly support facilitated diffusion through a specific transporter as the mechanism of FLC import . To determine whether other azole drugs utilize the same transporter as FLC , import of [3H]-FLC was assayed in de-energized SC5314 cells in the presence of unlabeled azole drugs including FLC , ketoconazole ( KTC ) , voriconazole ( VRC ) , and itraconazole ( ITC ) . As seen in Fig . 3A , all four azoles inhibited the import of [3H]-FLC at 10 fold and 100 fold excess , suggesting that the same transporter is involved in the uptake of all four azole drugs . 50% inhibitory concentration ( IC50 ) values were calculated for KTC and ITC by measuring the rate of [3H]-FLC import in the presence of excess unlabeled KTC or ITC . The calculated IC50 values are 65 nM and 48 nM for KTC and ITC respectively ( Fig . 3B and data not shown ) . As 50 nM of [3H]-FLC is used in the assay , the IC50 values suggest that FLC , KTC and ITC compete at equimolar concentrations for import into the cell , suggesting that they compete equivalently as substrates for the transporter ( s ) . The antifungal 5-flucytosine ( 5FC ) and the fluorescent dye rhodamine 6G ( R6G ) have been hypothesized to share the same import mechanism as FLC [10] , [20] . 5FC is a nucleoside analog [20] and R6G is a dye known to be effluxed by Cdr1p and Cdr2p [10] . However , neither 5FC ( 5 µM ) nor R6G ( 5 µM ) at 100 fold molar excess reduced the import of [3H]-FLC ( Fig . 3A ) . These observations suggest that 5FC and R6G have independent import mechanisms , and that the assay does not measure drug export , since R6G acts as a substrate for the efflux pumps . As FLC is unlikely to be the natural substrate for the import mechanism , additional compounds were tested as competitors for FLC import . The following types of compounds were tested in molar excess and were shown not to compete with FLC: sugars , nucleosides , amino acids , unrelated antifungals , salts and unrelated drugs ( complete list of compounds is in Table 2 ) . It is known that uptake of ergosterol increases under low-oxygen conditions in Saccharomyces cerevisiae [21] . FLC import also increases under low-oxygen conditions ( Table 1 , row 6 ) . These observations suggested that FLC and ergosterol might share an import mechanism . However , 10 and 100 fold excess of unlabeled ergosterol did not compete for [3H]-FLC import in this assay ( data not shown ) . In analyzing the competition data with FLC , ITC , KTC and VRC ( Fig . 3A ) , it became evident that all of the azoles share two structural moieties in common – a ) a halogenated aromatic 6 carbon ring and b ) an imidazole [2 nitrogen ( N ) ] or triazole ( 3 N ) 5-member ring ( Table 3 , structures shown in Supplemental Figure S1 based on the compound structures in the NIH PubChem Compound Database [22] ) . Using this as a starting point , we have expanded our understanding of the structural components that are required to compete for FLC import ( Table 3 ) . Rows 1 to 8 show the molecules that compete for FLC import , including the clinically important FLC , VRC , ITC , KTC , and posaconazole , as well as a fluorescein labeled ITC , and two agricultural azoles , paclobutrazol and azaconazole . All of these compounds contain a 6-carbon ring halogenated with fluorine ( F ) or chlorine ( Cl ) and an imidazole ( 2N ) or triazole ( 3N ) 5-member ring . Importantly , fluorescein labeled ITC competes for FLC import , but it has no antifungal activity on its own ( data not shown ) . Many of these compounds contain additional ring structures within the molecule ( Table 3 and Supplemental Figure S1 ) . Four other drugs were tested because they are active against the sterol biosynthetic pathway and have antimicrobial activities , including Tipifarnib , the Tipifarnib derivative 2g [23] , STN54 [24] , and Benznidazole . None of these molecules compete ( Table 3 , rows 9–12 ) . All four have a modified imidazole rings , and two have aromatic rings without halogenation , suggesting that one or both of these two structures are important for competition of import . To analyze these two components separately , the compounds difluoro-benzene and imidazole were tested in the import assay independently and together ( Table 3 rows 13–15 ) . The compounds individually or in combination did not compete , suggesting that a compound must have both moieties physically linked to compete for FLC import . Finally , the imidazole ring resembles histidine ( H ) and the aromatic ring resembles phenylalanine ( F ) . These two amino acids alone do not compete for FLC import ( Table 3 , rows 16–17 ) . To test if physical linkage of H and F would compete for FLC import , oligopeptides were prepared including the dipeptides HF and FH , and larger peptides separated by one , two or three alanines ( A ) ( Table 2 ) . None of these oligopeptides competed for FLC import . To address concerns about peptide degradation , the peptides were tested for competition of FLC import in the presence of protease inhibitors , and at vast molar excess ( 10 , 000X ) with no appreciable competition . FLC and other azoles are used in the treatment of many pathogenic fungi . De-energized Cryptococcus neoformans , S . cerevisiae and C . krusei , which is an intrinsically azole resistant Candida species , were all tested for import of [3H]-FLC and import levels were compared to [3H]-FLC import of C . albicans ( SC5314 ) . As shown in Table 4 , each species was able to import and accumulate [3H]-FLC at levels similar to C . albicans . Similar experiments with equivalent OD units of E . coli cells did not detect FLC import above background ( data not shown ) , supporting the conclusion that FLC import is not by passive diffusion . Collectively , these data suggest that azoles are imported by a specific transporter that may be conserved across fungal species . S . cerevisiae shows FLC import similar to C . albicans ( Table 4 ) . To ensure that known pumps are not involved in import , two S . cerevisiae strains , AD1-8 and AD1-9 [25] , [26] , which are deleted for eight and nine efflux pumps respectively , were tested for import . Import was not significantly different from the wild type strain ( data not shown ) . A collection of over 5 , 000 Saccharomyces strains containing deletions in non-essential genes is available to the research community [27] . The collection was screened biochemically , using a 96 well format ( see Materials and Methods ) . Two gene deletions from the collection were identified that had significantly reduced fluconazole import ( Fig . 4 ) . SNF7 and DOA4 are involved in protein transport at the plasma membrane ( reviewed in [28] , [29] ) . SNF7 is a member of the ESCRT III complex ( Endosomal Sorting Complex Required for Transport ) that is involved in recycling or degrading membrane proteins . DOA4 is a de-ubiquitination enzyme involved in the same process and physically interacts with SNF7 . As these proteins are both cytoplasmic and are involved in surface protein processing , they are not likely to be directly involved in import , but might be involved in subsequent processing of the import protein . Other ESCRT proteins in the screen did not have a significant loss of FLC import ( data not shown ) . One explanation for the role of SNF7 and DOA4 in FLC import is that they may interfere with degradation of efflux pumps , resulting in increased efflux . To test for efflux pump activity in snf7 and doa4 strains , R6G efflux was monitored over time ( Fig . 5 ) . Mutant strains of snf7 and doa4 effluxed R6G at similar rates to the wild type cells , while heat killed cells did not efflux R6G . The similar rates of efflux for wild type and snf7 and doa4 mutants indicate that reduced FLC import in these mutants ( Fig . 4 ) is not the result of increased efflux , even through the assay was preformed with deenergized cells . In C . albicans , the SNF7 and DOA4 mutants did not show altered FLC import when compared to wild type strains ( data not shown ) . This suggests that the role of SNF7 and DOA4 is not conserved between the two species . It is not unusual for genes to have different functions between the two species ( i . e . [30]–[33] ) . It also indicates that the role of SNF7 and DOA4 in S . cerevisiae is not central to the mechanism of FLC import . To date , changes in azole import have not been reported as a mechanism of antifungal resistance in C . albicans or other pathogenic fungi . However , it is plausible that a mutation in the azole transporter would lead to azole resistance . A collection of unmatched clinical isolates of C . albicans was tested for their ability to import [3H]-FLC in the absence of glucose . Approximately 50% of the resistant isolates in this collection have unknown mechanisms of resistance [34] . As seen in Fig . 6 , of the 35 isolates tested , four exhibited statistically significant alterations in [3H]-FLC import when compared to the median import value for all 35 isolates . Three of the isolates revealed significant decreases in import , while one had a significant increase . In addition , there is considerable variation in import between the other clinical isolates , both above and below the mean . Of the isolates exhibiting significantly decreased import , all exhibited other known but diverse mechanisms of resistance , including mutations in ERG11 and overexpression of CDR1 , CDR2 ( both of which are inactive in the import assay ) or MDR1 . The isolate with significantly increased import had no known mechanism of resistance . It is possible that these strains have altered import as well as other mechanisms of azole resistance , as it is not uncommon for clinical isolates to have multiple mechanisms of resistance [7] , [34] .
The biochemical analyses clearly demonstrate that FLC import into C . albicans is not the result of passive diffusion but the result of facilitated diffusion . The biochemical evidence includes the following: There is strong biochemical evidence that FLC is not simply binding to the cells . First , the heat-killed and methanol-killed cells do not show FLC accumulation ( Fig . 1 and Table 1 ) . If the drug were binding to the cell wall , then the inactivated cells should also bind to drug . In fact , these treatments might expose more cell wall , resulting in more drug binding . However , FLC accumulation is not observed in these cells . Second , the cell wall component beta 1 , 3 glucan , available commercially as laminarin , does not compete for FLC binding ( Table 2 ) . It has been shown previously that cells in biofilms bind to FLC and this binding can be competed with laminarin [36] . However , laminarin had no effect on the FLC import assay used in this study ( Table 2 ) , confirming that the import assay and the biofilm assay are measuring separate phenomena . There is strong evidence that the results of these analyses are not the result of the efflux pumps . In the assay , FLC import kinetics were studied under de-energizing conditions ( Fig . 2 ) , in which the efflux pumps are not active , and the results were not altered by the addition of 2 deoxy-glucose ( data not shown ) . In addition , the pump mutant DSY1050 in which MDR1 , CDR1 and CDR2 are all deleted shows the same FLC import as wild type strains ( Fig . 1 ) . Saccharomyces strains deleted for 8 to 9 efflux pumps that are known to be associated with FLC efflux are still able to accumulate FLC ( data not shown ) . Finally , inhibitors of mitochondrial function , including DNP and CCCP , that would eliminate function of MDR1 , as well as CDR1 and CDR2 , had no effect on import ( Table 2 ) . Azoles , including the clinically important FLC , KTC , ITC , VRC and POS , as well as the agriculturally important paclobutrazol and azaconazole , compete with labeled FLC for import ( Fig . 3A , Table 3 ) . The imidazole KTC and the triazole ITC compete at approximately equal molar concentrations ( Fig . 3B and data not shown ) . This supports the conclusion that both imidazoles and triazoles utilize the same import mechanism as FLC . Unrelated antifungals , including terbinafine , fenpropimorph and amphotericin B , do not compete and thus are unlikely to use the same transporter . By testing related drugs , the structural moieties within a compound that allow it to compete with FLC for import were defined ( Table 3 ) . Compounds that compete contain both a halogenated ( F or Cl ) aromatic 6 member ring , and a triazole or imidazole 5 member ring . Compounds that do not compete have a methylation or other modification of the 5 member ring , potentially coupled to non-halogenated aromatic rings . The 5 member triazole or imidazole ring is necessary for import , as compounds containing aromatic rings alone do not complete: . Similarly , neither His nor imidazole compete , consistent with the fact that neither contains an aromatic ring . The 6 member ring and the 5 member ring structures must be contained on the same molecule , as a mixture of difluoro-benzene and imidazole does not compete . Given that His and Phe do not compete separately , it was of interest to test the two amino acids together . Oligopeptides containing His and Phe separated by 0 to 4 Ala did not compete for FLC import ( Table 2 ) suggesting that linked His and Phe , which contain an imidazole ring and an aromatic ring , are not sufficient for recognition of the FLC import mechanism . Both 5FC and R6G have been suggested to be co-transported with FLC . In a recent study , Noel et al . [20] characterized a series of Candida lusitaniae isolates that were cross-resistant to 5FC and FLC . The isolates tested were resistant to 5FC and susceptible to FLC unless the compounds were used simultaneously , in which case cross-resistance was observed . Noel et al . hypothesized that 5FC and FLC shared a common transporter and that extracellular 5FC was acting as a competitive inhibitor of FLC uptake transport [20] . Later reports indicated that the cross resistance was in fact due to mutations in genes encoding for cysteine permease [37] and cytosine deaminase [38] indicating that cross resistance is not due to a shared import mechanism . Data in this study support this conclusion as 5FC does not compete for [3H]-FLC import ( Fig . 3 ) . R6G is a dye known to be effluxed from the cell by the pumps Cdr1p and Cdr2p . R6G has been shown to be capable of competing for FLC efflux [10] . It has therefore been hypothesized that FLC is also imported by Cdr1p or Cdr2p acting in reverse and that R6G could possibly compete for FLC import as well as export [10] . Data from this study clearly indicate that R6G does not compete for FLC import ( Fig . 3 ) . Azoles are used to treat a wide variety of human ( Candida , Cryptococcus , Aspergillus ) and agricultural ( Pichia , Rhodoterula , Saccharomyces ) fungal pathogens [39] . Based on this widespread use , it was of interest to determine if other fungal species import and accumulate [3H]-FLC . It was shown that C . neoformans , S . cerevisiae and C . krusei import [3H]-FLC with similar kinetics ( data not shown ) and to final levels similar to C . albicans ( Table 4 ) . In addition the agricultural triazole paclobutrazol [40] , [41] and the agricultural imidazole azaconazole [42] , [43] compete with FLC for import ( Table 3 ) . These data suggest that the putative azole transporter is conserved across various fungal species . Interestingly , a recent study by Muller et al . [39] has shown that fungi found in an agricultural environment ( including , but not limited to , various Candida , Cryptococcus and Saccharomyces species ) are routinely treated with fluquinconazole , penconazole , tebuconazole or triadimenol . A significant portion of the isolates from different species that are resistant to these agricultural azoles were cross resistant to medical azoles including FLC , ITC , KTC or VRC . The cross-resistance could be the result of over-expression of the efflux pumps , but it is also possible that these isolates contain an alteration in an azole importer that is conserved across various fungal species and confers cross resistance to all azoles since all clinically significant azoles compete for FLC import ( Table 3 and Fig . 3 ) . The biochemical screen for an importer in the haploid gene deletion strain collection failed to identify a potential import protein . The lack of a clear candidate suggests that the import protein is either an essential gene , which can not be deleted and would not be represented in the strain collection , or is present in more than one version - two paralogs with the same function or a gene family . In that case , deletion of one of the genes would not eliminate FLC import . It is possible that with two paralogs , import would be reduced 50% but that was not observed in the screen , despite the use of two different time points . If the gene family members or paralogs had substantially different kinetics , that would have been detected in the kinetic analysis . However , if the multiple copies of the importer all behave similarly , and the wild type strain has all of the genes expressed , then the differences would not be detected in the kinetic analysis . Further kinetic analysis awaits the identification of the import protein . The biochemical screen did identify SNF7 and DOA4 , two components of the ESCRT III complex involved in recycling and degradation of surface proteins through the endosomes and multi-vesicular bodies ( MVB ) . SNF7 encodes one of the four subunits of the ESCRT III complexes , and DOA4 encodes ubiquitin isopeptidase that is closely associated with the complex . However , the C . albicans snf7 and doa4 mutants did not exhibit a reduction in import , suggesting that the role of SNF7 and DOA4 in S . cerevisiae is not central to the import mechanism . The two gene deletion strains , snf7 and doa4 , do not have an altered MIC to FLC . However , the efflux pump PDR5 is highly active in Saccharomyces wild type strains , and may mask any effect of snf7 and doa4 on FLC MIC . It is curious that other ESCRT proteins do not have an altered FLC import . Further work awaits the identification of the FLC import protein . To date , there has been no report of altered FLC import as a mechanism of antifungal resistance . The data in this study suggests that all azoles utilize the same import mechanism mediated by a transporter . Therefore , it is possible that a mutation in the putative transporter would lead to azole cross-resistance . 35 unmatched clinical isolates in which known mechanisms of resistance had been documented [34] were evaluated for FLC import ( Fig . 6 ) . Four of the 35 isolates exhibit significantly altered [3H]-FLC import . One of these four has no known mechanism ( s ) of resistance , while the other three are known to overexpress ERG11 , MDR1 , CDR1 and/or CDR2 or contain a mutation in ERG11 . However , it is common for clinical isolates to exhibit multiple mechanisms of resistance [7] , [8] , [34] . Therefore , it is likely that these isolates have mutations that affect [3H]-FLC import , in addition to other mechanisms of resistance . These data suggest that loss , reduction , or alteration of azole import may be a previously unknown mechanism of antifungal resistance . In conclusion , this study uses biochemical analyses to demonstrate that FLC import is not via passive diffusion but is in fact via facilitated diffusion . The data presented here represents the first comprehensive analysis of FLC import in C . albicans . This work also demonstrates that azoles share a common transport mechanism and azole import is conserved between several pathogenic fungi . Future directions will be focused on identifying and characterizing the protein responsible for this newly identified FLC facilitated diffusion .
Candida albicans SC5314 ( from W . Fonzi; [44] ) is the wild type lab strain used in this study . C . albicans isolates #1 ( 2–76 ) and #17 ( 12–99 ) are a matched set of FLC susceptible and resistant clinical isolates in which #17 overexpresses ERG11 , CDR1 , CDR2 and MDR1 . #1 and #17 are from a series of 17 oral isolates from a single HIV positive patient [45] . C . albicans DSY1050 ( from D . Sanglard [19] ) is a FLC hyper-susceptible strain containing homozygous deletions of CDR1 , CDR2 and MDR1 . Cryptococcus neoformans H99 ( from J . Lodge; [46] ) , Candida krusei ( our collection; [47] ) and Saccharomyces cerevisiae W303 ( our collection; [48] ) were all used to determine if other fungal species are capable of importing [3H]-FLC . A collection of 35 un-matched clinical isolates ( from D . Stevens; [34] were used to determine import of [3H]-FLC in isolates with known and unknown mechanisms of resistance . The Saccharomyces cerevisiae haploid gene-deletion library was originally obtained from Research Genetics ( Huntsville , AL ) . Strains were maintained on YEPD ( 10 g of yeast extract , g of peptone , 20 g of dextrose , with or without 15 g of Bacto Agar per liter ) , or on CSM complete medium ( 0 . 75 g CSM [Bio 101; Vista , CA] , 1 . 7 g yeast nitrogen base without amino acids or ammonium sulfate , 5 g ammonium sulfate , 20 g dextrose per liter ) . All isolates were stored at −80°C in CSM or YEPD containing 10% glycerol . Overnight cultures were inoculated from a single colony on a YEPD agar plate and inoculated into YEPD broth and grown overnight at 30°C , 180 rpm , unless otherwise noted . Medium components were obtained from Fischer Scientific ( Pittsburgh , PA ) or Bio 101 ( Vista , CA ) . General chemicals were obtained from Fisher Scientific , or Sigma-Aldrich ( St . Louis , MO ) . Itraconazole , voriconazole , ketoconazole , paclobutrazol , azaconazole , flucytosine and R6G were obtained from Sigma-Aldrich ( St . Louis , MO ) . FLC was a generous donation from Pfizer , New York , NY . Fluorescein-labeled ITC was the generous gift of J . Lui and collaborators ( Johns Hopkins U ) . POS , benznidazole , Tipifarnib , Tipifarnib 2g , and STN54 were the generous gifts of Fred Buckner , ( U Washington ) . Oligopeptides were obtained from Neo BioScience ( www . neobiosci . com ) . FLC uptake was determined using [3H]-FLC ( specific activity 740 GBa/mmol , 20 Ci/mmol , 2×104 CPM/pmol , 1 uCi/uL; 50 uM FLC; custom synthesis by Amersham Biosciences , UK ) . Cells were grown overnight in CSM complete medium at 30°C to a density typically between OD600 6 . 0 to 8 . 0 , unless otherwise noted . Cells were subsequently harvested by centrifugation ( 3000×g , 5 m ) and washed three times with YNB complete ( 1 . 7 g yeast nitrogen base without amino acids or ammonium sulfate , 5 g ammonium sulfate per liter , pH 5 . 0 ) without glucose ( for starvation ) and without supplementation , unless otherwise noted . Cells were resuspended at an OD600 of 75 in YNB for 2–3 h for glucose starvation . Reaction mixes consisted of 250 uL of YNB , 200 uL of cells ( 75 OD ) and 50 uL of [3H]-FLC ( 1/100 dilution of stock ) . The resulting [3H]-FLC concentration is 50 nM ( 0 . 015 ug/ml ) , which is significantly below the MIC for all strains . Samples ( 100 ul ) were removed at various time points and placed into 5 ml stop solution ( YNB +20 µM [6 ug/ml] FLC ) , filtered on glass fibre filters ( 24 mm GF/C; Whatman; Kent , UK ) pre-wetted with stop solution and washed with 5 ml of stop solution . Filters were transferred to 20 ml scintillation vials . Scintillation cocktail ( Ecoscint XR , National Diagnostics , Atlanta GA ) was added ( 15 ml ) and the radioactivity associated with the filter was measured with a liquid scintillation analyzer ( Tri-Carb 2800 TR; Perkin Elmer; Waltham , MA ) and normalized to CPM/1×108 cells . Rate of [3H]-FLC uptake was determined by incubating samples in the presence of increasing concentrations of unlabeled FLC ( unless otherwise noted ) and samples were analyzed for [3H]-FLC accumulation at designated time points . These data were analyzed using linear regression to determine the rate of [3H]-FLC uptake . GraphPad Prism 4 . 0 was used to determine linear regression , Michaelis-Menten import kinetics ( Vmax and Km ) and 50% inhibitory concentration ( IC50 ) values . Uptake of [3H]-FLC by a collection of 35 Candida clinical isolates , and by the Saccharomyces gene deletion library was determined by following the above protocol with the exception that cells ( 1 . 5 ml ) were grown for 48 h in 2 . 0 ml 96-deep well plates ( Masterblock; Greiner bio-one; Monroe , NC ) . The reactions were half of the size: 125 uL of YNB , 100 uL of cells , and 25 uL of [3H]-FLC ( 1/100 dilution of stock ) . Samples ( 100 ul ) from the reaction were removed 24 h post incubation and filtered over 96-well multiscreen HTS filter plates ( Opaque non-sterile with lid , 1 . 2 µm glass fibre type C filter; Millipore , Billerica , MA ) , wells were dried and the bottoms were sealed with sealing tape ( Perkin Elmer; Waltham , MA ) . Scintillation fluid ( 150 ul ) was added to each well , the tops were sealed with sealing tape and the plates were counted on a 96-well liquid scintillation counter using both top and bottom counting ( Liquid Scintillation and Luminescence Counter , WALLAC/Jet , 1450 Microbeta ) . Isolates were grown to exponential phase in 5 ml YEPD at 30°C with 180 RPM shaking . Cells were collected by centrifugation ( 3000×g 5 m ) and washed three times in sterile water . Cells were resuspended to OD600 = 0 . 4 in 50 mM phosphate buffer pH 6 . 0 with 5 mM 2-deoxy-D-glucose . Cells were incubated for 60 m at 30°C with shaking . R6G was added to a final concentration of 10 µm and cells were incubated for 90 minutes at 30° with shaking . Cells were collected by centrifugation and washed twice in 50 mM phosphate buffer pH 6 . 0 . Cells were resuspended to OD600 = 0 . 2 in 50 mM phosphate buffer pH 6 . 0 . 500 µl of cells was diluted 1∶2 in 50 mM phosphate buffer pH 6 . 0 , and accumulation was measured by fluorescence activated cell sorting ( FACS ) . Glucose was added to a final concentration of 40 µm and efflux was measured by analyzing an aliquot of cells diluted 1∶10 in ice-cold 50 mM phosphate buffer pH 6 . 0 by FACS at the time intervals indicated using a Beckman Coulter EpicsXL-MCL 4-colour cell analyzer . The geometric mean of the fluorescence of each sample was calculated using FlowJo software . Further studies were done to determine the effect of changes in the incubation conditions have on FLC import: | Azole antifungals are used to treat a wide variety of fungal infections of humans , animals and plants . A great deal is known about how the azoles interact with their target enzyme within fungal cells and how the azoles are exported from the fungal cell through various efflux pumps . Altered interactions with the target enzyme and altered efflux pump expression are common mechanisms of azole resistance in fungi . However , the mechanism by which azoles enter a fungal cell is not clear—many have assumed that azoles passively diffuse into the cell . This study demonstrates that azoles are not passively diffused , or actively pumped , into the cell . Instead , azoles are imported by facilitated diffusion , mediated by a transporter . Facilitated diffusion of azoles is saturable . All clinically important azoles , and many structurally related compounds , compete for FLC import , while structurally unrelated drugs do not compete . Azole import by facilitated diffusion is shown in four species of fungi , suggesting that it is common for most if not all fungi . Altered facilitated diffusion is observed in a collection of clinical isolates , suggesting that altered import is a previously uncharacterized mechanism of resistance . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"infectious",
"diseases/fungal",
"infections",
"microbiology/medical",
"microbiology",
"infectious",
"diseases/antimicrobials",
"and",
"drug",
"resistance"
] | 2010 | Azole Drugs Are Imported By Facilitated Diffusion in Candida albicans and Other Pathogenic Fungi |
Fluoroquinolones are antibacterial drugs that inhibit DNA Gyrase and Topoisomerase IV . These essential enzymes facilitate chromosome replication and RNA transcription by regulating chromosome supercoiling . High-level resistance to fluoroquinolones in E . coli requires the accumulation of multiple mutations , including those that alter target genes and genes regulating drug efflux . Previous studies have shown some drug-resistance mutations reduce bacterial fitness , leading to the selection of fitness-compensatory mutations . The impact of fluoroquinolone-resistance on bacterial fitness was analyzed in constructed isogenic strains carrying up to 5 resistance mutations . Some mutations significantly decreased bacterial fitness both in vitro and in vivo . We identified low-fitness triple-mutants where the acquisition of a fourth resistance mutation significantly increased fitness in vitro and in vivo while at the same time dramatically decreasing drug susceptibility . The largest effect occurred with the addition of a parC mutation ( Topoisomerase IV ) to a low-fitness strain carrying resistance mutations in gyrA ( DNA Gyrase ) and marR ( drug efflux regulation ) . Increased fitness was accompanied by a significant change in the level of gyrA promoter activity as measured in an assay of DNA supercoiling . In selection and competition experiments made in the absence of drug , parC mutants that improved fitness and reduced susceptibility were selected . These data suggest that natural selection for improved growth in bacteria with low-level resistance to fluoroquinolones could in some cases select for further reductions in drug susceptibility . Thus , increased resistance to fluoroquinolones could be selected even in the absence of further exposure to the drug .
Fluoroquinolones are potent antibacterial drugs [1] that bind to bacterial type II topoisomerases ( DNA gyrase and topoisomerase IV ) when they are in complex with DNA . The drugs inhibit chromosome re-ligation after enzyme-mediated cleavage [2] . Fluoroquinolones are effective against many bacteria including invasive E . coli , but resistance is increasing with 28 of 29 countries in Europe reporting a significant rise between 2001–2007 [3] . The rapid increase is surprising because clinically relevant levels of resistance in E . coli require multiple genetic changes , including mutations altering topoisomerases and up-regulating drug efflux [4] , changes that are associated with reduced bacterial fitness in vitro and in vivo [5]–[7] . Inappropriate use of fluoroquinolones , or co-selection of resistant bacteria with the use of other antimicrobial drugs [8] may be factors driving the increase in resistance but these may not be the sole causes [9] . For example , resistant isolates of E . coli increased from ∼7–19% between 2001 and 2007 in the UK [3] , while outpatient fluoroquinolone use remained unchanged from 1997–2003 [9] . To develop an effective strategy to restrict the increase in resistance frequency will require that we have a full understanding of the factors driving the increase . The aim of this paper was to investigate whether selection for improved fitness in bacteria might itself be a factor promoting increased resistance . The fitness costs of drug resistance can be reduced by selection of low-cost mutations or by the accumulation of secondary fitness-compensating mutations that do not reduce resistance . During experimental evolution of clinical isolates of E . coli for decreased susceptibility to fluoroquinolones most lineages ( 16/18 ) suffered reduced growth competitiveness after only two or three selection steps [5] . However , lineages selected for further decreases in susceptibility were occasionally associated with a relative restoration of fitness [5] . A similar reversal was noted in some constructed strains of Streptococcus pneumoniae carrying one or two resistance mutations [10] . These data suggested that some resistance mutations might be selected because they decrease susceptibility to the drug and simultaneously reduce the fitness costs associated with existing resistance mutations . No cause for the phenomenon has been demonstrated , and it's relevance to bacterial fitness in vivo in not clear . To examine the phenomenon we constructed isogenic strains carrying various combinations of five resistance mutations found commonly in fluoroquinolone-resistant clinical E . coli , and measured their drug-susceptibility and fitness . The relationships between the number of resistance mutations and bacterial fitness were complex and the addition of a resistance mutation was shown in some cases to improve bacterial fitness . These findings have implications for the evolution of fluoroquinolone resistance in the absence of antibiotic exposure .
The E . coli urinary tract infection isolate C1186 [4] is highly-resistant to fluoroquinolones ( MIC for ciprofloxacin ≥32 µg/ml ) and carries resistance mutations altering topoisomerases ( gyrA Ser83→Leu , Asp87→Asn; parC Ser80→Ile ) , and up-regulating drug efflux ( marOR small deletion , and amino acid substitution; acrR IS1 insertion ) . These 5 mutations are typical of highly resistant clinical isolates [4] . C1186 has a growth rate similar to a laboratory wild-type . Thus , these resistance mutations may individually be low-cost , as found for some rifampicin-resistant patient isolates [11] , [12] or the strain may carry additional fitness-compensatory mutations [7] . A third possibility is that some resistance mutations reduce existing fitness costs while simultaneously decreasing susceptibility to the drug . This last possibility is highly relevant to the multi-step nature of fluoroquinolone-resistance development . We constructed 28 isogenic derivatives of the wild-type MG1655 , each mimicking in part the complex resistance genotype of C1186 ( Table 1 ) . Mutations were initially isolated spontaneously ( gyrA , parC ) or constructed by λ-red recombineering ( ΔmarR , ΔacrR ) and then separately introduced into MG1655 by P1 transduction . The gyrA and parC mutations were transduced by selection for a linked genetic marker ( introduced using λ-red recombineering , see Materials and Methods ) and at least 20 transductants of each cross were subsequently screened by phenotype ( MIC ) and DNA sequencing for the linked mutation . In every case only two phenotypic and genotypic classes were found , showing that the gyrA and parC mutations were not associated with other mutations . MIC for ciprofloxacin was measured for each strain ( Table 1 ) . The margin of error of MIC values is ±1 half-doubling step . Accordingly , any change that is at least 2-fold is significant . Single mutations in gyrA , S83L and D87N , increased MIC 24-fold and 16-fold respectively . Knockout mutations in marR and acrR increased MIC only 2–3-fold , while the substitution S80I in parC had no effect on MIC . Double mutation combinations had MIC's that were 8–64-fold wild-type level , with the combination ΔmarR+ΔacrR having the smallest increase . Certain combinations with parC were not tested because in E . coli parC mutations only selected after the prior occurrence of a mutation in gyrA . Triple mutation combinations have MIC's that were 31–2000-fold wild-type level . Most strains with three resistance mutations ( 5/9 ) , and all strains with 4 or 5 mutations had MIC's above the 1 µg/ml breakpoint that defines clinical resistance in Europe [13] equivalent to 64-fold wild-type MIC in these strains . On average there was a positive correlation between the number of resistance mutations carried by a strain and the MIC for ciprofloxacin ( Figure 1 ) . The 28 mutant strains were tested in growth competitions against wild-type to measure their Malthusian fitness [14] , [15] as a function of the resistance mutations they carried ( Table 1 ) . The fitness value associated with having either one or two resistance mutations ranged from ∼1 down to 0 . 82 per generation . Some single mutations ( gyrA S83L , gyrA D87N , and parC S80I ) were statistically neutral whereas others ( ΔmarR and ΔacrR ) caused a significant reduction in fitness ( 0 . 83 and 0 . 91 per generation , respectively ) . Strains carrying three resistance mutations had fitness values that ranged from ∼1 down to as low as 0 . 60 , with 8/9 strains suffering a fitness deficit of ≥5% per generation . Interestingly , the addition of a fourth resistance mutation to these strains restricted the minimum fitness value to 0 . 66 per generation , higher than that measured with three mutations ( Figure 1 ) . When all five resistance mutations were present the fitness value was 0 . 68 . Thus , some strains carrying 4 or 5 resistance mutations have a higher fitness than some strains carrying only 3 mutations . The major negative effect on fitness was associated with the presence of the marR and acrR mutations . Thus , fitness did not decrease as a simple function of the number of resistance mutations , but instead depended critically on the nature of those mutations . In general the addition of a resistance mutation to a strain was either neutral with respect to MIC and fitness , or it caused an increased MIC and / or decreased fitness ( Figure 1 , Table 1 ) . Across all the strains constructed decreased fitness was very strongly associated with the presence of one or both efflux mutations . In contrast , strain LM693 ( gyrA S83L , D87N; parC S80I ) has high level resistance ( MIC 32 ) with no significant reduction in fitness relative to the wild-type ( Table 1 ) . LM693 shows that it is possible to evolve high level resistance with no , or minimal , fitness costs . However , mutations up-regulating drug efflux are highly relevant to resistance evolution because they arise at a rate hundreds of times higher than mutations in the structural genes for topoisomerases . This is because the genetic target for knockout mutations in efflux regulating genes is much larger than the target for the specific amino acid substitutions required in topoisomerase genes . Thus , even though , as shown here , efflux mutations are fitness-costly and contribute relatively small increases in resistance , they occur very frequently , and are found in many resistant clinical isolates [4] . For three low-fitness strains , each carrying three resistance mutations including a marR mutation , the addition of an extra resistance mutation increased both MIC and fitness: LM695→LM707; LM882→LM707; and LM871→LM707 ( Figure 1 and Table 1 ) . In these strains the added mutation affected gyrA or parC . The increased fitness was statistically significant ( Students t-test , two-tailed , p<0 . 05 ) in two of the three cases , LM695>LM707 and LM882→LM707 ( Table 2 ) . The robustness of these results was verified by independently reconstructing each of these strains and re-measuring their MIC and fitness values . No significant differences were found from the original values . In addition , we tested these critical strains by de-construction experiments: replacing gyrA and/or parC mutations with equivalent wild-type genes and determining that the MIC and fitness values of the de-constructed strains were as expected . Based on these two experiments , re-constuction , and de-construction , we are confident that the isogenic strains do not carry any additional mutations affecting MIC or fitness . Thus , the addition of a single resistance mutation can increase growth fitness by 5–10% per generation and simultaneously increase MIC more than 40-fold ( Table 2 ) . The data in the previous section showed that the addition of a fourth resistance mutation to either LM695 or LM882 could significantly increase competitive growth fitness measured in vitro . This result would be more interesting from a clinical viewpoint if the measured increase in fitness was not exclusively an in vitro phenomenon . To test this competition experiments were made in a mouse UTI infection model [16] . Each of the strains ( LM695 , LM883 , and LM707 ) was competed against the isogenic wild-type in the mouse model and relative fitness expressed as a competitive index ( C . I . ) . Three different measures of C . I . could be obtained in this model: from the urine; the bladder; and the kidneys . For each strain and tissue there was a clear positive correlation between relative fitness in vivo and in vitro ( Figure 2 ) . Thus , the relative order of fitness values of these strains , initially measured in vitro , was confirmed in the physiologically more complex in vivo environments . The marR mutation makes the single largest contribution to loss of fitness in these 28 strains ( Table 1 ) . The marR mutation alone reduced fitness to 0 . 83±0 . 3 ( LM202 ) , and the average fitness of all strains carrying marR is 0 . 76 ( SD 0 . 08 ) . MarR protein regulates , directly and indirectly , transcription of many genes [17] and the reduction in fitness associated with loss of function mutations in mar is most likely because of disruption of gene regulation . This suggests a possible mechanism of fitness compensation associated with resistance mutations in topoisomerase genes . Thus , some topoisomerase mutations that alter chromosomal supercoiling levels may partially restore expression of growth-limiting gene ( s ) regulated by mar , reversing the fitness deficit . This hypothesis predicts that the improvement in fitness measured in LM695→LM707 should be associated with a change in global supercoiling level . This was tested by electrophoresis of pUC18 and pUC19 plasmids purified from MG1655 , LM695 and LM707 in chloroquine-agarose gels [18] . With this method we were unable to detect differences in plasmid topoisomer patterns ( data not shown ) . We also tested for differences in supercoiling using a reporter gene assay [19] , [20] . This was done by introducing plasmids carrying a luciferase reporter gene fused to each of two different promoters ( ptopA and pgyrA ) whose expression is sensitive to , respectively , increased or decreased supercoiling levels [20] . In this assay the ratio ( ptopA / pgyrA ) of reporter gene expression from these two promoters , defined as the quotient of supercoiling ( Qsc ) , is a measure of the relative level of negative supercoiling in isogenic strains [19] , [20] . Qsc was 2 . 9 ( SD 0 . 7 ) for MG1655 ( n = 6 ) , 2 . 6 ( SD 0 . 7 ) for LM695 ( n = 12 ) , and 3 . 6 ( SD 1 ) for LM707 ( n = 13 ) . In this assay neither of the mutant strains differed significantly from the wild-type , but LM695 and LM707 differed significantly from each other ( P value 0 . 01 , Students t-test , two tailed ) . From the expression data it was clear that the major cause of this difference in Qsc between LM695 and LM707 was due to a 25% decrease in expression from the gyrA promoter construct in LM707 relative to LM695 . Thus the introduction of the parC S80I mutation to make strain LM707 caused a significant increase in the Qsc , coincident with the improvement in fitness . Although we cannot explain the absence of an effect in the chloroquine-agarose assay , the reporter gene assay suggests that some changes in chromosomal supercoiling associated with the acquisition of topoisomerase mutations may provide a mechanism linking bacterial fitness and decreased susceptibility to fluoroquinolones . However , an alternative explanation is that the combination of mutations in LM707 reduces expression of gyrA by a mechanism that does not change global supercoiling levels , and that it is a consequence of the reduced expression of gyrA that causes the increase in growth fitness . Additional experiments will be required to distinguish between these models . In the sections above it was shown that the transfer of the parC S80I mutation into LM695 increased its MIC for ciprofloxacin and its competitive growth fitness versus the wild-type . This predicted that the resulting constructed strain , LM707 , should outcompete LM695 in a head-to-head competition , and also raised the following questions: ( i ) could a mutant with an increased MIC for ciprofloxacin and a higher growth fitness be selected spontaneously from LM695 in the absence of drug; and ( ii ) would such a mutant be exclusively associated with of the acquisition of the parC S80I mutation . These predictions and questions were addressed experimentally . First , in head-to-head growth competition experiments in LB medium with no drug , LM707 ( four resistance mutations ) outcompeted LM695 ( three resistance mutations ) , gaining ∼8% per generation , in good agreement with the relative differences in growth fitness of each strain versus the wild-type ( Table 1 ) . Second , 96 independent lineages of LM695 were grown in rich medium in the absence of drug for 4 growth cycles . Each growth cycle was inoculated with 2×106 cfu and grown to a total of 2×108 cfu ( ∼7 generations per growth cycle ) in a volume of 200 µL . An aliquot of 5×106 cfu from each lineage was tested , after the completion of 2 , 3 , and 4 cycles of growth , for the presence of resistant mutants on solid medium ( 3 µg/mL ciprofloxacin , 4×MIC ) . Growth of LM695 is completely inhibited on 3 µg/mL ciprofloxacin and the spontaneous mutation rate to resistance on this media , measured in fluctuation tests , is 2×10−8 . Accordingly , we expected that virtually none of the 96 independent lineages would contain a resistant mutant at the initiation of the experiment , but that a small number of resistant mutants would arise in each lineage during each growth cycle . If these mutants out-competed the parental LM695 they would be expected to increase relative to LM695 , and thus have a greater probability of being transferred to the next growth cycle . The number of lineages from which resistant mutants were obtained was found to increase with successive growth cycles: from 2/96 ( cycle 2 ) →8/96 ( cycle 3 ) →15/96 ( cycle 4 ) . Because the spontaneous mutation rate to resistance ( 2×10−8 ) if much lower than the number of cells being tested from each lineage ( 5×106 ) it is very unlikely that these mutants arose on the selective media . Instead , the most reasonable conclusion is that the resistant mutants arose spontaneously and randomly during the growth of lineages in the absence of drug and were enriched because they out-competed the parent population . Three random drug-resistant mutants were chosen from independent lineages and tested by DNA sequencing for the presence of mutations in gyrA and B , in parC and E , and in marR and acrR . In each case the mutations originally present in LM695 were confirmed and in addition each of the mutants was found to have acquired a new mutation in parC ( S80R in two cases , and E84K in one case ) . The MIC CIP of each of the mutants had increased from 0 . 75 to >32 µg/mL , and the exponential doubling time in rich medium had increased significantly , by ∼10% per generation ( Table 3 ) . Thus , the selection of a spontaneous parC mutation in LM695 decreased its susceptibility to ciprofloxacin and increased its growth rate . From these experiments we concluded that the phenotypes generated directly by strain construction ( LM695→LM707; reduced drug susceptibility and increased growth fitness ) could also be generated by a variety of spontaneous mutations , and that the growth advantage phenotype could be enriched by growth selection in the absence of drug .
A set of 28 isogenic E . coli strains was constructed and used to measure the relationship between the accumulation of fluoroquinolone resistance mutations , drug susceptibility , and growth fitness . The question was whether the accumulation of up to five resistance mutations , commonly found in resistant clinical isolates , would progressively reduce bacterial fitness . Most of the published data on the relationship between drug resistance and bacterial fitness would predict two possibilities: ( i ) that these mutations would cause little or no fitness cost , explaining their high frequency among resistant isolates; or ( ii ) that their accumulation would cause a progressive decrease in bacterial fitness , and require additional fitness-compensating mutations to restore fitness [6] , [7] . The data from this study support , in part , each of these scenarios ( Table 1 , and Figure 1 ) . However , they also revealed , for some combinations of resistance mutations , a positive relationship between reduced drug-susceptibility and increased bacterial fitness . This positive relationship could be another driving force in the development of increased resistance to these antibacterial drugs . Although this may be the first demonstration in bacteria that in the absence of an antimicrobial , selection can increase resistance to that antimicrobial , a similar phenomenon has been reported for HIV resistance to a protease inhibitor [21] . Thus , the phenomenon described here may have a broad biological significance . The main conclusions from the data set were the following: Molecular details of how particular parC and gyrA mutations together improve the fitness of a resistant strain with a marR mutation are beyond the scope of this study but we can suggest the outlines of a model . Growth rate depends on the rate of transcription regulated in accordance with physiological demands [22] . The MarR protein regulates , directly and indirectly , the transcription of many genes [17] and it is probable that the severe reduction in fitness associated with ΔmarR ( Table 1 ) is because it causes inappropriate patterns of transcription regulation . We suggest that specific mutant forms of DNA gyrase and topoisomerase IV , possibly by acting in concert to influence the level of superhelicity in DNA , restore appropriate levels of gene expression at some loci where the loss of MarR regulation has a negative impact on growth rate [23] , [24] . The particular gyrA and parC resistance mutations studied here are clinically relevant , being among the most common found in fluoroquinolone-resistant clinical isolates of E . coli [4] . Among 30 resistant UTI isolates analyzed: 30/30 had the gyrA S83L mutation; 18/30 had the double mutation gyrA S83L , D87N; 22/30 had the parC mutation S80I; and 15/30 had the triple combination gyrA S83L , D87N , parC S80I ( 23/30 had some form of triple mutation combination including other substitutions at position 87 of gyrA and/or position 80 of parC ) . Thus , the combination of target mutations studied here is typical of resistant clinical isolates . As measured by organic solvent tolerance , drug efflux was phenotypically up-regulated in 15/30 of these resistant isolates , associated in most cases with mutations in acrR and/or marOR [4] . The observed frequency of the efflux phenotype is very low relative to the frequency of specific gyrase and topoisomerase mutations in the same isolates , and given the much higher expected probability of mutations that knock out the function of efflux regulator genes it suggests selection against such mutations . This under-representation is consistent with our finding that mutations up-regulating efflux pumps carry significant fitness costs . The growth fitness of clinical isolates cannot be meaningfully compared with the data presented in this paper , in part because clinical isolates are not isogenic , and in part because clinical isolates will already have evolved to ameliorate the fitness costs , if any , associated with their resistance determinants . We believe that our data provide an insight into the likely initial effects on relative fitness and drug susceptibility of different pathways of resistance development . In particular , that one consequence of a following a high-probability but low-fitness evolutionary pathway may be that one or more steps may be driven by selection for increased fitness in the absence of drug exposure . How fluoroquinolone resistance evolves in nature will depend on the genotype being selected and on the selective environment [25] but it is likely include mutational steps that reduce bacterial fitness . Bacteria that progress down an evolutionary path with reduced fitness relative to a competing population may be driven to extinction , or may , given the opportunity by mutation , acquire a change that increases their relative fitness thus improving their chances of survival . Evolutionary paths that could be taken on the road to extinction or antibiotic resistance are outlined in Figure 3 . The low-fitness mutants LM695 and LM882 each have MICs that lie under the resistance breakpoint for ciprofloxacin [26] , [27] . Such mutants are at a critical stage in resistance development: having low fitness they may be driven to extinction by natural selection; being under the resistance breakpoint they may avoid detection in a clinical setting; however , they are , as shown here , only one mutational step away from a high-level resistance phenotype with increased fitness , without additional exposure to the drug . The magnitude of these co-selected changes in fitness and susceptibility are significant ( Table 2 ) . These data argue in favor of testing anti-mutant dosing strategies , or other measures that could prevent the enrichment of low-level resistant mutants [28] .
C1186 is a multiply mutant fluoroquinolone resistant UTI isolate previously described [4] . E . coli K12 MG1655 wild-type was the starting strain for all constructions ( Table 1 ) . Liquid growth medium was Luria broth ( LB ) while solid medium was Luria-Bertani agar ( LA ) . Strains were grown at 37°C . Ciprofloxacin ( Bayer HealthCare AG , Wuppertal , Germany ) was dissolved in 0 . 1 M NaOH at 100 µg/mL then further diluted in LA in selective plates . MIC was determined by Etest ( AB BIODISK , Solna , Sweden ) on Mueller-Hinton agar plates incubated for 16 to 18 h at 37°C with quality control reference strains [4] as recommended by the Clinical and Laboratory Standards Institute ( www . clsi . org ) . Spontaneous resistance mutations in gyrA and parC were selected sequentially in E . coli MG1655 in LB with ciprofloxacin at 2–8-fold MIC and mutations were identified by DNA sequencing . Individual mutations were always moved into a clean genetic background ( MG1655 ) after initial selection . Deletion-replacement mutations in marR , acrR , yfaH , metC and araB were made by λ-red recombineering [29] in NC397 ( a Lac+ Nad+ derivative of DY329 ) using PCR amplified linear DNA from pCP16 with an FRT-bounded TcR cassette [30] . The PCR reaction protocol was 95°C 5 min followed by 30 cycles of 95°C 15 sec , 55°C 20 sec , 72°C 240 sec . PCR primer sequences with details of the deletion-replacement boundaries are shown in Supporting Information ( Table S1 ) . Isogenic derivatives of MG1655 were constructed by phage P1-mediated transduction . Transduction of the deletion-replacement mutations in marR and acrR was selected directly on LA+Tc . When transducing the gyrA and parC mutations selection was made for the linked markers , yfaH<>Frt::TcR::Frt and metC<>Frt::TcR::Frt , each ∼10 kb from gyrA and parC , respectively . Note that we are using the symbol <> to indicate a replacement generated by λ red homologous recombination technology . After transduction the TcR marker was removed by Flp-catalyzed excision expressed following transformation with pCP20 [30] . All strain constructs were confirmed by DNA sequencing . LM347 ( ΔaraB<>FRT ) was used as the standard wild type strain in growth competitions against which each of the constructed mutant strains was competed . This strain was tested against its parent MG1655 showing that the ΔaraB mutation was neutral ( relative growth rate 1 . 002±0 . 005 ) . To support statistical analysis each competition was tested in at least 5 independent experiments ( Table 1 ) . To initiate growth competitions , each strain was grown in LB at 37°C 12 h , mixed in a 1∶1 ratio , diluted 10−3 into LB , then grown 23 h to complete a growth cycle . Each successive growth cycle was initiated by diluting the mixture 10−3 into LB . Each competition experiment ( 4 cycles ) was made the number of times indicated in Table 1 . For some low-fitness strains such as LM595 the mutant population became too low to detect after the second or third cycle of competition . After the initial mixing , and after each growth cycle , appropriate dilutions of the mixture were plated onto MacConkey agar plates containing 1% arabinose . Plates were incubated 37°C overnight . Red ( ara+ ) and white ( ΔaraB ) colonies were scored . The change in the ratio of mutant/wild-type was used to estimate the selection coefficient per generation of each of the constructed strains according to the formula: S = ln2 ( mutant/wild-type ) / generation [31] . Relative fitness per generation with respect to the wild-type LM347 is defined as S+1 . Note that the fitness defects could be due to defects at any stage of the growth cycle ( lag , exponential , stationary phase ) . Relative bacterial fitness in vivo was measured using an established urinary tract infection model [16] . Details , including mouse strain and ethical permission , are given in Text S1 . Competitive index was calculated as the geometric mean of the ratio of mutant/wild-type bacteria isolated from each organ ( urine , bladder , kidneys ) of 8 mice per experiment , normalised to the ratio at the time of inoculation . Exponential doubling times were calculated by measuring the increase in optical density at 600 nm ( OD600 ) at 10-min intervals , using a BioscreenC machine ( Oy Growth Curves Ab Ltd . Helsinki , Finland ) . The mutation rate of LM695 to resistance on LA+3 µg/mL ciprofloxacin ( CIP ) is 2×10−8 measured by fluctuation test . Independent lineages were inoculated in 96×200 µL wells , and grown 24 h at 37°C to ∼109 CFU/mL . 2 µL was transferred to initiate a new growth cycle . 5 µL was plated on LA+3 µg/mL CIP to assay for resistant mutants . The relative supercoiling degree was determined using a published assay . The quotient of supercoiling , Qsc [19] is defined as the ratio of luciferase activity from two different supercoiling sensitive promoters , ptopA-luc , and , pgyrA-luc , [20] . Details are given in Text S2 . | The increasing frequency of human pathogens resistant to important classes of antibiotics poses a serious and growing challenge for medicine and society . We need improved strategies to reduce the rate of resistance development , for established and novel drugs , based on knowledge of the factors that drive the increase in resistance . Resistance to fluoroquinolones in most bacteria develops via a series of sequential genetic changes affecting several different genes . These are selected and enriched in bacterial populations by exposure to the drug . Relevant factors driving this increase include overuse , and inappropriate use , of these drugs . In this paper we show that mutant bacteria with low-level resistance ( not itself a problem to treat with standard drug doses ) can evolve by natural selection ( for improved growth rate ) to acquire mutations that dramatically increase their level of drug resistance . This means that we may need to consider how to reduce inappropriate drug use that can enrich for bacteria with low-levels of resistance , because at that stage some of the mutant bacteria in the population may continue to evolve higher level resistance even in the absence of any further drug exposure . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"infectious",
"diseases/antimicrobials",
"and",
"drug",
"resistance",
"microbiology"
] | 2009 | Interplay in the Selection of Fluoroquinolone Resistance and Bacterial Fitness |
Understanding how genetic variation interacts with the environment is essential for understanding adaptation . In particular , the life cycle of plants is tightly coordinated with local environmental signals through complex interactions with the genetic variation ( G x E ) . The mechanistic basis for G x E is almost completely unknown . We collected flowering time data for 173 natural inbred lines of Arabidopsis thaliana from Sweden under two growth temperatures ( 10°C and 16°C ) , and observed massive G x E variation . To identify the genetic polymorphisms underlying this variation , we conducted genome-wide scans using both SNPs and local variance components . The SNP-based scan identified several variants that had common effects in both environments , but found no trace of G x E effects , whereas the scan using local variance components found both . Furthermore , the G x E effects appears to be concentrated in a small fraction of the genome ( 0 . 5% ) . Our conclusion is that G x E effects in this study are mostly due to large numbers of allele or haplotypes at a small number of loci , many of which correspond to previously identified flowering time genes .
The transition from vegetative to reproductive growth is a key developmental step in the life cycle of higher plants , and its timing is tightly regulated by both genes and environment , often in an interactive manner , so that the effect of genetic variants depends on the environment [1 , 2] . Such genotype by environment interactions ( G x E ) have long been of interest to quantitative geneticists , as they are crucial for local adaptation [3 , 4] and for improving agricultural yield . In particular , understanding G x E variation is considered essential for predicting the effects of climate change on ecology and agriculture [2 , 5] . Analytically , G x E can be described in terms of “reaction norms” as genetic variation in the phenotypic response to the environment [2] . The phenotypic variation can be decomposed into genetic effects that are the same across environments ( G ) , effects that are different across environments ( G x E ) , and non-genetic environmental effects ( E ) . Many approaches have been proposed to identify loci contributing to G x E variation [2 , 6] . In the context of genome-wide association studies ( GWAS ) , Korte et al . [7] proposed a multi-trait mixed model ( MTMM ) that can also be used to study G x E [2 , 5 , 7] . Attempts to map G x E variation , whether using classical linkage mapping or GWAS [4 , 5 , 8–10] , have generally revealed loci explaining only a small fraction of the G x E variation . The most likely explanation for this “missing” G x E heritability is that the underlying genetic architecture involves either rare alleles of relatively large effect [2] , or large numbers of polymorphisms of small effect [5 , 8 , 9] . Here we present a GWAS for flowering time at two temperatures ( 10°C and 16°C; see Methods ) in a population of 173 A . thaliana lines from Sweden [11] ( S1 Fig , S1 Table ) . Our goal was twofold: first , we wanted to investigate our ability to map polymorphisms responsible for G x E interactions; second , we wanted to characterize the main determinants of flowering time variation in Sweden , because although many GWAS have mapped genes responsible for flowering time variation in A . thaliana [5 , 12–15] , this has almost always been done in global samples , and there is reason to believe that the relatively small number of significant associations in these attempts is due to excessive genetic heterogeneity in these samples . The genetics of flowering time in local samples could be simpler , increasing the power of GWAS [12] .
The increase in growing temperature from 10°C to 16°C had a dramatic effect on flowering behavior , significantly accelerating flowering in 29% of the lines , significantly decelerating flowering in 16% of the lines , and generally increasing the variance both within and between lines ( t-test , q-value < 0 . 01; Fig 1; S1–S2 Tables ) . Broad-sense heritabilities ( H2 ) were extremely high ( over 90% ) at both temperatures ( albeit significantly lower at 16°C , p < 0 . 01 ) , demonstrating strong genetic effects , in agreement with published results ( Table 1 ) [12 , 16 , 17] . We partitioned the variance in flowering time using a model with four components: genotype ( G , the variance attributable to genome-wide relatedness ) , environment ( E ) , G x E , and noise ( see Methods ) . This analysis revealed massive G x E effects . The G x E effects are largely due to the differences in the reaction norm between the subsets in Fig 1 . For example , 67 . 9% of the variation among lines with accelerated flowering is due to direct genetic effects ( Table 2 ) . We attempted to map the polymorphisms responsible for the G x E effect using genome-wide association using a mixed model that allows multiple correlated traits ( MTMM [7] ) . Three different association tests were carried out: a “full SNP test” that compares a full model including the effect of marker genotype and its interaction with environment against a model with no ( fixed ) SNP effect; “common SNP effect test” that compare a model with genetic marker ( a genetic model ) against no SNP effect , and; “interaction ( GSNP x E ) effect test” that compares the full model against the genetic model [7] . In agreement with previous results , MTMM appeared to correct for confounding population structure well , whereas a standard multi-linear regression model ( MLR ) produced massively skewed p-values ( S2 Fig ) . The full SNP test identified two peaks with genome-wide significance ( Fig 2A ) . The strongest association was centered around position 3 , 180 , 721 on chromosome 5 , in the promoter region of the well-known flowering regulator FLOWERING LOCUS C ( FLC ) ( Fig 2B ) , which has previously been shown to play a major role in natural variation for flowering time , but has generally been difficult to map using GWAS [5 , 12 , 13] , presumably because of extensive genetic heterogeneity [18 , 19] . Interestingly , the FLC peak can be seen using both the common SNP and the GSNP x E effect tests , but was significant in neither , suggest that it has a weak GSNP x E effect as well as a weak common SNP effect . The behavior of the second strong association is very different . This association , centered on position 9 , 005 , 735 on chromosome 2 , is more significant under the common SNP effect test , and is not present under the GSNP x E effect test , suggesting that the polymorphism has the same effect in both temperatures . The peak is quite broad ( Fig 2C ) and contains approximately 13 genes , none of which are known to be involved in regulating flowering time . However , one of them , FIONA1 ( FIO1 ) , is related to the circadian clock , and the null mutant shows early flowering [20] . Furthermore , GWAS using indel markers identified the most significant association ( p-value = 2 . 97E-08; Fig 2D ) as a insertion of two nucleotides in the 9th ( last ) exon of FIO1 , which would result in a frameshift , however , this exon appears not to be present in mRNA-seq data from leaves [21] , and appears to be specific to A . thaliana . A stop codon is found 26-amino acids upstream of the insertion in the closely related Arabidopsis lyrata and Capsella rubella . The putative frameshift polymorphism is due to eight vs nine GA repeats , and is in strong linkage disequilibrium with several non-synonymous polymorphisms , which are slightly less strongly associated with flowering time ( S3 Fig ) . Although definitive proof in the form of transgenic experiments ( allele swapping ) is missing , polymorphism in FIO1 is a strong candidate for the major common effect on chromosome 2 . The common SNP effect test revealed no further significant associations , and the GSNP x E effect test revealed no significant associations at all , despite the fact that G x E effects account for 66% of the phenotypic variance ( Fig 2 , Table 2 ) . Our GWAS identified two associations with genome-wide significance , one of which corresponds to a clear a priori candidate ( FLC ) . Given that the number of a priori candidates ( genes known to be involved in flowering time ) is on the order of a percent of total genes ( S3 Table ) , one out of two is obviously more than expected by chance . To investigate whether there is an overrepresentation of a priori candidates among associations that do not reach genome-wide significance as well , we calculated the enrichment as a function of significance threshold [12] . Because an association that is significant at a certain level will generally be surrounded by many SNPs that are less strongly associated ( giving rise to a peak of association ) , we calculated enrichment at a given level after removing all peaks ( defined as 30 kbp windows ) containing SNPs that were already significant using a more stringent threshold . For the full SNP test , a significant enrichment of a priori candidates persists as we increase the significance threshold ( i . e . , lower the stringency ) to 10−5 ( Fig 3 ) . Although associations at this level are far from significant in the genome-wide sense , the enrichment of a priori candidates implies that the false-discovery rate ( FDR ) among these candidates is less than 20% [12] . Three a priori candidates were identified using this approach ( Table 3 ) : FLC ( which also reaches genome-wide significance ) ; SHORT VEGETATIVE PHASE ( SVP ) , which mediates ambient temperature signaling by regulating FLOWERING LOCUS T ( FT ) [22] , and has been shown to be involved in natural variation in other samples [23]; and VERNALIZATION INSENSITIVE 3 ( VIN3 ) , which is involved in the epigenetic silencing of FLC during vernalization , but has hitherto not been identified in natural populations [20 , 24] . Some of the associated SNPs were found in promoter regions ( common SNP effects of FLC , VIN3 ) . These SNPs are excellent candidates for being causal , and it seems likely that we simply lack the power to pick them up in a genome-wide scan . What the FDR is among the approximately 10 peaks that do not correspond to a priori candidates but are significant using the same threshold is not known ( S4 Table ) . The results for the common SNP effect test were very similar to the full SNP test , and the same a priori candidates were identified ( Fig 3 , Table 3 ) . However , the GSNP x E effect test showed no evidence for significant enrichment at any p-value threshold , suggesting that if low power is the reason for the missing G x E associations , then the power is low indeed . Finally , we note that if causal variants are strongly correlated with global relatedness , power to detect them may be greatly decreased [25 , 26] . We therefore scanned for associations without correction for relatedness ( using MLR ) , as well . The associations from such an analysis are of course extremely inflated , but it is possible to use the enrichment analysis described above , as it does not rely on well calibrated p-values ( S2 , S4 Figs ) . However , this approach identified only a subset of the candidate genes already identified using MTMM . Statistical power in GWAS may be decreased by allelic heterogeneity , which reduces the marginal contribution of individual polymorphisms at a genetic locus . One possible way around this is to consider the joint effect of all polymorphisms at a genetic locus using a mixed model . Instead of mapping individuals SNPs as fixed effects , we estimate the variance component that is due to local relatedness around each gene ( using a 15 kbp window on each side of the coding region ) and compare that to the variance component that is due to the rest of the genome [21] . We refer to these effects as “local” and “global” , respectively , and we also include environmental and G x E components . Three different tests were carried out: a “full local test” that compares a full model , including local and global effects and their interactions with E , with a null model that does not include any local effect; a “common local effect test” that compares a local model that does not include a Glocal x E with the null model , and; an “interaction ( Glocal x E ) effect test” that compares the full model with the local model . For each test , log-likelihood ratios were calculated ( see Methods ) . Result for the full local and the common local effect tests were strongly correlated with their corresponding GWAS results ( presented above ) , especially for genes with reasonably strong association with flowering , while GSNP x E and Glocal x E showed much lower correlation ( S5 Fig ) . Because the variance component likelihood ratios are not calibrated , it is difficult to say whether any particular effect is significant . However , we can assess this using overrepresentation of a priori candidates as for MTMM above . In all tests ( full local , common local and Glocal x E ) , a significant enrichment of a priori candidates exist for likelihood ratios of 5 or higher , for which FDR is less than 20% ( Fig 4 ) . Notably , this effect was observed for the Glocal x E effect test as well , whereas GSNP x E showed no evidence of overrepresentation ( Fig 3 ) . Thus the variance component analysis appears to capture G x E effects not captured by the marginal SNP GWAS . A total of four flowering time genes showed significant peaks at the log-likelihood threshold of 5 ( Table 4 ) . FLC and VIN3 showed high common local effect as well as common SNP effect , while FPA , an FLC suppressor in the autonomous pathway [28] , showed up as a Glocal x E locus . Furthermore , CENTER CITY ( CCT ) was significant in using the full local test . CCT , also known as CRYPTIC PRECOCIOUS ( CRP ) , is a flowering regulator that acts as a promoter of FT and a suppressor of FLC [29 , 30] . It is closely linked to the well-known flowering time locus FRIGIDA ( FRI ) and has previously been detected in GWAS [12] . Fig 5 shows the distribution of common ( i . e . , G ) and G x E signals across the genome , for SNPs as well as for local variance components . The three highest peaks of Glocal ( S5 Table ) overlap peaks of common GSNP effect centered around FIO1 on chromosome 2 , and FLC on chromosome 5 , and position 23 , 544 , 472 on chromosome 5 . This overlap suggests that a small number of SNPs identified by MTMM might be responsible for the local variance components . Although there are no obvious flowering time candidates in the final region on chromosome 5 , a recent study reported that gene in the region , MULTICOPY SUPRESSOR OF IRA 1 ( MSI; AT5G58230 ) delays the transition to flowering [31] . The most significant peak of Glocal x E only was found at the top of chromosome 1 ( 963 , 400-1 , 053 , 719 ) and includes eight genes , none of which are known to be involved in flowering . Finally , we consider the question of genetic architecture . For a Mendelian trait , all the phenotypic variation is due to a single locus , whereas for a truly Fisherian trait , the contribution of a genomic region should be proportional to its size ( relative to the entire genome ) . Flowering time is clearly neither . As shown in Table 5 , the 144 SNPs identified using MTMM ( with the full SNP test using the 20% FDR defined in Fig 3 ) jointly explain 22% of the phenotypic variation as common ( to both environments ) genetic variation ( G ) , and 31% as G x E variation . The remaining 3 . 7 million SNPs ( of which 1 million have a minor allele frequency less than 0 . 1 ) explain only 6% as G and 35% as G x E . If we instead turn to the local variance components , the identified regions , comprising roughly 2% of the genome , explain 26% as G and 67% as G x E ( randomly chosen regions explain on average at total of 7 . 5%; p = 0 . 001; S6 Fig ) , supporting the observation that the local variance component approach seems to have significantly greater power to capture G x E effects , but does not do better when it comes to common effects . Importantly , the local variance components explain essentially all the available genetic variation , and combining SNPs and local variance components yield almost no improvement ( Table 5 ) . It is also worth noting that the less than 10% of the identified regions that contain one of the a priori candidates explain almost 40% of the variation , a clearly significant overrepresentation ( p = 0 . 001; S6 Fig ) .
The main purpose of this study was to investigate the genetic architecture of G x E variation using a population and experimental setting where such variation was massive . Roughly 66% of the variation for flowering time among lines across environments in this study is due to G x E ( Table 2 ) , yet a standard GWAS method failed to detect a single significant SNP association . Indeed , even when considering enrichment for a priori candidates using less stringent thresholds , there is no trace of G x E associations . The same was true using various summaries of the traits , like the slope of the reaction norm . In contrast , there is ample evidence for polymorphisms that do not interact with the environment ( include two that reach genome-wide significance ) , although this type of variation is only 28% of the phenotypic variation . The much-discussed “missing heritability” problem in human genetics refers to the fact that individually identifiable ( mappable ) SNPs do not explain the genetic variation [32] . Although many explanations have been proposed , the simplest one is that the marginal contributions of the underlying variants are too small ( due to a combination of allele frequency and effect size ) for them to be identified given the statistical power of the study . This explanation is supported by studies that increase power by increasing sample size [33] or that use variance components to estimate the joint contribution of all SNPs rather than trying to identify marginal effects [34] . In the present study , we have no “missing heritability” for common genetic variation , since the SNPs we identified account for almost all of this ( 22% vs 28%; Tables 2 and 5 ) . However , we do have “missing heritability” for G x E variation , where the identified SNPs explain less than half of the existing variation ( 31% vs 66%; Tables 2 and 5 ) . Why this difference between G and G x E ? The obvious explanation is again power . Under some scenarios , G x E effects are more difficult to detect for purely statistical reasons [7] , and it is also possible that the distribution of allele frequencies and/or effect sizes differ . Simulation studies have likewise suggested that substantial genetic risk score-by-environment interactions may exist , although marginal G x E effects are undetectable [35] . The notion that power is involved is supported by the fact that we are able to account for the missing G x E variation fully using variance component methods that estimate the joint contribution of multiple SNPs ( Tables 2 and 5 ) . However , these results also demonstrate that the G x E variation is not Fisherian in the sense of being spread out infinitessimally thinly across the genome . Instead , 8 small regions , comprising about 0 . 5% of the genome , appear to explain almost all the G x E variation ( S5 Table ) . This suggests that G x E variation for flowering is due to a relatively small number of genes harboring a large number of functionally distinct alleles ( or haplotypes ) , i . e . , allelic rather than genetic heterogeneity . This is consistent with what is known about allelic variation at several flowering time loci [18 , 36 , 37] , and perhaps also with the general observation that different linkage mapping experiments , which are insensitive to allelic heterogeneity , consistently seem to identify the same small number of flowering loci , several of which have not been identified using GWAS [38 , 39] . Dissecting these complex regions and haplotypes further will likely require painstaking experimental work , as linkage disequilibrium is typically too extensive for fine-mapping [12 , 18] . It should be noted that the extensive allelic heterogeneity for G x E is in contrast to several examples from crops [40 , 41] . A possible explanation for this is that domestication and breeding increased the frequency of rare alleles . The pattern in A . thaliana , on the other hand , suggests strong local adaptation . There is no obvious correlation between flowering time and geography in our data , but this is not surprising given the strong G x E effects , and the existence of micro-scale climate variation . In order to elucidate the selective forces acting on flowering time variation , field experiments will be required [14 , 42] . A secondary purpose of this project was to investigate the genetics of flowering time variation in a local population sample from Sweden . From an a priori list of more than hundred flowering time genes , we identified five genes , FLC , SVP , VIN3 , CCT and FPA at an FDR of less than 20% ( S5 Table ) . FLC , in particular , clearly has a major effect , in agreement with its role as a major flowering repressor and central player in the vernalization response [43] . Although flowering time is determined by the interaction of huge networks that include the photoperiod , gibberellin , vernalization , temperature , autonomous pathways [44] , we found that all identified flowering time genes in our analysis were tightly related to the regulation of FLC and FT ( S7 Fig ) . Briefly , floral initiation starts immediately by upregulation of FT when warm temperature returns after FLC is epigenetically silenced by VIN3 during a cold period [20 , 24] . CCT and FPA suppresses FLC in the autonomous pathway [29 , 30 , 45] . SVP has been reported as another flowering regulator that suppresses FT independent of FLC [46] . It should be noted that CCT is closely linked to FRIGIDA ( FRI , distance is 13 . 97 kbp ) , a strong up-regulator of FLC [47–49] known to harbor , strong allelic heterogeneity and massive haplotype sharing in global samples ( over 250 kbp [50 , 51] ) . Although FRI is not known to be segregating in the Swedish population , it is clearly possibly that FRI alleles could lead to confounding at CCT [12] . In addition to known flowering time genes , we also identified one possible novel gene . Although our FDR approach only works for a priori candidates , the peak in FIO1 is clearly significant at the genome-wide level , and the association is currently being confirmed experimentally . With the exception of FLC and SVP , none of the genes identified here have previously been shown to be important in natural variation . This demonstrates the advantages of using a local sample for GWAS when working on a trait important in local adaptation , and is in agreement with the G x E results above . Given that allelic heterogeneity can have a major effect on the power of GWAS even within Sweden , it should come as no surprise that flowering time is recalcitrant to GWAS in global samples [12] .
173 Swedish lines and Col-0 were used for experiments ( S1 Table ) . These lines , and all genome information , including SNPs and short indels , are described elsewhere [11] . Seeds were sown on soil and stratified for three days at 4°C in the dark . They were then transferred into a single pot after germination . All plants were grown in MTPS144 Conviron walk-in growth chambers ( Winnipeg , MB , Canada ) set to long-day conditions ( 16 h photoperiod ) under 10°C or 16°C constant temperatures . Periods from germination to presence of first buds were recorded as flowering time for multi-individuals for each line . Measurements were taken twice a week , until 190 days from germination . | Many traits are influenced by genetic variation in interaction with the environment , so called G x E variation . In agriculture , for example , different varieties are optimal in different environments . In evolution , G x E is also crucial for local adaptation . Identifying the genes underlying G x E has proven extremely challenging , however . Using a collection of inbred lines of the model plant Arabidopsis thaliana , we meausured flowering time under two temperature regimes , and scanned the genome for polymorphisms responsible for variation in this trait . Although most of the variation is due to G x E , genome-wide scans using SNPs only revealed direct genetic effects ( G ) , and failed to reveal any significant G x E associations . In contrast , scanning the genome using local windows of polymorphism suggested that almost all the observed variation can be explained by 2% of the genome . Previously identified flowering time genes are strongly overrepresented in these regions , and our results are compatible with a model under which G x E is mainly due to many alleles at a relatively small number of loci . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | "Missing" G x E Variation Controls Flowering Time in Arabidopsis thaliana |
Between October 2013 and April 2014 , more than 30 , 000 cases of Zika virus ( ZIKV ) disease were estimated to have attended healthcare facilities in French Polynesia . ZIKV has also been reported in Africa and Asia , and in 2015 the virus spread to South America and the Caribbean . Infection with ZIKV has been associated with neurological complications including Guillain-Barré Syndrome ( GBS ) and microcephaly , which led the World Health Organization to declare a Public Health Emergency of International Concern in February 2015 . To better understand the transmission dynamics of ZIKV , we used a mathematical model to examine the 2013–14 outbreak on the six major archipelagos of French Polynesia . Our median estimates for the basic reproduction number ranged from 2 . 6–4 . 8 , with an estimated 11 . 5% ( 95% CI: 7 . 32–17 . 9% ) of total infections reported . As a result , we estimated that 94% ( 95% CI: 91–97% ) of the total population of the six archipelagos were infected during the outbreak . Based on the demography of French Polynesia , our results imply that if ZIKV infection provides complete protection against future infection , it would take 12–20 years before there are a sufficient number of susceptible individuals for ZIKV to re-emerge , which is on the same timescale as the circulation of dengue virus serotypes in the region . Our analysis suggests that ZIKV may exhibit similar dynamics to dengue virus in island populations , with transmission characterized by large , sporadic outbreaks with a high proportion of asymptomatic or unreported cases .
Originally identified in Africa [1] , the first large reported outbreak of Zika virus ( ZIKV ) disease occurred in Yap , Micronesia during April–July 2007 [2] , followed by an outbreak in French Polynesia between October 2013 and April 2014 [3] , and cases in other Pacific countries [4 , 5] . During 2015 , local transmission was also reported in South American countries , including Brazil [6 , 7] and Colombia [8] . Transmission of ZIKV is predominantly vector-borne , but can also occur via sexual contact and blood transfusions [9] . The virus is spread by the Aedes genus of mosquito [10] , which is also the vector for dengue virus ( DENV ) . ZIKV is therefore likely to be capable of sustained transmission in other tropical areas [11] . As well as causing symptoms such as fever and rash , ZIKV infection has also been linked to increased incidence of neurological sequelae , including Guillain-Barré Syndrome ( GBS ) [12 , 13] and microcephaly in infants born to mothers who were infected with ZIKV during pregnancy [14] . On 1st February 2015 , the World Health Organization declared a Public Health Emergency of International Concern in response to the clusters of microcephaly and other neurological disorders reported in Brazil , possibly linked to the recent rise in ZIKV incidence . The same phenomena were observed in French Polynesia , with 42 GBS cases reported during the outbreak [13 , 15] . In addition to the GBS cluster , there were 18 fetal or newborn cases with unusual and severe neurological features reported between March 2014 and May 2015 in French Polynesia [16] , including 8 cases with microcephaly [17] . Given the potential for ZIKV to spread globally , it is crucial to characterize the transmission dynamics of the infection . This includes estimates of key epidemiological parameters , such as the basic reproduction number , R0 , defined as the average number of secondary cases generated by a typical infectious individual in a fully susceptible population , and how many individuals ( including both symptomatic and asymptomatic ) are typically infected during an outbreak . Such estimates could help assist with outbreak planning , assessment of potential countermeasures , and the design of studies to investigate putative associations between ZIKV infection and other conditions . Islands can be useful case studies for outbreak analysis . Small , centralized populations are less likely to sustain endemic transmission than a large , heterogeneous population [18] , which means outbreaks are typically self-limiting after introduction from external sources [19] . Further , if individuals are immunologically naive to a particular pathogen , it is not necessary to consider the potential effect of pre-existing immunity on transmission dynamics [20] . Using a mathematical model of vector-borne infection , we examined the transmission dynamics of ZIKV on six archipelagos in French Polynesia during the 2013–14 outbreak . We inferred the basic reproduction number and the overall size of the outbreak , and hence how many individuals would still be susceptible to infection in coming years .
We used weekly reported numbers of suspected ZIKV infections from the six main regions of French Polynesia between 11th October 2013 and 28th March 2014 ( Table 1 ) , as detailed in the Centre d’hygiène et de salubrité publique situation reports [21 , 22] . Confirmed and suspected cases were reported from sentinel surveillance sites across the country; the number of such sentinel sites varied in number from 27–55 during the outbreak ( raw data are provided in S1 Dataset ) . Clinical cases were defined as suspected cases if they presented to health practitioners with rash and/or mild fever and at least two of the following signs: conjunctivitis , arthralgia , or oedema . Suspected cases were defined as a confirmed case if they tested positive by RT-PCR on blood or saliva . In total , 8 , 744 suspected cases were reported from the sentinel sites . As there were 162 healthcare sites across all six regions , it has been estimated that around 30 , 000 suspected cases attended health facilities in total [21] . For each region , we calculated the proportion of total sites that acted as sentinels , to allow us to adjust for variation in reporting over time in the analysis . Population size data were taken from the 2012 French Polynesia Census [23] . In our analysis , the first week with at least one reported case was used as the first observation date . We used a compartmental mathematical model to simulate vector-borne transmission [24 , 25] . Both people and mosquitoes were modelled using a susceptible-exposed-infectious-removed ( SEIR ) framework . This model incorporated delays as a result of the intrinsic ( human ) and extrinsic ( vector ) incubation periods ( Fig 1 ) . Since there is evidence that asymptomatic DENV-infected individuals are capable of transmitting DENV to mosquitoes [26] , we assumed the same for ZIKV: all people in the model transmitted the same , regardless of whether they displayed symptoms or were reported as cases . The main vectors for ZIKV in French Polynesia are thought to be Ae . aegypti and Ae . polynesiensis [12] . In the southern islands , the extrinsic incubation period for Ae . polynesiensis is longer during the cooler period from May to September [27] , which may act to reduce transmission . Moreover , temperature can also influence mosquito mortality , and hence vector infectious period [28] . However , climate data from French Polynesia [29] indicated that the ZIKV outbreaks on the six archipelagos ended before a decline in mean temperature or rainfall occurred ( S1 Fig ) . Hence it is likely that transmission ceased as a result of depletion of susceptible humans rather than seasonal changes in vector transmission . Therefore we did not include seasonal effects in our analysis . In the model , SH represents the number of susceptible people , EH is the number of people currently in their incubation period , IH is the number of infectious people , RH is the number of people that have recovered , C denotes the cumulative number of people infected ( used to fit the model ) , and N is the human population size . Similarly , SV represents the proportion of mosquitoes currently susceptible , EV the proportion in their incubation period , and IV the proportion of mosquitoes currently infectious . As the mean human lifespan is much longer than the outbreak duration , we omitted human births and deaths . The full model is as follows: d S H / d t = - β H S H I V ( 1 ) d E H / d t = β H S H I V - α E H ( 2 ) d I H / d t = α H E H - γ I H ( 3 ) d R H / d t = γ I H ( 4 ) d C / d t = α H E H ( 5 ) d S V / d t = δ - β V S V I H N - δ S V ( 6 ) d E V / d t = β V S V I H N - ( δ + α V ) E V ( 7 ) d I V / d t = α V E V - δ I V ( 8 ) Parameter definitions and values are given in Table 2 . We used weakly informative prior distributions for the human latent period , 1/αH , infectious period , 1/γ , extrinsic latent period , 1/αv , and mosquito lifespan , 1/μ . For these prior distributions , we made the assumption that human latent period was equivalent to the intrinsic incubation period , i . e . that no transmission typically occurs before symptom onset . A systematic review of the incubation period for ZIKV in humans estimated a mean value of 5 . 9 days [30]; the infectious period , 1/γ , lasted for 4–7 days in clinical descriptions of 297 PCR-confirmed cases in French Polynesia [22]; the extrinsic latent period has been estimated at 10 . 5 days [1]; and mosquito lifespan in Tahiti was estimated at 7 . 8 days [31] . We therefore used these values for the respective means of 1/αH , 1/γ , 1/αv and 1/δ in our prior distributions . These parameters were estimated jointly across all six regions; as mentioned above , we assumed that the parameters remained fixed over time , as temperature and rainfall levels did not change substantially during the outbreak . The rest of the parameters were estimated for each region individually; we assumed uniform prior distributions for these . Serological analysis of samples from blood donors between July 2011 and October 2013 suggested that only 0 . 8% of the population of French Polynesia were seropositive to ZIKV [33]; we therefore assumed that the population was fully susceptible initially . We also assumed that the initial number of latent and infectious people were equal ( i . e . E 0 H = I 0 H ) , and the same for mosquitoes ( E 0 V = I 0 V ) . The basic reproduction number was equal to the product of the average number of mosquitoes infected by the typical infectious human , and vice versa [24]: R 0 = β V γ × α V δ + α V β H δ . ( 9 ) We fitted the model using Markov chain Monte Carlo ( MCMC ) , where incidence in week t , denoted ct , was the difference in the cumulative proportion of cases over the previous week i . e . ct = C ( t ) − C ( t − 1 ) . In the model , the total number of cases included asymptomatic and subclinical cases—which would not be detected at any site—as well as those that displayed symptoms . Hence there were two sources of potential underreporting: as a result of limited sentinel sites; and as a result of cases not seeking treatment . We adjusted for the first source of underreporting by defining κt as the proportion of total sites that reported as sentinels in week t . We assumed that the population was uniformly distributed across the catchment areas of the healthcare sites . Under this assumption , the proportion of total sites that reported cases as sentinels in a particular week , κt , was equivalent to the expected fraction of new cases that would be reported in that week if the reporting proportion , r , was equal to 1 . The parameter r accounted for the second source of under-reporting , and represented the proportion of cases ( both symptomatic and asymptomatic ) that did not seek treatment . To calculate the likelihood of observing a particular number of cases in week t , yt , we assumed the number of confirmed and suspected cases in week t followed a negative binomial distribution with mean rκt ct and dispersion parameter ϕ , to account for potential variability in reporting over time [34] . The dispersion parameter reflected variation in the overall proportion reported , as well as potential variation in size and catchment area of sentinel sites . Hence the log-likelihood for parameter set θ given data Y = { y t } t = 1 T was L ( θ|Y ) = ∑t log P ( yt|ct ) . As a sensitivity analysis ( see Results ) , we also extended the model so the likelihood included the probability of observing 314/476 seropositive individuals in Tahiti after the outbreak , given that a proportion Z were infected in the model . Hence for Tahiti , L ( θ|Y ) = ∑t log P ( yt|ct ) + log P ( X = 314 ) , where X ∼ B ( n = 476 , p = Z ) . The joint posterior distribution of the parameters was obtained from eight replicates of 25 , 000 MCMC iterations , each with a burn-in period of 5 , 000 iterations ( S2–S8 Figs ) . The model was implemented in R version 3 . 2 . 3 [35] using the deSolve package [36] . We implemented a simple demographic model to examine the replacement of the number of susceptible individuals over time . In 2014 , French Polynesia had a birth rate of b = 15 . 47 births/1 , 000 population , a death rate of d = 4 . 93 deaths/1 , 000 population , and net migration rate of m = −0 . 87 migrants/1 , 000 [37] . The number of susceptible individuals in year τ , S ( τ ) , and total population size , N ( τ ) , was therefore expressed as the following discrete process: N ( τ ) = N ( τ - 1 ) + b N ( τ - 1 ) - d N ( τ - 1 ) - m N ( τ - 1 ) ( 10 ) S ( τ ) = S ( τ - 1 ) + b N ( τ - 1 ) - d S ( τ - 1 ) - m S ( τ - 1 ) ( 11 ) We set S ( 2014 ) as the fraction of the population remaining in the S compartment at the end of the 2013–14 ZIKV outbreak , and propagated the model forward to estimate susceptibility in future years . The effective reproduction number , Reff ( τ ) , in year τ was the product of the estimated basic reproduction number , and the proportion of the population susceptible: Reff ( τ ) = R0S ( τ ) . We sampled 5 , 000 values from the estimated joint posterior distributions of S ( 2014 ) and R0 to obtain the median and credible intervals .
Across the six regions , estimates for the basic reproduction number , R0 , ranged from 2 . 6 ( 95% CI: 1 . 7–5 . 3 ) in Marquises to 4 . 8 ( 95% CI: 3 . 2–8 . 4 ) in Moorea ( Table 3 ) . Our results suggest that only a small proportion of ZIKV infections were reported as suspected cases: sampling from the fitted negative binomial reporting distributions for each region implied that 11 . 5% ( 95% CI: 7 . 32–17 . 9% ) of infections were reported overall . Estimated dispersion in reporting was greatest for Marquises ( S1 Table ) , reflecting the variability in the observed data ( Fig 2 ) , even after adjusting for variation in the number of sentinel sites . Dividing the 8 , 744 cases reported at sentinel sites by the total estimated infections , we also estimated that 3 . 41% ( 95% CI: 3 . 32–3 . 55% ) of total infections were reported at the subset of health sites that acted as sentinel sites . Our posterior estimates for the latent and infectious periods in humans and mosquitoes were consistent with the assumed prior distributions ( S2 Fig ) , suggesting either that there was no strong evidence that these parameters had a different distribution , or that the model had limited ability to identify these parameters from the available data . As a sensitivity analysis , we therefore considered two alternative prior distributions for the incubation and infectious periods for humans and mosquitoes . First , we examined a broader prior distribution . We used the same mean values for the Gamma distributions specified in Table 2 , but with σ = 2 . These priors produced similar estimates for R0 , proportion reported , and total number of infections ( S2 Table ) , although the estimated parameters for humans were further from zero than in the prior distribution ( S9 Fig ) . As a second sensitivity analysis , we used prior distributions with mean values as given in studies of dengue fever , and σ = 0 . 5 . As there is evidence that human-to-mosquito transmission can occur up to 2 days before symptom onset [38] , and the intrinsic incubation period for DENV infection is 5 . 9 days [39] , we assumed a mean latent period of 5 . 9–2 = 3 . 9 days . We also assumed an infectious period of 5 days [38]; an extrinsic latent period of 15 days [39]; and a longer mosquito lifespan of 14 days [28] . Again , these assumptions produced similar estimates for key epidemiological parameters ( S3 Table ) , with posterior estimates for incubation and infectious periods tracking the prior distributions ( S10 Fig ) . The estimated proportion of the population that were infected during the outbreak ( including both reported and unreported cases ) was above 85% for all six regions ( Table 3 ) , and we estimated that 94% ( 95% CI: 91–97% ) of the total population were infected during the outbreak . A serological survey following the French Polynesia ZIKV outbreak found 314/476 children aged 6–16 years in Tahiti were positive for ZIKV in an indirect ELISA test for IgG antibody , corresponding to an attack rate of 66% ( 95% CI: 62–70 ) [17] . To test whether this seroprevalence data could provide additional information about the model parameters , we extended the model to calculate the likelihood of observing 314/476 seropositive individuals in Tahiti after the outbreak , as well as the observed weekly case reports . We obtained a much lower R0 estimate for Tahiti , but similar results for other regions , and the median reporting rate remained unchanged for all areas ( S4 Table ) . However , the model was unable to reproduce the Tahiti epidemic curve when the overall attack rate was constrained to be consistent with the results of the serological survey ( S11 Fig ) . During the 2013–14 outbreak in French Polynesia , there were 42 reported cases of GBS [13] . This corresponds to an incidence rate of 15 . 3 ( 95% binomial CI: 11 . 0–20 . 7 ) cases per 100 , 000 population , whereas the established annual rate for GBS is 1–2 cases per 100 , 000 [10] . In total , there were 8 , 744 confirmed and suspected ZIKV cases reported at sentinel sites in French Polynesia , which gives an incidence rate of 480 ( 95% CI: 346–648 ) GBS cases per 100 , 000 suspected Zika cases reported at these sites . However , when we calculated the GBS incidence rate per estimated total ZIKV cases , using the model estimates based on the prior distributions in Table 2 , we obtained a rate of 16 . 4 ( 95% CI: 11 . 5–21 . 4 ) per 100 , 000 cases . These credible intervals overlap substantially with the above incidence rate calculated with population size as the denominator , indicating that the two rates are not significantly different . Using a demographic model we also estimated the potential for ZIKV to cause a future outbreak in French Polynesia . We combined our estimate of the proportion of the population that remained susceptible after the 2013–14 outbreak and R0 with a birth-death-migration model to estimate the effective reproduction number , Reff , of ZIKV in future years . If Reff is greater than one , an epidemic would be possible in that location . Assuming that ZIKV infection confers lifelong immunity against infection with ZIKV , our results suggest that it would likely take 12–20 years for the susceptible pool in French Polynesia to be sufficiently replenished for another outbreak to occur ( Fig 3 ) . This is remarkably similar to the characteristic dynamics of DENV in the Pacific island countries and territories , with each of the four DENV serotypes re-emerging in sequence every 12–15 years , likely as a result of the gradual accumulation of susceptible individuals due to births [19 , 40] .
Using a mathematical model of ZIKV transmission , we analysed the dynamics of infection during the 2013–14 outbreak in French Polynesia . In particular , we estimated key epidemiological parameters , such as the basic reproduction number , R0 , and the proportion of infections that were reported . Across the six regions , our median estimates suggest that between 7–17% of infections were reported as suspected cases . This does not necessarily mean that the non-reported cases were asymptomatic; individuals may have had mild symptoms and hence did not enter the healthcare system . For example , although the attack rate for suspected ZIKV disease cases was 2 . 5% in the 2007 Yap ZIKV outbreak , a household study following the outbreak found that around 19% of individuals who were seropositive to ZIKV had experienced ZIKV disease-like symptoms during the outbreak period [2] . Our median estimates for R0 ranged from 2 . 6–4 . 8 across the six main archipelagos of French Polynesia , and as a result the median estimates of the proportion of the populations that became infected in our model spanned 87–97% . This is more than the 66% ( 95% CI: 62–70% ) of individuals who were found to be seropositive to ZIKV in a post-outbreak study in Tahiti [17] . When we constrained the model to reproduce this level of seroprevalence as well as the observed weekly reports , however , we obtained a much poorer fit to the case time series ( S11 Fig ) . The discrepancy may be the result of population structure , which we did not include within each region; we used a homogeneous mixing model , in which all individuals had equal chance of contact . In reality , there will be spatial heterogeneity in transmission [41] , potentially leading to a depletion of the susceptible human pool in some areas but not in others . Additionally , there is evidence that Ae . aegypti biting rate can vary between individual human hosts [42] . Whereas in the model everyone in regions with ZIKV infected mosquitoes had equal probability of infection , in reality there is likely to be individual-level heterogeneity in probability of infection , which could alter the proportion who seroconvert to ZIKV after the outbreak . As we used a deterministic model , differences in the estimate for the reporting dispersion parameter for different regions may to some extent reflect the limitations of the model in capturing observed transmission patterns , as well as true variability in reporting . The ZIKV outbreak in French Polynesia coincided with a significant increase in Guillain-Barré syndrome ( GBS ) incidence [13] . We found that although there was a raw incidence rate of 480 ( 95% CI: 346–648 ) GBS cases per 100 , 000 suspected ZIKV cases reported , the majority of the population was likely to have been infected during the outbreak , and therefore the rate per infected person was similar to the overall rate per capita . This could have implications for the design of epidemiological studies to examine the association between ZIKV infection and neurological complications in island populations . If infection with ZIKV confers lifelong immunity , we found it would take at least a decade before re-invasion were possible . In the Pacific island countries and territories , replacement of DENV serotypes occurs every 4–5 years [19 , 40] , and therefore each specific serotype re-emerges in a 12–15 year cycle . The similarity of this timescale to our results suggest that ZIKV may exhibit very similar dynamics to DENV in island populations , causing infrequent , explosive outbreaks with a high proportion of the population becoming infected . In September 2014 , Chikungunya virus ( CHIKV ) caused a large outbreak in French Polynesia [43] , and is another example of a self-limiting arbovirus epidemic in island populations [5] . However , it remains unclear whether ZIKV could become established as an endemic disease in larger populations , as DENV and CHIKV have . For immunising infectious diseases , there is typically a ‘critical community size’ , below which random effects frequently lead to disease extinction , and endemic transmission cannot be sustained [18 , 44] . Analysis of dengue fever outbreaks in Peru from 1994–2006 found that in populations of more than 500 , 000 people , dengue was reported in at least 70% of weekly records [41] . Large cities could have the potential to sustain other arboviruses too , and understanding which factors—from population to climate—influence whether ZIKV transmission can become endemic will be an important topic for future research . We did not consider seasonal variation in transmission as a result of climate factors in our analysis , because all six outbreaks ended before there was a substantial seasonal change in rainfall or temperature . Such changes could influence the extrinsic incubation period and mortality of mosquitoes , and hence disease transmission . If the outbreaks had ended as a result of seasonality , rather than depletion of susceptibles , it would reduce the estimated proportion of the population infected , and shorten the time interval before ZIKV would be expected to re-emerge . There are some additional limitations to our analysis . As we were only fitting to a single time series for each region , we also assumed prior distributions for the incubation and infectious periods in humans and mosquitoes . Sensitivity analysis on these prior distributions suggested it was not possible to fully identify these parameters from the available data . If seroprevalence data from each region were to become available in the future , it could provide an indication of how many people were infected , which may make it possible to constrain more of the model parameters , and evaluate the role of spatial heterogeneity discussed above . Such studies may require careful interpretation , though , because antibodies may cross-react between different flaviviruses [12] . Our results suggest that ZIKV transmission in island populations may follow similar patterns to DENV , generating large , sporadic outbreaks with a high degree of under-reporting . If a substantial proportion of such populations become infected during an outbreak , it may take several years for the infection to re-emerge in the same location . A high level of infection , combined with rarity of outbreaks , could also make it more challenging to investigate a potential causal link between infection and concurrent neurological complications . | Since the first reported major outbreak of Zika virus disease in Micronesia in 2007 , the virus has caused outbreaks throughout the Pacific and South America . Transmitted by the Aedes genus of mosquitoes , the virus has been linked to possible neurological complications including Guillain-Barré Syndrome and microcephaly . To improve our understanding of the transmission dynamics of Zika virus in island populations , we analysed the 2013–14 outbreak on the six major archipelagos of French Polynesia . We found evidence that Zika virus infected the majority of the population , but only around 12% of total infections on the archipelagos were reported as cases . If infection with Zika virus generates lifelong immunity , we estimate that it would take at least 12–20 years before there are enough susceptible people for the virus to re-emerge . Our results suggest that Zika virus could exhibit similar dynamics to dengue virus in the Pacific , producing large but sporadic outbreaks in small island populations . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"invertebrates",
"dengue",
"virus",
"medicine",
"and",
"health",
"sciences",
"microcephaly",
"pathology",
"and",
"laboratory",
"medicine",
"demography",
"pathogens",
"geographical",
"locations",
"microbiology",
"vector-borne",
"diseases",
"mathematical",
"models",
"animals"... | 2016 | Transmission Dynamics of Zika Virus in Island Populations: A Modelling Analysis of the 2013–14 French Polynesia Outbreak |
Fish-borne zoonotic trematodes ( FZT ) infections including liver- and minute intestinal flukes are common in Southeast Asia in both humans and domestic animals eating raw fish and since 2010 , the liver flukes are recognised as neglected tropical diseases by WHO . Mass drug treatment with praziquantel is advised for humans , but no recommendations for control of the FZT in the reservoir hosts exist . A study was conducted to assess the ability of praziquantel treatment for control of FZT in farm dogs in an endemic area in Northern Vietnam . Initially , 101 dogs from 73 households were examined for small trematode eggs in their faeces . Forty seven copro-positive dogs were included in the study . Thirty eight dogs received treatment with a single dose of 40 mg/kg of praziquantel . A group of nine dogs were left untreated . Coprological examination for small trematode eggs was performed on day 0 , 3 , 10 , 30 , 60 , 90 and 120 post treatment . Farmers were questioned about dog keeping practises . All dogs were copro-negative for small trematode eggs on both day 3 and 10 post treatment . From day 30 onwards previously negative dogs became positive again . The reinfection rates were 26 . 3% ( day 30 ) , 45 . 5% ( day 60 ) , 53 . 1% ( day 90 ) , 61 . 3% ( day 120 ) . The nine untreated dogs remained positive throughout the study period . There was no difference in the intensity of infection at day 0 and 120 neither in the treated or untreated group . Dogs had easy access to raw fish and did not receive treatment against flukes by their owner . More than 50% of the dogs were reinfected 3 months post treatment . We do not recommend controlling FZT infections in dogs by anthelmintic treatment alone since reinfection occurs fast under the existing farm management systems .
The fish-borne zoonotic trematodes ( FZT ) in South-east Asia comprise of the liver flukes Clonorchis sinensis and Opisthorchis vivirrini and a large group of minute intestinal flukes ( MIF ) with more than 35 species mainly belonging to the family Heterophyidae [1] . Humans and fish-eating animals acquire these zoonotic infections by consumption of raw or undercooked fish . It is estimated that 40 to 50 million people are currently infected with one or several species of FZT [2] . The number of infected animals remains unknown . Mixed infections with MIF and liver flukes are often found both in Thailand , Laos and Vietnam [3]–[5] . However , as the eggs of MIF and liver flukes are indistinguishable by light microscopy , studies based on coprological examinations should not report prevalence of a particular species [6] . E . g . a study performed in Northern Vietnam claimed a prevalence of C . sinensis of 26% based on faecal examination [7] . However , examinations in fish in the same area showed a prevalence of only 1 . 5% for C . sinensis and 55 . 6% for the MIF , Haplorchis pumilio [8] . Furthermore , several other studies performed in Northern Vietnam showed that MIF are not only prevalent in humans [4] , but also in reservoir hosts [9] , [10]: Dogs and cats hosted a range of 12 different FZT species , with prevalence of 35–70% in the Nam Dinh and Nghe An provinces . C . sinensis was also found but only in very few individuals . Moreover , these studies documented the importance of the reservoir host , especially dogs , in maintaining transmission of the flukes by contributing substantially to contamination of the environment with eggs [9] , [10] . Besides preventive chemotherapy in humans , a part of WHO's strategy to overcome neglected tropical diseases is to focus on veterinary public health [11] . Indeed , given the importance of the reservoir hosts in contamination of the environment , FZT cannot be controlled by treatment of humans alone . Therefore , this study was conducted to provide evidence-based information for recommendations for treatment and control of FZT in domestic animals . For humans the recommended dose is 3×25 mg praziquantel per kg body weight , or a single dose of 40 mg/kg for mass drug treatment [12] . Since a single dose would be the most applicable treatment for free roaming farm dogs it was chosen in the present study . Based on the following knowledge , we suspected that reinfection with FZT would occur fast when dogs were constantly exposed to a diet including raw , infected fish . The pre-patent period for C . sinensis is approximately 30 days [13] whereas it is just 9 days for the prevalent MIF , H . pumilio [14] . A previous field trial with praziquantel also gave indications of dogs becoming copro-positive for small trematode eggs 30 days after treatment [15] . The specific objective was to evaluate the ability of a single praziquantel treatment to control FZT infection in farm dogs in Northern Vietnam . We determined the effectiveness of drug treatment on day 3 and 10 post treatment , the time until reinfection occurred as well as the intensity of infections before and after treatment . Fluctuations in egg excretion in a group of untreated dogs were also monitored and information on the practice of keeping dogs on farms was obtained through a questionnaire to describe the dog's exposure to raw fish .
The study was carried out from July–December 2011 in Nghia Lac commune in the Nam Dinh province in Northern Vietnam , an area endemic for FZT , where aquaculture is widespread . Initially , faecal samples from 101 randomly selected dogs from 73 farming households were examined for presence of FZT . Dogs were regarded positive for FZT when finding ‘small trematode eggs’ , a term commonly used for FZT egg being shorter than 50 µm [4] , [16] , in the faeces . The study was designed as a randomised intervention study . Forty-seven dogs yielding faecal samples positive for small trematode eggs were chosen by convenience and assorted into either treated or untreated groups . The treated group consisted of 27 young ( <1 year ) and 11 older dogs ( ≥1 year ) , whereas the untreated group consisted of 4 young and 5 older dogs , respectively . Thirty-eight dogs were treated and nine dogs were kept as controls . Dogs were weighed using a scale and treated with praziquantel tablets ( Distocide , ShinPoong Pharma Co . , Ltd . , Seoul , South Korea ) at a single dose of 40 mg/kg body weight at day 0 and faecal samples were examined at day 0 , 3 , 10 , 30 , 60 , 90 and 120 post treatment ( untreated dogs were not examined on day 3 and 10 ) . Dog in the treated group were regarded as cured , when no intact eggs were found in their faecal samples on day 3 and 10 post infection and were regarded as reinfected , when intact small trematode eggs were found in their faeces . Dogs younger than 2 months of age , pregnant bitches , dogs showing clinical signs of disease and vicious dogs were excluded from the study . The animal protocol followed the EU-guidelines for animal experiments [17] . An ethical review of the study was conducted by the project management of FIBOZOPA project in Vietnam , which provided the funds and by the National Institute of Veterinary Research . These two institutions approved the animal protocol , since no official governmental animal ethics committee exist in Vietnam . Informed , oral consent was obtained from the farmers . Dog were treated at the end of the study . A member of each household visited for the initial screening for FZT-positive dogs was interviewed ( in majority of households the person feeding the dogs ) to obtain information about the household's general conditions and practices of keeping dogs , including the feeding of raw fish . Information included presence of fish ponds in household or neighbouring household; total number of dogs in the household; roaming behaviour of dogs; and feeding practices , including whether raw fish or leftover meals with fish were fed to dogs . Farmers were also asked about anthelmintic treatment of their dogs during the last month , how often treatment was performed and finally the type of drug used . Information about the 47 dogs in the treatment study included: sex , age , breed , and body condition score ranging from 1 being gaunt , 3 being ideal and 5 being obese [18] was gathered by trained personal . The body weight was measured at the day of treatment ( day 0 ) using a scale and the dogs were assigned an ID number . Faecal samples were collected rectally from the dogs and analysed for FZT eggs by a modification of the method by Willingham et al . [19] . Five gram of faeces was dissolved to make a 100 ml 0 . 9% saline solution . The sample was washed with saline through three sieves: 400 , 100 and 45 µm mesh size . Between sieving the samples were left to sediment ( 15–20 min ) in a 250 ml conical beaker . After washing through the 45 µm sieve the sediment in the conical beaker was collected in a 15 ml centrifuge tube and centrifuged for 1 min at 100 g . The sediment was mixed with saline to form a 2 . 25 ml suspension . Small trematode eggs were counted in a total of 0 . 45 ml ( dilution of the sample was often necessary ) of each sample ( 100× magnification ) . This volume represented 1 g of faeces , and the counts were expressed as eggs per g of faeces ( epg ) . In some cases less than five grams of faeces were obtained . The epg was then calculated as: Egg count×2 . 25 ( ml ) /amount of faeces ( g ) ×0 . 45 ( ml ) . The actual reinfection rates were calculated on day 30 , 60 , 90 and 120 as the cumulative number of dogs being copro-positive for FZT on the respective day divided by the number of dogs examined on the actual day . The Kaplan-Meier product limit method [20] , commonly used in survival analysis to model the time to an event , was used to calculate the probability of being reinfected at a certain time point given that the dog stayed in the study up to that point . The model took into account information on dogs leaving the study before termination of the study due to various reasons and dogs still uninfected by termination of the study , through point censoring [20] . The median time till reinfection ( the time when 50% of the dogs were expected to be reinfected ) was estimated and difference between young ( <1 year ) and older dogs ( ≥1 year ) was tested using the Kaplan-Meier model . The arithmetic mean of the intensities of infection among FZT-positive dogs on day 0 and 120 were compared within and between the treated group and likewise for the untreated group by Mann Whitney test ( a non-parametric test had to be used since data was not normally distributed ) . Intensities of infection day 0 were also compared between young and older dogs using the same test as above . The effectiveness of the treatment was determined as number of treated dogs negative for undamaged eggs both on day 3 and 10 post treatment divided by the total number of treated dogs , multiplied by 100 .
A total of 73 farming households were visited for questionnaire interviews during the initial screening . The majority ( 86% ) of the households had a fish pond in their close proximity ( them self or neighbours ) . The households kept between 1 and 7 dogs , the majority having 1 or 2 dogs ( 38 and 36% , respectively ) . More than 80% of the households described that their dogs sometimes or always roamed freely . Forty-one percent reported that their dogs ate raw fish occasionally and 52% said the dogs ate intestines and other body parts when fish were prepared for household members . Other types of dog feed mentioned were rice , kitchen left overs and different types of cooked food . A typical situation with cleaning and preparation of fish for human consumption is illustrated in Fig . 1 . Only 6 households had dewormed their dogs within the last month , 64% never dewormed their dogs , 21% dewormed the dogs once when they were young and a single farmer gave anthelmintic treatment twice per year . Most farmers did not remember which drug was used for treatment , though four farmers mentioned levamisol and a single farmer tetramisol as being the drug of choice . No drugs active against trematodes were mentioned . Twenty six female dogs and 18 males were included in the study ( in three cases the sex was not recorded ) . The age of dogs <1 year ranged between 2–8 months ( mean 2 . 4 ) , dogs ≥1 year ranged between 1–7 years ( mean 2 . 6 ) . The body condition score ( mean ± SD ) was 2 . 4±0 . 5 meaning that majority of dogs were somewhat thinner than the ideal measure . All dogs were of mixed breed . This low number of older dogs in the treated group was due to some of the initially selected dogs being unavailable and selection of other dogs instead . The initial examination found 77/101 ( 76% ) dogs positive for small trematode eggs in their faeces . Of the 38 dogs treated with praziquantel , 7 dogs dropped out prior to termination of the study , 12 remained uninfected and 19 became reinfected within the 120 days . The actual reinfection rates were 26 . 3% ( day 30 ) , 45 . 5% ( day 60 ) , 53 . 1% ( day 90 ) , 61 . 3% ( day 120 ) . This corresponded to 10 dogs becoming reinfected before day 30 , additional 5 dogs between 30–60 days , 2 dogs between 60–90 and , finally , 2 dogs between 90–120 days post treatment . The reinfection rates and the Kaplan-Meier estimates can be seen in Fig . 2 . The median time till reinfection was 105 days , ( 95% CI 60–120 ) . The time till reinfection did not differ significantly between young and older dogs . Nineteen dogs were point censored in the study , of these , 7 dogs had left the study prior to the last sampling ( day 30: 0 , day 60: 5 , day 90: 1 , day 120: 1 dogs , respectively ) because the dogs were sold or died . At the day of praziquantel treatment , the intensity of FZT infection in the dogs was 65±198 epg ( mean ± SD ) in the treated group and 4±3 epg in the untreated group , which differed significantly ( P<0 . 05 ) . No significant difference was seen in the intensity of infection between day 0 and 120 in any of the groups probably due to a large variation in egg counts . The fluctuations in intensity of infection and reinfection can be seen in Fig . 3 . The effectiveness of the treatment was 100% . However , three days after treatment , few damaged eggs ( 1–4 ) with black spots or empty egg shells without operculum were found in three treated dogs and a similar findings was seen in three dogs at day 10 ( one dog had such eggs both days ) . These damaged eggs were believed to be reminiscence of eggs trapped in the mucosa from the first infection and not a result of treatment failure . Two of these dogs showed , undamaged eggs at day 30 and one at day 60 . The remaining two dogs dropped out of the study on day 60 and 120 , respectively , without being reinfected . The untreated dogs remained positive throughout the study period .
The present study documented that reinfection of dogs with FZT took place within 30 days after praziquantel treatment . More than 25% of the dogs were reinfected day 30 and more than 50% at day 90 post treatment , respectively . The intensity of FZT infection four months after drug treatment was similar to the intensity found before treatment took place . Infected dogs kept excreting eggs for 4 months . In general , dogs were often exposed to raw fish; more than 50% through the feed and more than 80% roamed freely thereby having access to fish ponds . These practices are risk factors for infections [9] , [10] . The initial prevalence found in this study showed that FZT infections are common in the dog population in this area as demonstrated in a study four years earlier ( 76% vs . 57% found in 2007 ) and the intensity of infection found initially in the treated dogs was higher than this earlier study [9] . Based on the Kaplan-Meier estimate more than 25% of treated dogs were reinfected one month after treatment and half of the treated dogs would be reinfected 105 days after treatment , a slightly more conservative estimate than the actual reinfection rates , due to the inclusion of censored dogs in the analysis . The results suggest that even if dogs were treated with praziquantel regularly , the likelihood of becoming reinfected soon after treatment would be very high , and thus continued contamination the environment with eggs would take place . The lack of apparent influence of the praziquantel treatment on the intensity of the reinfection is also important since it suggests that no immunity , or immunity of limited strength , is acquired following treatment . Twelve dogs ( 32% ) remained uninfected at day 120 perhaps due to lack of , or infrequent , exposure to raw fish . The present findings are also alarming from a human treatment perspective: If reinfections occur within months after treatment , obviously more frequent treatments would be needed to control the FZT . This again would lead to higher risk for development of resistance against praziquantel . Monitoring mass drug treatment programs in humans , which primarily aim at targeting the liver flukes , are further complicated by the presence of MIF egg which cannot be distinguished from the liver fluke eggs [6] . Hence proper diagnostic tools to differentiate between liver fluke and MIF eggs are urgently needed . For dogs , several factors favour reinfection: Fish ponds are common in the villages and most dogs roam freely with access to ponds and fish . A previous study in the same area found that the main risk factor for having infection with FZT was if the dogs were fed raw fish [9] . In the present study more than 50% of the households gave their dogs access to raw fish left over . Most of the dog in the current study was less than one year old . This is a result of the farm practice in the area , where many dogs are kept for meat , hence are sold for consumption at around one year of age . Despite the skewed age profile , no significant difference was found in reinfection rate or intensity of infection comparing dogs younger and older than one year . When calculating the effectiveness a criterion of negativity for intact eggs on day 3 and 10 post infection was used . Hence , the 5 dogs excreting damaged eggs were regarded cured of the FZT infection and the effectiveness of the treatment in the field to be 100% . In a previous study in dogs and cats where two different doses of praziquantel were evaluated , two cats receiving 75 mg/kg also showed damaged eggs 3 days post treatments that were judged non-viable [15] . We therefore reason that our assumption is valid . The finding of damaged eggs could be due to worms or eggs being trapped in the mucosa and therefore being expelled over some days . Excretion of non-viable eggs for days after treatment is known to occur e . g . with Schistosoma japonicum [21] , [22] . Earlier studies have assessed the cure rate of a single praziquantel dose of 40 mg/kg against opisthorchiasis in humans and have found it to be 91 and 95 . 5% [23] , [24] . No strict rules apply for defining effectiveness and based on faecal examinations alone , no true cure rate can be obtained . Future studies should evaluate the true cure rate of praziquantel on MIF by doing necropsies and worm recovery after treatment . A recent intervention study for reducing prevalence and intensity of fish-borne zoonotic trematode infections in The Red River Delta in Vietnam , confirmed the importance of including dogs and cats in the control programs [25] . However , the authors also pointed out , that the practicality of control programs must be considered , and in this regard , treatment of domestic animals and humans is costly and difficult to undertake . Combined with the knowledge gained in the present study about the high reinfection rate and the existing treatment practise of dogs for helminths , recommending farmers to treat their dogs for trematodes with frequent intervals is not a sustainable way to overcome infections with FZT . Awareness about the need for an integrated approach towards control of FZT is rising e . g . as described by Sithithaworn et al . [26] who suggest a community-orientated approach including treatment , improved sanitation and information , education and communication . Our recommendation is to include teaching about feeding practise of the dogs and other domestic animals in such a program to avoid reoccurring reinfections to take place . To conclude , dogs have easy access to raw fish and do not receive treatment against flukes by their owner . After a single treatment more than 50% of the dogs were reinfected after 90 days . The effectiveness of the single dose of praziquantel was found to be 100% . Treatment of individual dogs with symptoms of FZT infection is of course always recommendable , however , repeated mass treatments are hardly applicable and cannot stand alone for control of FZT in dogs . | Fish-borne zoonotic trematodes ( FZT ) including liver and small intestinal flukes in humans and domestic animals are a growing problem in Southeast Asia . WHO has recently recognized the problem , listed the liver flukes as neglected tropical diseases and recommends mass drug treatment of humans to overcome the diseases . Domestic animals including dogs , cats and pigs contribute to transmission and the infection must be controlled in these animals if the transmission cycle shall be broken . We studied the effect of a single praziquantel treatment in a group of natural infected dogs in an FZT endemic area in Northern Vietnam . The dogs were followed over 4 months after treatment and management practices were described . We found that dogs had easy access to raw fish and did not receive treatment against flukes by their owners . More than fifty per cent of the dogs became reinfected with FZT within 3 months after treatment . We conclude that repeated mass treatments are not applicable and cannot stand alone for control of FZT in dogs . The reinfection pattern in humans treated for FZT needs to be investigated to monitor effect of mass drug treatments and avoid development of drug resistance . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2014 | Reinfection of Dogs with Fish-Borne Zoonotic Trematodes in Northern Vietnam following a Single Treatment with Praziquantel |
Proteins perform their function or interact with partners by exchanging between conformational substates on a wide range of spatiotemporal scales . Structurally characterizing these exchanges is challenging , both experimentally and computationally . Large , diffusional motions are often on timescales that are difficult to access with molecular dynamics simulations , especially for large proteins and their complexes . The low frequency modes of normal mode analysis ( NMA ) report on molecular fluctuations associated with biological activity . However , NMA is limited to a second order expansion about a minimum of the potential energy function , which limits opportunities to observe diffusional motions . By contrast , kino-geometric conformational sampling ( KGS ) permits large perturbations while maintaining the exact geometry of explicit conformational constraints , such as hydrogen bonds . Here , we extend KGS and show that a conformational ensemble of the α subunit Gαs of heterotrimeric stimulatory protein Gs exhibits structural features implicated in its activation pathway . Activation of protein Gs by G protein-coupled receptors ( GPCRs ) is associated with GDP release and large conformational changes of its α-helical domain . Our method reveals a coupled α-helical domain opening motion while , simultaneously , Gαs helix α5 samples an activated conformation . These motions are moderated in the activated state . The motion centers on a dynamic hub near the nucleotide-binding site of Gαs , and radiates to helix α4 . We find that comparative NMA-based ensembles underestimate the amplitudes of the motion . Additionally , the ensembles fall short in predicting the accepted direction of the full activation pathway . Taken together , our findings suggest that nullspace sampling with explicit , holonomic constraints yields ensembles that illuminate molecular mechanisms involved in GDP release and protein Gs activation , and further establish conformational coupling between key structural elements of Gαs .
G protein-coupled receptors ( GPCRs ) mediate a large variety of physiological events throughout the body by activating intracellular signal transduction pathways [1] . Signaling molecules , such as hormones and neurotransmitters , can induce conformational changes in GPCRs , which allow it to complex with intracellular protein partners such as heterotrimeric guanine nucleotide-binding protein G . β2 Adrenergic Receptor ( β2AR ) , a so-called class A receptor , initiates activation of stimulatory protein Gs mainly through interactions with the latter’s α subunit ( Gαs ) . Upon activation , Gs interacts with effector proteins in the cell which , ultimately , leads to a cellular response . However , a precise characterization of the activation mechanism of Gs has remained elusive [2] . Molecular dynamics ( MD ) simulations can structurally characterize the dynamics of biomolecules in great detail [3] . However , as increasingly sophisticated experimental techniques yield ever bigger molecular systems and complexes , the computational demands to ensure adequate sampling of the conformational landscape often require highly specialized hardware and algorithms [4] . In parallel , time-independent or non-deterministic sampling-based algorithms together with simplified macromolecular representations have also led to tremendous insights . Conformational sampling with CONCOORD has provided seeds for subsequent MD simulations to overcome large energy barriers in the characterization of recognition dynamics of ubiquitin [5 , 6] . Rapid exploration of conformational space in internal coordinates with a traditional mechanical force field via a biased Monte Carlo approach [7] accurately predicted agonist binding modes for GPCRs [8] . Exhaustive sampling has predicted ensembles of low-energy conformers for GPCRs associated with ligand binding and activation [9] . Rosetta-based sampling and energy analysis provided a structural basis for rhodopsin-mediated GDP release from Gi , a inhibitory protein highly related to Gs [10] . Vibrational modes of a biomolecule are well-approximated with a so-called Elastic Network Model ( ENM ) , in which non-bonded interactions are replaced with a harmonic pseudo-potential [11] . Analysis of ENMs with NMA , which relies on a Hamiltonian in which the kinetic energy is also quadratic , yields the equations of motion around a minimum of the potential energy of the system . While low-frequency modes are generally associated with biological activity , the second order approximation underlying NMA limits its ability to access conformational substates and observe larger , diffusional motions . Nonetheless , NMAs are enormously successful and have , for instance , proposed GPCR activation mechanisms [12] . Combined with short MD trajectories NMA also predicted a molecular mechanism for GDP release from Gi [13] . Kinogeometric sampling ( KGS ) treats a biomolecule as a branched polymer , with rotatable bonds as degrees of freedom ( DoFs ) and non-covalent ( hydrogen ) bonds as distance constraints [14–16] . Hydrogen bonds define nested , closed loops that require coordinated changes of DoFs to avoid breaking the bonds . Kinogeometric sampling maps structural perturbations onto a subspace of conformation space that accounts for the reduced flexibility of these closed loops . This procedure intrinsically favors certain directions on the conformational landscape , namely those that avoid , collectively , native hydrogen bond dissociation . Additionally , representing biomolecular systems with fewer DoFs enables better exploration of conformation space and , ultimately , allows fitting sparse experimental data sets while reducing the risk of overfitting . Distance constraints from hydrogen bonds can completely rigidify substructures of biomolecules . For instance , an α-helix is often rigidified owing to its backbone hydrogen bonding network . Kinogeometric and similar sampling-with-constraint techniques have relied on combinatorial constraint counting to explicitly identify rigid substructures in the molecule that result from the hydrogen bonds [17] . Perturbing a molecular conformation with constraints generally required breaking constraints and subsequently reclosing them [18] , or iteratively refitting the perturbed conformation and the rigid substructures [19] . Here , we extend our kinogeometric computational techniques by abandoning explicit constraint counting to proteins . Our procedure efficiently samples conformational degrees of freedom in a lower-dimensional subspace in which instantaneous distance constraints are preserved exactly [20] . The advantage of our method is that a single , exact mathematical analysis both provides constraint satisfaction and infinitesimal , coordinated directions of motion for the degrees of freedom of the protein . It naturally couples motions throughout the protein by many interconnected and interdependent cycles , making few additional assumptions on interactions . As a result , collective motions emerge which deform the protein along preferred dimensions . We apply our algorithm to compute a broad conformational distribution of the inactive and active states of the α subunit of free ( i . e . not receptor-bound ) , apo ( i . e . nucleotide-free ) Gαs . We demonstrate that our ensemble identifies detailed molecular mechanisms implicated in domain opening and activation of protein Gs . We compare the findings to an ensemble obtained with a state-of-the art torsional ENM . An ENM representation with torsional degrees of freedom is conceptually similar to our approach , and is known to better represent protein conformational changes than Cartesian ENMs [21 , 22] . We selected an implementation , the iMC module of iMod , that is capable of generating large domain motions by sampling along low-frequency normal modes [23] .
The linear , branched structure of proteins naturally forms a kinematic linkage , i . e . a chain with rigid groups of atoms , or rigid bodies , as links and rotatable bonds or degrees of freedom ( DoF ) , as revolute joints . The DoFs are the backbone torsion angles ( ϕ and ψ ) and the side-chain torsion angles ( χi ) . Bond lengths , bond angles and the peptide torsion angle ω are assumed fixed at their initial values in this study . Rigid bodies are the largest sets of atoms in a protein without internal , rotational degrees of freedom ( S1 Fig ) . We initially set each atom or group of double-bonded atoms as a rigid body . The rigid bodies of atoms connected by a non-rotatable covalent bond are merged . Hydrogen atoms are explicitly included in the model . A vector q ∈ 𝕊n , q = ( q1 , … , qn ) T completely specifies a conformation for a molecule with n rotational degrees-of-freedom . We represent the kinematic linkage as a rooted , directed spanning tree , i . e . an acyclic graph G = ( V , E ) that connects all vertices V such that each one , except the root , has only one incoming , directed edge E . Vertices Vi , i = 1 , …B represent rigid bodies , and edges Ej , j = 1 , … , n represent DoFs . Hydrogen bonds are encoded as distance constraints , resulting in closed loops or so-called kinematic cycles in G ( Fig 1 ) . A cycle-closing hydrogen bond connects two subtrees propagating from a common ancestor rigid body Vc ( Fig 1 ) . To avoid hydrogen bond dissociation , a perturbation Δq should leave the positions of the hydrogen bond donor atom h and acceptor atom A unchanged with respect to a local coordinate frame placed at A and h . We denote the DoFs subject to constraints as cycle DoFs . For each cycle i = 1…m , we can define endpoint maps f : 𝕊 k ↦ ℝ 3 , x h , AL , R = fh , AL , R ( q ) , which map the ncycle DoFs of the molecular conformation q to the hydrogen bond acceptor A and donor h positions xh , A , along the left ( L ) or right ( R ) sub trees stemming from Vc . The six holonomic closure constraints fhL ( q ) −fhR ( q ) =0 , fAL ( q ) −fAR ( q ) =0 ( 1 ) define a constraint manifold 𝓜 , which is in general ( ncycle − 5m ) -dimensional . Motions on 𝓜 result in coordinated changes to DoFs that satisfy the distance constraints , and thus maintain hydrogen bonds . However , such motions are difficult to calculate since the constraint manifold is complex . We approximate the manifold locally by its tangent space Tq𝓜 at q . Differentiating Eq ( 1 ) yields ddt ( fhL ( q ) −fhR ( q ) ) = ( dfhLdq−dfhRdq ) q˙=0 , ddt ( fAL ( q ) −fAR ( q ) ) = ( dfALdq−dfARdq ) q˙=0 , ( 2 ) which we can rewrite as J q . = 0 . The 6m × ncycle Jacobian matrix , J , gives the instantaneous relationship between the cycle degrees of freedom and the end-point positions and orientations . Entries of the Jacobian matrix are efficiently computed as J ij = u j × ( r - r O j ) , where u is a unit vector along DoF j , r denotes the position of the donor or acceptor atom of the cycle-closing bond , and rOj denotes the position of the tail atom of DoF j . Perturbing a molecular conformation with any vector selected from a sufficiently small neighborhood of the origin in the null-space of J , i . e . Ker ( J ) = {q ∈ 𝕊n:Jq = 0} will maintain hydrogen-bond distances . The right-singular vectors of the singular value decomposition J = U Σ VT form a basis , N , of the null-space of the Jacobian . Note that N is orthonormal , and that NNT is the orthogonal projection onto Ker ( J ) . A null-space perturbation projects a trial-vector Δq onto the null-space , ΔqTq𝓜 = NNTΔq . Previous sampling-with-constraint procedures relied on an elegant combinatorial pebble game algorithm [17] to identify exactly all rigid and flexible substructures in the molecule [15 , 24] . The pebble game algorithm , originally developed for 2D network glasses and later validated for 3D molecular graphs by the molecular conjecture [25] , explicitly counts constraints and degrees of freedom . Our projection method does not require constraint counting , recognizing that the subset of rigidified degrees of freedom Vrigid span the nullspace of the projection matrix Ker ( NNT ) in our method: V rigid = { q : NN T q = 0 } Note that Ker ( NNT ) never requires explicit computation in our method . Mapping a trial move Δq onto Ker ( J ) by NNTΔq naturally projects out the rigidified DoFs . In addition to cycle DoFs , proteins generally have free DoFs that are not part of any closed loop and , therefore , not subject to constraints . Note that the designation free or cycle DoF is independent of the choice of the root R . Bond lengths and angles are assumed fixed in our kinematic representation , representing bonded energy terms . Non-bonded van der Waals interactions are represented by a hard-sphere , repulsive potential that is scaled for each atom type . We use an efficient grid-indexing method for detecting clashes [26] . While no explicit dihedral energy term is present , disallowed dihedral combinations are avoided by clashes . To validate our algorithm , we selected the three proteins with the largest RMSD between apo and holo conformations from a data set curated for predicting apo conformations from holo conformations [27] . Hydrogen bonds shared between apo and holo conformations were included as constraints . The domains were determined as follows: L-Leucine binding protein ( leub ) domain 1 residues 1–119 and 251–327 , domain 2 residues 120–250 and 328–345; Osmo protection protein ( osmo ) domain 1 residues 6–109 and 213–275 , domain 2 110–212 . Alginate binding protein ( algi ) domain 1 residues 1–133 and 310–400 , domain 2 residues 134–309 and 401–490 . For each holo conformation , 20 , 000 random samples were generated with exploration radius of 8Å for leub , 6Å for osmo and 10Å for algi , see the section KGS sampling below for details . To analyze the results , the centers of mass of the holo domains were first aligned with the z-axis of the laboratory coordinate system . Domain 1 in each sample in the conformational ensemble was aligned with domain 1 of the holo conformation before the zenith and azimuth angle of domain 2 of the sample were calculated [28] . We used the ligand-free ( PDB 2ZIJ ) and bound ( PDB 1BB5 ) crystal structures of human lysozyme as starting conformations . We made the L96A mutation to the bound structure to match the wild-type sequence of the ligand-free conformation [29] . Hydrogen bonds shared between the starting conformations were included as constraints . We generated 20 , 000 random samples with an exploration radius of 4Å . To analyze the results , ensemble conformations were aligned to the backbone heavy atoms of the bound structure . The breathing angle θ was computed from the centers of mass of the CA atoms from three protein regions [29] . The RMSD of the CA atoms of secondary structure elements from the bound structure was computed for each ensemble [29] . The angle θ and RMSD were binned in 0 . 5 degrees and 0 . 1Å to calculate ‘free-energy’ landscapes of these reaction coordinates . KGS takes as input a constraint file , which allows users to identify which distance constraints to maintain . In this study , hydrogen bonds belonging to our modeled Linkers I and II were removed . In both systems , the intersection of the sets of hydrogen bonds for active and inactive states was retained . Eventually , KGS sampling of both states was performed with 130 hydrogen bonds in total . A structural representative for activated apo Gαs was extracted from the crystal structure of β2AR:Gs complex with PDB id 3sn6 [30] and inactive apo Gαs was obtained from 1azt [31] . The crystal structure of the inactive state of Gαs had three residue gaps: 1 − 34 , 70 − 86 , 391 − 402 . Residues 70 − 86 ( Linker I , Fig 2 ) were added by Xpleo [16] and subsequently refined in Coot [32] . Finally , the structure was truncated to include residues 35 to 391 ( 357 residues ) . The crystal structure of active Gαs had four residue gaps: 1 − 8 , 60 − 87 , 203 − 204 , 256 − 262 . Residues 60 − 87 were built by Xpleo , 203 − 204 were added in Coot , 254 − 265 were copied from the inactive structure of Gαs after alignment , and the sequence was also truncated to include residues from 35 to 391 . The α subunit of Gs consists of a Ras-like domain and an α-helical ( AH ) domain ( Fig 2 ) . The Ras-domain is about 260 residues , which is connected to the AH-domain ( about 112 residues ) by two linkers . The long α1–αA linker I , spans residues 65 to 88 , and a shorter αF–β2 linker II spans residues 200 to 206 ( Fig 2 ) . These structures were then parametrized by the CHARMM27 all-atom force field [33] including the CMAP correction [34] and solvated in an octahedral unit cell with 19 , 737 TIP3 water molecules and electrostatically neutralized by 22 Na and 12 Cl ions ( concentration of 0 . 05 M and no ions within 6Å of any protein atom ) for a total of 65 , 000 atoms . The resulting system was minimized with Gromacs 4 . 6 . 3 [35 , 36] by a series of steepest descent and conjugate gradient algorithms by gradually reducing constraints on the protein atoms . The minimized structures of active and inactive apo Gαs served as the input models for the sampling algorithms . The Gs α subunit was represented by 1769 rigid bodies and 1768 directed edges corresponding to the dihedral DoFs ϕ , ψ , and χi . While any rigid body in the molecule can serve as the root R , we set R as the first rigid body at the N-terminus of the molecule . There were 767 cycle DoFs in the system . To ensure rapid and broad diffusion of the sampled ensemble , the sampling protocol inspired by Rapidly-exploring Random Trees of previous work was used [14 , 15 , 37] , which we briefly summarize . The sampling pool was initialized with the minimized conformations of active or inactive apo Gαs qinit . We generated a pool of 20 , 000 samples in an exploration sphere of fixed radius ( 20Å RMSD ) from qinit , which was subdivided into shells 𝓢i , i ∈ {1 , … , 100} of width 0 . 2Å , as follows . At each sampling step , a shell 𝓢k was selected at random from the subset of shells containing at least one conformation . Next , an entirely random conformation qrandom was generated . The conformation that was RMSD-closest to qrandom in 𝓢k was selected as qseed , and qrandom was discarded . A random perturbation Δq to qseed was proposed , that was then projected onto the constraint manifold and applied to qseed to obtain a new conformation qnew , i . e . qnew = qseed+NNTΔq . If qnew did not contain clashes , it was added to the pool in the shell corresponding to its RMSD from qinit , else it was discarded . The exploration radius and shell width are adjustable parameters . The combination of values selected for this study were found to balance broad exploration and uniform coverage . The collision factor that scales VdW radii during collision detection was set to 0 . 75 . The maximum rotation of a DoF was scaled to 0 . 29 degrees , which was found to reflect a good balance between fast divergence from initial structure and a high acceptance ratio . To test if the sampling trajectories had converged , we additionally generated a conformational distribution of 50 , 000 samples around the inactive and active states . All analyses are based on 20 , 000 samples , unless otherwise stated . We carried out ENM normal modes vibrational analysis ( NMA ) in internal coordinates ( IC ) with the software package iMOD [23] . After first obtaining the IC normal modes for each structure with the iMODE tool , we generated a conformational ensemble of 20 , 000 samples with the default NMA Monte Carlo sampling procedure enabled by the iMC module [23] . We obtained the first 20 normal modes by using all default settings , except enabling χ dihedral angles as DoFs to better agree with the KGS DoFs . By default iMC selects from the 5 lowest frequency modes for a Monte Carlo step . S2 Fig . displays all the modes . We used coarse-grained all heavy-atom representation and a sigmoid function pairwise interaction potential with default parameterization . We scaled the parameter a ( ’linear factor to scale motion’ ) ten-fold to better match the amplitude of domain motions suggested by experimental measurements . Further increasing the scaling did not lead to better agreement . To examine if a sigmoid function potential possibly over-constrained the system , we also sampled using a coarse-grained , CA-only representation with an essential dynamics ( ED ) potential function . A scale factor of a = 10 agreed with experimental data , but led to distortions in the models . ( S3 Fig ) . Thus , to enable a direct , one-to-one comparison between KGS and ENM , we selected an all-heavy atom , sigmoid function representation for iMC with amplitudes scaled by a = 10 , notwithstanding its slightly overconstrained model . We additionally generated conformational ensembles with the distance-restraint based sampling procedure CONCOORD [38] . We used the default , heavy-atom CONCOORD settings for structure and distance bounds generation with OPLS-AA parameters . We used near-default parameters for sampling , using the following command line: disco -on disco . pdb -n 20000 -i 2500 -viol 1 . -bump . The software is implemented in C++ . Calculations were performed on a single , 2 . 6GHz Intel processor core . Average time to obtain an accepted conformation was 8 . 9s , at an average acceptance ratio of 30% . Depending on the size of the molecule , computations take from several hours to a few days . No attempts were yet made to optimize the code . The performance limiting step is currently the repeated ( 𝓞 ( n2 ) ) calculation of RMSD within shells to ensure broad sampling . The shell width balances performance with broad diffusion . The RMSD calculation would be trivially replaced by more modern algorithms that are two orders of magnitude faster [39] . Our SVD calculation is optionally GPU-accelerated . The software and sampling trajectories are available from http://smb . slac . stanford . edu/~vdbedem .
To validate our algorithm , we computed conformational distributions for three two-domain protein crystal structures that were determined in both holo and apo conformations . For each protein , the domains open , re-orient and conformationally adjust upon adopting the apo conformation . We observed conformational distributions directed along holo-apo pathways . Starting from the holo conformation , we found that conformational ensembles on the constraint manifold defined by interconnected cycles were highly biased toward the apo conformation ( Fig 3a ) . Polar plots of the distribution of zenith ( θ ) and azimuth ( ϕ ) angles of relative positions of the centers of mass of the two domains reveal domain opening and collective , reorientating motions toward the apo conformation . No conformational pathways connecting the holo substate to the apo substate were observed , but it is unknown if ligand-free holo-apo conformational interconversion occurs for these proteins in solution . Additionally , sampling limitations or steric barriers between the states can prevent end-to-end pathways . Reaching sparsely populated , ‘excited’ substates often demand additional ( experimental ) restraints on conformational sampling techniques [14 , 27 , 29] . We furthermore tested whether conformational distributions owing to collective motions on the constraint manifold can accord with free energy landscapes observed in solution . Apo human lysozyme displays large breathing motions , characterized by the angle θ between the α and β domains . The free-energy landscape for the reaction coordinates θ and RMSD to the holo crystal structure of apo and holo ( triNAG-bound ) human lysozyme was recently characterized from replica-averaged , RDC-restrained molecular dynamics simulations [29] . While the free energy of apo lysozyme has a single minimum , the holo state revealed a second , sparsely populated ‘unlocked’ state centered on ( 49° , 1 . 5Å ) in addition to the main ‘locked’ state around ( 58° , 0 . 9Å ) ( Fig 3b ) . The holo protein is capable of sampling a wider range of θ angles than the apo structure , presumably to facilitate product release . Our conformational distributions starting from the ( ligand-free ) holo and apo structures revealed surprisingly similar conformational distributions compared to those from RDC-restrained simulations ( Fig 3b , left panel ) . The holo distribution samples more broadly , and more towards closed conformations ( smaller θ angles ) than the apo distribution , in agreement with the free-energy landscape observed from RDC restrained simulations . Additionally , weak local maxima were observed in the holo distribution , corresponding to the ‘locked’ and ‘unlocked’ state ( Fig 3b , right panel ) . The unlocked state corresponds to a sparsely populated , intermediate state , which was validated experimentally . Thus , collective motions on the constraint manifold enable quick diffusion away from the initial state along biologically-relevant directions that map the conformational landscape of the protein . β2AR can form a complex with heterotrimeric stimulatory protein Gαsβγ[40] . While the precise mechanism of protein Gs-activation remains poorly understood , interaction with the activated receptor is incidental with the dissociation of GDP and the βγ subunits [41] . Gαs , which binds GTP after the release of GDP , subsequently interacts with many effector proteins in the cell . It is hypothesized that its profound conformational flexibility plays a crucial role in signal modulation [42] . The activated ( nucleotide-free ) state of Gαs involves a large motion of the AH-domain with respect to the stable Ras-domain [43] . Additionally , the α5-helix of Gαs translates and rotates upward to interact with the cytoplasmic core of the receptor . The domain opening purportedly facilitates the release of GDP . The β6–α5 loop , which binds the purine ring of GDP , and the β6 strand also change conformation ( Fig 2 ) . The large distance separating the crystal structures of the active and inactive states suggests that Gαs can access many different conformations [2 , 42 , 44] . However , structurally characterizing and determining the sequence of events in the activation pathway by experimental means has proved challenging . Simulations suggest a coupled motion between the AH-domain and helix α5[13 , 45] . Additionally , the opening angle of the AH-domain upon activation is the subject of intense debate . Several lines of evidence suggest that crystal lattice formation may have played a role in selecting an extreme opening angle for the AH-domain [10 , 44] . Distance distributions obtained in solution indicate that the conformational variability of the AH-domain of Gi protein in complex with rhodopsin is more limited than that observed in the crystal structure of β2AR:Gs [10] . We first examined the conformational variability of the AH-domain between the active and inactive states with the methods KGS , iMC , and CONCOORD . The RMS deviations for KGS samples starting from the inactive state of the AH-domain of Gαs was 13 . 5Å , while for the active state it was 5 . 8Å ( Fig 4a ) . For iMC , the observed values were 5 . 2Å and 11 . 1Å , and for CONCOORD 9 . 7Å and 15 . 8Å ( Fig 4b ) . In addition , all three methods identify large motions of helix α5 concurrent with the domain motions . The maximum opening angle Θmax between the two domains was 27 . 2 degrees ( Fig 5 and S4 Fig , 37 . 9 degrees for 50 , 000 samples ) for the inactive state KGS ensemble , compared to 91 degrees for the activated crystal structure ( Fig 2 ) . iMC reported a maximum opening angle for the inactive AH-domain of around 9 . 9 degrees ( Fig 5 , left panel ) . The CONCOORD conformational ensemble reported a range of 15 − 20 degrees of an opening angle around the active state , and 18 degrees around the inactive state ( Fig 5 , right panel ) . Both iMC and CONCOORD sample with nearly uniformly fixed radius around the active starting conformation , which is rationalized by their reliance on an equilibrium conformation ( Fig 5 ) . In contrast , KGS , by design of its RRT-based sampling avoiding steric collisions , mimics a trajectory diffusing out of the starting conformation . KGS sampling of the activated state exhibited a slowing rate of change , while the opening angle of the inactive state still appeared to increase slightly at 20 , 000 samples , leveling of at 50 , 000 samples ( S4 Fig ) . The lack of full convergence did not appreciably change the conformational distributions , but can moderately limit interpreting the ensemble as a collection of exchanging conformational substates . Interestingly , while KGS sampling of the active conformation initially exhibits greater conformational diversity away from the inactive state , later samples are directed more towards the inactive state . The KGS ensemble for free , apo Gαs compares very well with the RMSD and opening angle reported from experimental observations in solution . Double Electron-Electron Resonance ( DEER ) spectroscopy measurements suggest that the average displacement of the apo AH-domain of Gi protein complexed with rhodopsin is 15Å [10] . From the nine models of receptor-bound Gi conformations reporting on the DEER observations , we measured an equivalent average opening angle Θ of 25 . 5 degrees ( Θmax = 48 . 8 degrees ) after alignment to the Gαs Ras-domain . The KGS ensemble suggests that ligand-free Gαs is structurally and evolutionary designed to access a broad range of opening angles . However , a set of discrete samples connecting the inactive with the active state of Gαs ( Fig 4a and 4b ) was not observed . The sample acceptance ratio in KGS , i . e . samples not rejected owing to collisions between atoms , also differed substantially between the inactive and active states ( 35% vs 15% ) . These findings could signify a steep conformational barrier between the inactive and active crystal structures between 40 to 80 degrees of domain opening angle . For instance , in the activated state of β2AR:Gαs , the α1-helix of the Ras-domain is partially melted to accommodate the large motion . In the remainder , we focus on a direct comparison of the directional conformational variability of KGS and iMC since these methods are conceptually most alike . We calculated a distribution of angles between the mean displacements of Cα atoms of the Gα AH-domain for the two ensembles ( Fig 6a and S5 Fig ) . The mean displacement is the vector connecting the center of mass of all AH-domain Cα atoms of the initial structure to the averaged center of mass of the ensemble , after alignment to the stable part of the Ras-domain . The angles of mean displacement for the AH-domain are visualized in Fig 6 . The angles of mean displacements did not align but were significantly shifted for both the active ( 57 . 6 degrees , Fig 6a yellow bar ) and inactive state ( 70 . 2 degrees , Fig 6a gray bar ) . The long tails for the angle distributions , in particular for the inactive state , identify a significant number of residues for which the angle differ by more than 90 degrees . Thus , large-amplitude motions of Gαs are described differently by the two procedures . The conformational distributions from KGS starting from the inactive form of Gαs aligns with the proposed activation mechanism of the β2AR:Gαs after GDP release . The direction of motion for the KGS inactive ensemble corresponds to a domain opening motion in the viewing plane , with a small component orthogonal to the viewing plane ( Fig 6b ) . The iMC motion is nearly orthogonal to the viewing plane , resulting in a transverse ‘rocking’ motion , with a moderate component downward towards a domain opening motion . Floquet and coworkers observed a similar , pivoting motion for the AH domain around the αA helix , which is implicated in GDP release , from Cartesian NMA with the CHARMM27 force field for protein Gi [13 , 46] . The size of the vectors reflects the difference in RMSD amplitude of the ensembles . For the active state both methods have a significant component orthogonal to the viewing plane . For neither method the main displacement in the active state appears to be along the activation pathway , signifying that additional mechanisms , such as GTP hydrolysis , likely play a key role in Gαs . The direction of mean displacement for iMC is nearly identical for the inactive and active ensembles . A possible explanation is that local structural changes in the AH-domain between the active and inactive state are modest , leaving interactions defined by the ENM largely unchanged between the states . Receptor-induced conformational changes in helix α5 are believed to contribute to GDP release [10 , 47] . Concomitant with activation , helix α5 undergoes a rotation and translation towards β6 . The magnitude and direction of these fluctuations in the KGS ensemble are striking , coinciding with those observed in MD simulations [45] . Fig 6c shows the view looking towards the cytoplasm from the receptor core . The top panel shows the α5 helix in its active conformation , and the bottom panel in its inactive conformation . The distribution of magnitudes and directions of the KGS displacement vectors along the helix in the inactive state ( bottom panel ) correspond remarkably well to a translation and rotation along a path to reach the active state ( top panel ) . RMS amplitudes of 8 . 3Å and 7 . 9 Å were observed for α5-helix in the KGS ensemble of the active and inactive states . By contrast , the iMC displacement vectors are slightly smaller in magnitude in the inactive state ( indicated by RMS spread of 1 . 4Å and 3 . 6 Å ) and have a component nearly opposite to the activation pathway . Note that while in general normal mode vectors indicate undirected displacement , our displacement vectors were calculated directly from the ensembles . Next , we analyzed displacements at the residue level for both domains of Gαs . Fig 7 ( S5 Fig ) shows the normalized magnitude of the mean CA atom displacement vectors of the ensembles . Each displacement vector was calculated as the average RMSD vector of all conformations after alignment to the stable part of the Ras-domain ( as above ) , and normalized within the angle values of its own ensemble . Mean displacements for the KGS and iMC sampled Gαs ensemble exhibit a clear pattern; they are larger for the AH-domain and vanishingly small for the Ras-domain in both the active and inactive state . The coordinated perturbations of DoFs by KGS resulted in intra-domain displacements shared by groups of contiguous residues . Three regions of the AH-domain separately display collective features indicated by elevated mean displacements , corresponding to the C-terminus of αA and αB , αC and αD , and αE and αF . A remarkably similar pattern is observed for the iMC ensemble . Helices A − D are located towards the outer radius of the rotation of the AH-domain , explaining the elevated levels of mean displacement in both active and inactive state ( Fig 2b ) . Surprisingly , their relative orientation remains well-preserved despite a sparse inter-secondary structure hydrogen-bond network in the AH-domain . The pattern of displacements is similar for the active and inactive state . Analysis at the residue level reveals key details suggesting collective motion . In the Ras-domain , helix α5 shows a large displacement , exceeding the mean displacement values of the Ras-domain ( Fig 7 , right-most shaded bands ) . The growth in amplitude towards the C-terminus is characteristic for the rotational motion we observed in the previous section . Interestingly , the single , unique feature standing out in an otherwise flat Ras-domain is elevated displacement for helix α4 and loop αG–α4 ( residues 320–340 ) in both active and inactive state ( Fig 7 ) . Fig 6d shows the motion of α4 and the adjacent loop . Helices αG and α4 are implicated in GDP release . Similar motions were observed using Cartesian NMA with the CHARMM27 force field [13] in protein Gi . Strikingly , both the α5 and motions of α4 and the adjacent loop are absent in the iMC Gs active ensemble , but both are present in KGS . This strongly suggests these motions are conformationally coupled , but possibly shifted to higher modes in iMC . The mean displacements up to residue number 80 suggest anti-correlated motions in iMC and KGS in the active state ( Fig 7 , top ) . The mean displacement reported by iMC is elevated owing to restraints between the BC loop in the AH domain and ( truncated ) helix α1 . This results in collective motions of the β1-strand with the highly mobile AH domain . The amplitude of the iMC motions is likely overestimated , as it leads to significant distortions of the β-sheet in the Ras domain ( S6 Fig ) . Similarly , the proximity of Linker I to helix F leads to collective motions in iMC . The absence of explicit constraints , i . e , hydrogen bonds in Linker I suppresses collective motions in KGS . While the precise nature of Linker I motions remains unclear , the absence of well-defined electron density in the crystal suggests this loop is highly mobile . To examine the origin of collective motion , we analyzed the distribution of the DoFs in the conformational ensembles . We observed key differences between the two methods in the spatial distribution of flexibility throughout the protein . The mean RMSF for free and cycle DoFs are summarized in Table 1 . Cycle DoFs are uniformly distributed throughout the protein . In KGS , 43 . 4% of total DoFs are cycle DoFs and of those 41% are rigidified , indicated by vanishing RMSF for cycle DoFs ( Fig 8 ) . These DoFs are contained in the null space of the projection matrix NNT . Rigidified cycle DoFs correspond largely to secondary structure elements , where DoFs are overconstrained by short or overlapping cycles . Free DoFs have larger RMSF than cycle DoFs ( Table 1 ) . If rigidified DoFs are excluded from the RMSF , a modest reduction of 20 . 5% in flexibility from free to cycle DoFs is observed . By contrast , while iMC does not define free or cycle DoFs , we observed a reversed flexibility trend compared to the corresponding DoFs in KGS–the cycle DoFs are 1 . 8 times more flexible than free DoFs for iMC . One possible contributing factor to this somewhat paradoxical find is that normal modes are obtained from eigenvectors of the Hessian matrix defined by the potential function . Free DoFs , like those in surface side-chains , are , on average , subject to fewer restraints , and thus less likely to contribute to major modes . The magnitude of a trial move is scaled by the eigenvalues of the modes , and more constrained areas may thus dominate the size of the move . We also observed that large parts of the Ras-domain do not show elevated RMSF with iMC ( Fig 8 ) , signifying that many vibrational frequencies rather than a single mode dominate structural changes for this domain . iMC locates elevated flexibility mainly in loop residues ( Fig 8 ) . Linkers I and II stand out , as well as the β6–α5 loop . Note that the backbone DoFs for LI and LII are cycle DoFs owing to hydrogen bonds between , for instance , β1 and β2 . By contrast , elevated variability in KGS is concentrated less in loop areas , and distributed more uniformly throughout the protein . The magnitude of helix α5 RMSD spread is nearly identical in the two states . However , small , motional differences in specific helical sub-regions can signify different functionalities . Significant flexibility towards the C-terminal part of the helix would enhance α5-helix conformational selectivity for inactive-like conformers , while a more active-like conformation would promote specificity through small-scale deformations near the N-terminal part of the helix . For the inactive state we observe elevated variability in the KGS ensemble from the C-terminus of α4 , through β6 , up to the N-terminus of α5 ( Fig 8a , dashed rectangle I ) . The C-terminus of helix α5 and the α4–β6 loop interact with the receptor . A conserved TCAV motif in the β6–α5 loop binds the GDP guanine ring . Helix α5 and strand β6 transmit receptor-induced conformational changes to facilitate GDP release [40] . KGS elevated variability is present in the inactive state , but moderated in the activated state and shifted away from the β6 strand . The magnitude of variability is reduced from inactive to active state for both sampling techniques , suggesting that smaller changes dominate this area in the active state . This interpretation is supported by iMC motions , where elevated variability shifts from β6 to the N-terminus of α5 upon activation . A heat map of conformational changes reveals a hotspot of highly elevated flexibility near the GDP binding pocket in the inactive state ( Fig 8b , bottom left circled ) . Similarities with a heat map obtained from peptide amide hydrogen-deuterium exchange mass spectrometry ( DXMS ) experiments , which report on exchange rates of amide hydrogens , are striking [48] . While the increased exchange rates established general sensitivity to GDP release , our nucleotide-free analysis suggests that increase in dynamics or disordering of this segment is , at least partly , attributable to motion of helix α5 and the AH domain . We also observed conformational coupling of the N-terminus of helix α5 to α1 , and the adjacent β1–α1 loop ( P-loop ) , which binds the nucleotide phosphate ( Fig 8a , dashed rectangle II and circled in inactive state ) . How the elevated flexibility is further coupled is illustrated in Fig 8b , left panels . Coupling in the GDP binding pocket extends to include helix α1 , helix αF , Linker II ( SW I ) , and the N-terminus of αE . Functional , conformational coupling is revealed to a lesser extent by iMC ( Fig 7b , right panels ) . In particular , the close coupling around the GDP binding pocket appears absent , and elevated flexibility is mostly located in loop residues . For iMC , variability of αE is shifted towards the C-terminal end of helix αD .
Proteins interconvert between functional , often sparsely populated conformational substates at a multitude of spatiotemporal scales to perform their function and interact with other biomolecules [49–51] . Understanding how these substates probe the conformational landscape and how they are coupled through collective motions can provide insights into molecular mechanisms and protein function [52 , 53] . Our conformational sampling algorithm maps small random perturbations , highly suggestive of equilibrium fluctuations , onto a constraint manifold that is defined by the hydrogen bonding network . Our new method does not require explicitly calculating rigid substructures of the protein . Instead , DoFs are subject to coordinated motion on the constraint manifold , and DoFs in isostatic or overconstrained substructures are intrinsically rigidified . Cycle DoFs contribute significantly to the distribution of the resulting conformational ensemble . Cycle DoFs make up nearly half of the DoFs , are distributed throughout the molecule , and their RMSF is only moderately reduced compared to free DoFs . The coordinated motion and distribution of cycle DoFs can potentially provide new information about mechanisms of conformational coupling . Compared to iMC , we observed motions with larger amplitudes , but both methods were in agreement with accepted mechanisms . We were better able to distinguish molecular mechanisms , and locate the origin of conformational flexibility . Important rotational DoFs stand out , and are , surprisingly , located not just in loops to accommodate inter domain motion . Results for Linker I and ( activated ) loop residues 254–265 should be interpreted with care , as experimental evidence to support their initial conformation is limited . Conformational coupling in the iMC ensemble was less pronounced , and sometimes more difficult to distinguish owing to higher modes or reduced motional amplitudes . The limited range of motion of the iMC sampling procedure likely results from the assumption of harmonic vibrations around equilibrium positions in the ENM . Large deviations break the underlying assumptions and would perturb the topology of the initial conformation–drawbacks that KGS intrinsically avoids . In contrast to the KGS distributions , the RMSD for the inactive state are reduced compared to the active state . The interface between the AH-domain and Ras-domain is subject to restraints imposed by the ENM , which limits the amplitude of the motion along the activation pathway from the inactive state . Nonetheless , while the active state sampled ensemble exhibits a larger RMSD than the inactive state , an overall reduced amplitude with respect to KGS was observed . Interdomain ENM restraints in the direction of the activation pathway alone do not explain the reduced RMSD . We observed a KGS ensemble along a pathway associated with activation for the α sub-unit of protein Gs . Conformational interconversions can occur through a myriad of alternative transition pathways . Computationally probing a multi-state conformational landscape through extensive MD simulations to obtain a probable minimum free energy pathway is often prohibitively expensive . In addition , sampling is generally affected by limitations and imperfections of the force fields [54] . At the expense of highly accurate energy estimates , our method efficiently explores the conformational space accessible to a protein while it maintains exactly covalent and hydrogen bond geometry , and avoids steric clashes . Nonetheless , interpretation of the ensemble as a collection of exchanging conformational substates would require long sampling trajectories to satisfy ergodic properties . Our method illuminates coupled intra- and interdomain motions , complementary to rigid-body domain sampling and subsequent loop rebuilding [10] . Paired with sophisticated MD simulations or energy relaxation protocols [55 , 56] our conformational ensemble can , for instance , serve as starting points for detailed transition path sampling . An exceptionally striking feature of our KGS ensemble is the magnitude of fluctuation of helix α5 concurrent with the Gαs AH-domain motion . These coupled motions point to a potential molecular mechanism of concomitant , structural changes between two remote sites implicated in the release of GDP upon activation . There is increasing experimental evidence to support this mechanism , which was first predicted by computational means for protein Gi by Floquet and coworkers [13] . Their NMA-based analysis of GDP-bound Gi identified a motion for the AH-domain that pivots on the long axis of the αA helix . Surprisingly , this transverse motion qualitatively agrees with our nucleotide-free iMC analysis . The similarity of nucleotide-free and nucleotide-bound motions is likely owed to ENM interactions between the Ras and AH domain in iMC , which mimic interactions of the nucleotide with each domain . Essential Dynamics Analysis ( EDA ) of AH domain motions upon ejection of GDP on the phosphate side from selected nanosecond time scale Targeted Molecular Dynamics trajectories furthermore revealed close agreement with motions from NMA analysis [46] . The transverse motion likely plays a key role in GDP release and Gi activation at nanosecond time scales . By contrast , whereas the AH domain motion for Gαs observed from KGS analysis also exhibits the small transverse component , it is mainly directed along a domain opening trajectory in agreement with DEER measurements , potentially additionally identifying longer , micro- to millisecond time-scales motions . While it is speculative to join analyses from two different proteins , these observations do suggest an activation mechanism whereby a transverse ‘rocking’ motion facilitates or results from GDP release , which in turn leads to a domain opening motion . For inactive , apo Gαs , elevated mobility is centered on a hotspot near the GDP binding pocket , extending to the N-terminus of helix α5 , α1 , and the adjacent P-loop . The mobile helix α4 is conformationally coupled to the hub through β6 . Previous studies established that a conformational ensemble obtained by maintaining hydrogen bonds through iteratively refitting rigid substructures agrees well with MD simulations [57] . In our method , the coordinated motions on the constraint manifold resulting from hydrogen bond encode ‘natural’ modes of deformation . However , these coupled motions and broad diffusion by a carefully selected sampling strategy come at the expense of a greatly simplified energy function . It allows our method to overcome high-energy barriers , but can lead to conformations with high physical energies . Thus , care should be taken in interpreting individual ensemble members prior to extensive energy minimization . Hydrogen bonding networks enforce collective motions that couple conformational substates implicated in GDP release . Our results highlight that in addition to stabilizing tertiary structure , hydrogen bonding networks mediate molecular mechanisms and dynamics . Indeed , evidence is emerging that hydrogen bonds mediate longe-range , correlated motions [58] . Our nullspace sampling procedure with explicit , holonomic constraints can relate motion to function by revealing molecular mechanisms . It enables researchers to formulate testable hypotheses about networks of residues that facilitate motions implicated in GDP release and AH-domain motion . In addition , our procedure could be augmented with intra-molecular distance constraints obtained from experimental data . | Multi-cellular physiology is an emergent property , which depends critically on inter-cellular signaling pathways . Transmembrane G protein-coupled receptors ( GPCRs ) mediate a large variety of physiological events throughout the body , such as vision or cardiovascular regulation . It is thus no surprise that GPCRs are targeted by more than one third of all FDA-approved drugs . Molecules such as hormones and neurotransmitters transmit messages to cells via GPCRs complexed to cytosolic heterotrimeric G proteins . G proteins , upon activation , interact with other molecules to trigger a cellular response . Despite an increasing amount of structural data , the precise conformational dynamics and activation mechanism of G proteins remain poorly understood . The size of the multi-protein complexes and the time scales at which conformational changes occur hinder adequate sampling of the conformational landscape with molecular dynamics simulations . Here , we extend and use an efficient , robotics-inspired conformational sampling procedure to probe the conformational landscape of protein G during activation . Our procedure reveals coupled , molecular mechanisms of the activation pathway , which are absent in a comparative analysis with normal modes . Our exciting results can ultimately lead to modulation of biological activity by drug design or fine-tuning of conformational heterogeneity . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Nullspace Sampling with Holonomic Constraints Reveals Molecular Mechanisms of Protein Gαs |
Quantitative imaging has become a vital technique in biological discovery and clinical diagnostics; a plethora of tools have recently been developed to enable new and accelerated forms of biological investigation . Increasingly , the capacity for high-throughput experimentation provided by new imaging modalities , contrast techniques , microscopy tools , microfluidics and computer controlled systems shifts the experimental bottleneck from the level of physical manipulation and raw data collection to automated recognition and data processing . Yet , despite their broad importance , image analysis solutions to address these needs have been narrowly tailored . Here , we present a generalizable formulation for autonomous identification of specific biological structures that is applicable for many problems . The process flow architecture we present here utilizes standard image processing techniques and the multi-tiered application of classification models such as support vector machines ( SVM ) . These low-level functions are readily available in a large array of image processing software packages and programming languages . Our framework is thus both easy to implement at the modular level and provides specific high-level architecture to guide the solution of more complicated image-processing problems . We demonstrate the utility of the classification routine by developing two specific classifiers as a toolset for automation and cell identification in the model organism Caenorhabditis elegans . To serve a common need for automated high-resolution imaging and behavior applications in the C . elegans research community , we contribute a ready-to-use classifier for the identification of the head of the animal under bright field imaging . Furthermore , we extend our framework to address the pervasive problem of cell-specific identification under fluorescent imaging , which is critical for biological investigation in multicellular organisms or tissues . Using these examples as a guide , we envision the broad utility of the framework for diverse problems across different length scales and imaging methods .
Diverse imaging techniques exist to provide functional and structural information about biological specimens in clinical and experimental settings . On the clinical side , new and augmented imaging modalities and contrast techniques have increased the types of information that can be garnered from biological samples [1] . Similarly , many tools have recently been developed to enable new and accelerated forms of biological experimentation in both single cells and multicellular model organisms [2–10] . Increasingly , the capacity for high-throughput experimentation provided by new optical tools , microfluidics and computer controlled systems has eased the experimental bottleneck at the level of physical manipulation and raw data collection . Still , the power of many of these toolsets lies in facilitating the automation of experimental processes . The ability to perform real-time information extraction from images during the course of an experiment is therefore a crucial computational step to harnessing the potential of many of these physical systems ( Fig 1A ) . Even when off-line data analysis is sufficient , the capability of these systems to generate large , high-content datasets places a large burden on the speed of the downstream analysis . Automated image processing and the use of supervised learning techniques have the potential for bridging this gap between raw data availability and the limitations of manual analysis in terms of speed , objectivity and sensitivity to subtle changes [11] . In this area , many computer vision techniques , including some general object detection strategies , have been developed to address the detection and recognition of faces , vehicles , animals and household objects from standard camera images [12–17] . While this body of literature solves complex recognition problems within the domain of everyday objects and images , it is not clear how or whether they are generalizable to the imaging modalities and object detection problems that arise in biological image processing . While these techniques have garnered some important but limited adoption in biological applications[18–28] , there is not a systematic methodology by which these computational approaches can be applied to solving common problems in mining biological images [29] . Thus , the development or adaptation of these tools for specific problems has thus far been relatively opaque to many potential end-users and require a high degree of expertise and intuition . At the same time , there is a diverse array of specific object recognition problems that arise in biology . Specifically , extraction of meaningful information from biological images usually involves the identification of particular structures and calculation of their metrics , rather than the usage of global image metrics . Depending on the specimen and the experimental platform , this may range in scale from molecular or sub-cellular structure to individual cells or tissue structures within a heterogeneous specimen , or entire organisms . While toolsets have already been developed to address some common needs in biology [19–22 , 24 , 25 , 30–32] and while powerful algorithmic tools exist for pattern and feature discrimination and decision-making [33–35] , there are still many unaddressed needs in biological image processing . Here , we present a general scheme for the detection of specific biological structures applicable as a basis for solving a broad set of problems while using non-specific image processing modules . As opposed to finished , ready-to-use toolsets , which address a limited problem definition by design , the workflow we propose has the power to simultaneously address the need for accuracy , problem-specificity , and generalizability; end-users have the opportunity to choose platforms and customize as needed . We demonstrate the power of this approach for solving disparate biological image processing problems by developing two widely relevant toolsets for the multicellular model organism , Caenorhabditis elegans . To address the problems of extracting region- , tissue- and cell-specific information within a multicellular context , we developed an image processing algorithm to distinguish the head of the worm under bright-field imaging and a set of tools for specific cell identification under fluorescence imaging . These developments demonstrate the flexibility of our framework to accommodate different imaging modalities and disparate biological structures . The resulting toolsets contribute directing to addressing two fundamental needs for automated studies in the worm and contribute specific concepts and modules that may be applied to a broader range of biological problems .
Due to its relatively large size , only a limited portion of the adult worm body can be captured within the field of view under high-resolution imaging; yet it is necessary to target specific regions along the anterior-posterior axis of the worm to capture or apply experimental perturbations to specific cells or tissues of interest ( Fig 2A ) . Thus , image processing for orientation along the anterior-posterior axis of the worm is crucial to enabling the full potential of many of the toolsets for high-resolution imaging and physical , chemical and optical manipulation of the worm . To address this need , many ad hoc tactics such as the presence of fluorescent markers [5 , 24 , 38 , 39] or the assumption of forward locomotion in freely moving worms [22 , 25 , 32 , 40–43] are often used delineate between the head and tail and orient the anterior-posterior axis . However , reliance on exogenously introduced fluorescent markers can necessitate time-consuming treatment of the worms under study and can spatially interfere with other fluorescent readouts of interest . While the assumption of forward locomotion does not require additional treatments , it is only useful in experimental contexts where worms are freely mobile . Therefore , these tactics lack general applicability to many high resolution imaging experiments , where worms may lack appropriate fluorescent markers or are physically restrained or chemically immobilized . Additionally , not relying on fluorescent markers avoids unnecessary photobleaching of the sample before data acquisition and affords robustness against age and condition-specific autofluorescence in the worm body [44] . In order to approach this problem with minimal reliance on specific experimental conditions , we note several consistent morphological differences between the head and the tail of the worm that are observable in bright-field imaging . Bright-field is a commonly available imaging modality and often used for location and positioning of specimens prior to fluorescent imaging . While the shape of the head and the tail differs somewhat , these differences are difficult to detect due to low contrast and may be physically obscured by some experimental platforms [38] . Instead , the head of the worm is more clearly distinguished by the presence of the pharynx , which has a stereotypical morphology that includes a biological structure for masticating food called the grinder [45] . As shown in Fig 2A , the grinder is a dark , uniquely shaped , high-contrast structure under bright field imaging . The grinder can also be easily resolved by most digital cameras at imaging magnifications above 20X and maintains its shape and integrity for several days of early adulthood [46 , 47] . This stereotypical feature of the head , which is relatively consistent in the worm post-developmentally , can thus serve as the target biological structure for our two-layer classification scheme . To construct and validate our classification scheme , bright-field images of the worm head and tail were collected using a custom microfluidic device ( Materials and Exp . Methods ) , although similar images on agar pad would also suffice ( S1 Fig ) . Following our architecture in Fig 1B from left to right , application of the scheme involves three major steps: preprocessing of raw images to generate candidates for the structure of interest , selection and calculation features to describe these candidates at both layers of classification , and optimization and training of the two classifiers based on these feature sets . First , in the preprocessing step , we apply a minimum intensity projection to consolidate dark structures of multi-plane bright-field images into a single image ( MP in Fig 2B ) and use Niblack local thresholding to generate discrete binary particles as potential candidates for the grinder particle ( BW0 in Fig 2B ) . We employ the Niblack local thresholding procedure in both this and our subsequent cell identification application to robustly segment particles , despite the potential variability in local lighting , texture and background tissue intensity as there would be in different imaging setups ( Materials and Exp . Methods ) . Following initial thresholding , preliminary filtering of the binary particles is then applied to remove segmented regions that are either too small ( less than 37 . 5 μm2 ) or too large ( greater than 100 μm2 ) to reduce downstream computation ( BW1 in Fig 2B ) . The remaining particles are processed through our two-layer classification scheme to detect the presence of the pharyngeal grinder . Second , in the feature selection step , distinct mathematical descriptors that may help to describe and distinguish the structure of interest are calculated for each layer of classification . In the first layer of classification , intrinsic and computationally inexpensive metrics of the particles are computed and used as features ( Fig 2C and S2 Fig ) in classification of the grinder shape . These features represent a combination of simple , intuitive geometric features , such as area and perimeter , in addition to higher level measures of the object geometry and invariant moments suitable for shape description and identification [36] . Training and application of a classifier with this feature set eliminates candidates on the basis of intrinsic shape ( BW2 in Fig 2C ) . However , the resulting false positives in Fig 2D show that the information within these shape metrics is insufficient to distinguish the grinder with high specificity . To refine the description of the biological structure in the second layer classification , we utilize features that describe the relationship of candidate particles to nearby particles and texture ( Fig 2E and S3 Fig ) . Specifically , we note that the grinder resides inside the terminal bulb of the pharynx , which is characterized by a distinct circular region of muscular tissue ( Fig 2A ) . Based on this observation , we define second layer features based on distributions of particle properties within a circular region around the centroid of the grinder candidate particle ( S3 Fig ) . Noting that the pharyngeal tissue is characterized by textural ranges in the radial direction and relative uniformity in the angular direction , we build features sets describing both the radial and angular distributions the surrounding particles ( S3 Fig ) . Using the features outlined in Fig 2 , each classification step is a mathematical model that is trained to distinguish between structures of interest such as the pharyngeal grinder and irrelevant structures generated represented the textures and boundaries of other tissues in the worm . To allow for supervised training of both the layer 1 and layer 2 classifiers , we annotated a selection of images ( n = 1 , 430 ) by manually identifying particles that represent the pharyngeal grinder . The classifiers can then be trained to associate properties of the feature sets with the manually specified identity of candidate particles . However , in addition to informative feature selection and the curation of a representative training set , the performance of SVM classification models is subject to several parameters associated with the model itself and its kernel function [34 , 48] . Thus , to ensure good performance of the final SVM model , we first optimize model parameters based on five-fold cross-validation on the training set ( Fig 3A and 3B , Materials and Methods ) . In the parameter selection process , the optimization metric can be designed to reflect the goals of classification in each layer ( Fig 3B ) . In our application , for the first layer of classification , the goal is to eliminate the large majority of background particles while retaining as many grinder particles in the candidate pool as possible for refined classification in the second layer . In other words , we aim to minimize false negatives while tolerating a moderate number of false positives . Therefore , we optimize the SVM parameters via the minimization of an adjusted error rate that penalizes false negatives more than false positives ( Fig 3B ) . We show that with an appropriate parameter selection , the first layer of classification can eliminate over 90% of background particles while retaining almost 99% of the true grinder particles for further analysis downstream ( Fig 3B ) . To visualize feature and classifier performance , we use Fisher’s linear discriminant analysis to linearly project the 14 layer 1 features of the training set onto two dimensions that show maximum separation between grinder and background particles ( Fig 3C ) . A high degree of overlap between the distributions of the grinder and background particles and high error rates associated with the trained SVM in this visualization suggest that shape-intrinsic features are insufficient to fully describe the grinder structure . Nevertheless , the first layer of classification enriches the true grinder structure candidates in the training set from roughly 6 . 2% of the original particle set to 40% of the particle set entering into the second layer of classification ( Fig 3C ) . This enriched set of candidate particles is used to optimize and train the second layer of classification in a similar manner ( Fig 3D ) . With appropriate parameter selection , we show that the second layer of classification is capable of identifying the grinder with sensitivity and specificity above 95% ( Fig 3E ) . We train the final layer 2 classifier with the reduced training set and these optimized parameters to yield high classification performance in combination with layer 1 ( Fig 3F ) . Changes in experimental conditions , the genetic background of the worms under study or changes to the imaging system , can cause significant variation in the features , and thus degrade the classifier performance due to overfitting that fails to take into account experimental variation ( Fig 3 ) . To account for this potential variability , we include worms imaged at different ages and food conditions in the training set of images . To validate the utility and efficacy of the resulting classification scheme in a real-life laboratory setting , we analyze its performance on new data sets that were not used in training the classifier . First , in spite of morphological changes due to experimental conditions ( Fig 4A ) , we show the resulting classification scheme operates with consistently high performance in distinguishing the head and the tail of the worm in the new data sets ( Fig 4B ) . Second , while the training set only includes wildtype worms imaged under different conditions , the morphology and texture of the worm is also subject to genetic alteration ( Fig 4C ) . To see whether our classification scheme can accommodate some of this genetic variability , we validate the classification scheme against a mutant strain ( dpy-4 ( - ) ) with large morphological changes in the body of the worm ( Fig 4C ) . Finally , changes in the imaging system can alter the digital resolution of biological structures of interest ( Fig 4E ) . We show that the inclusion of a calibration factor adjusting for the pixel to micron conversion of the imaging system is sufficient for maintain classifier operation across a two-fold change in the resolution of the imaging system ( Fig 4F ) . Thus , this calibrated classification scheme can be easily adapted to systems with different camera pixel formats via the calculation of a new calibration factor . An ever-expanding array of fluorescent markers and biosensors [6] has made the identification of specific fluorescent objects and patterns a common biological image processing problem . Although fluorescent staining or tagging techniques can be used to target structures or molecules of interest , they often cannot offer perfect specificity . Furthermore , biological specimens can also include autofluorescent elements that confound the analysis of fluorescent images . Thus , sifting relevant information from fluorescent images can pose non-trivial image processing problems where background fluorescent objects can have similar intensities or spatial locations . The usage of fluorescent tools in C . elegans is no exception . Existing toolsets permit fluorescent labeling of different genetic outputs of subsets of cells and tissues . However , fluorescent tags also often label multiple cells , cellular processes or tissue structures that must be distinguished to address specific biological questions . Moreover , C . elegans exhibits significant gut autofluorescence that varies in intensity and can obscure the identification of fluorescent targets throughout the length of the worm [44] . Here , we demonstrate the use of our framework to address these common challenges in fluorescent image processing , using neuron identification in the worm as a broadly useful example . We first focus on the identification of the ASI neurons as a stereotypical example of a bilaterally symmetric neuron pair in the worm . Fig 5B shows a corresponding set of bright field and fluorescent images illustrating the positioning of the neuron pair within the head region of the worm . In addition to the cell bodies of interest , the raw fluorescent image also shows cellular processes and autofluorescent granules in the gut of the worm that can confound cell-specific image analysis . Similar to our approach for pharyngeal grinder detection in Fig 2B , we begin building our cell identification toolset via preprocessing of the raw images by maximum intensity projection , Niblack thresholding and preliminary filtering of the resulting candidate particles ( Fig 5C , Materials and Exp . Methods ) . In the selection of features for both layers of classification , we note that the layer 1 feature set we developed for the detection of the pharyngeal grinder can be generally applied to the description of particle shape within other contexts ( S2 Fig ) . Using this feature set , we optimize and train a layer 1 SVM classifier using a manually annotated training set ( n = 218 ) ( S4A Fig , Materials and Methods ) and show that it is sufficient for identifying cellular regions with relatively high sensitivity and specificity ( Fig 5D and S4A Fig ) . While the first layer of classification is effective at eliminating the large majority of background particles , variable background intensity within the tissues surrounding the neurons can generate confounding binary particles that pass layer 1 classification ( Fig 6A ) . To make a final identification of a true cell pair , we apply a second layer of classification based on the relational properties of potential pairs of particles that pass layer 1 classification ( Fig 6B and S5 Fig ) . To construct our layer 2 classifier , we optimize and train an SVM model based on these pairwise relational features ( S4B Fig ) . We note that while the relational features we utilize are computationally simple , embedding relational features on the second layer of classification dramatically reduces the size of the paired candidate set . For example , for detection of cell pairs amongst n particles , there are ( n2 ) =n ! 2 ( n−2 ) ! possible candidate pairs that require feature calculation . Validating the resulting cell pair classifier against new test images , we find robust single cell-pair detection in the majority of cases ( Fig 6C , left ) . However , in a minority of cases , multiple candidate pairs are identified as potential neuron pairs in each image ( Fig 6C , right ) . This is a common scenario as many promoters used in transgene markers are not necessarily specific to a single class of cells . In this case , the probability estimates from the SVM classifier [37 , 49] along with the selection of the most likely candidate in images with multiple positive classification results is used to eliminate these false positives . This boosts the specificity of the classifier without compromising the high sensitivity ( Fig 6D ) . This additional step incorporates the real-world constraint that , at most , one cell pair exists in each valid image and resolves any conflicts that may arise in direct classification . To demonstrate the ability of our framework to detect more complex cellular arrangements , we use the expression pattern of a worm insulin-like peptide gene ( ins-6 ) in two bilaterally symmetric neuron pairs ( Fig 7A ) [50] . In this case , the specificity offered by the ins-6 promoter is insufficient to offer full cell specificity , requiring the identification of different cells from the raw fluorescent image . Taking advantage of our modular two-layer architecture , we reuse the preprocessing and first layer classification tools that we have already constructed to identify a small number of cell-shaped objects shown in Fig 7B . To detect the tetrad of cells with specificity for the ASI and ASJ neurons , we construct a relational feature set based on combinations of neuron pairs ( S6 Fig ) . As shown in Fig 7C , accounting for both correct cell pair identification and non-repetition of individual cells within the tetrad set , there are ( n2 ) ( n−22 ) =n ! 4 ( n−4 ) ! tetrad sets that require feature calculation . Our two-layer architecture is therefore essential for the construction of such relational feature sets with larger numbers of targets . Without layer 1 classification , description of such complex sets quickly becomes intractable: even 10 candidate particles generates 1 , 260 different possible tetrad sets for feature calculation . To construct a new problem-specific layer 2 classifier based on relationships within these tetrad candidates , we optimize and train a SVM model based on a manually annotated training set ( n = 324 ) ( S4C Fig ) . Subsequent validation of our two-layer classifier against new test images shows that the two-layer classification scheme operates with higher specificity but lower sensitivity in comparison to our single cell-pair classification problem ( Fig 7D ) . Further analysis of the classifier performance within the test set of images shows that this lower sensitivity is mainly due to more degrees of freedom for variability associated with this particular image processing problem . As shown in Fig 7E , while the second-layer classifier accommodates some deviation from the stereotypical arrangement of the neurons shown in Fig 7A ( positive identification on the left ) , there is a trade-off between maintaining specificity and sensitivity ( rejecting larger deviations as illustrated by the negative identification on the right ) . If the stringency is important , i . e . maintaining specificity and reducing misidentification rate , the users would have to tolerate a small amounts of false negatives . Users would need to determine a comfortable level of rejection rate for each specific problem to tune the classifier .
Beyond the specific applications we discuss here , we envision that our methodology can be a powerful way to tackle a broad range of biological image processing problems . For instance , we consider our scheme to be a generalization of the previously reported application of SVMs towards the understanding of synaptic morphology in C . elegans [24] . In this application , individual pixels within the image form the pool of candidates for potential synaptic pixels in the first layer classification . The second layer of classification then refines this decision on the basis of relational characteristics between candidates . Here , we formalize this classification approach and demonstrate that it can be adapted towards detection of disparate structures imaged under different imaging modalities . The imaging processing approach we present here has inherent structural advantages in terms of conceptual division , modularization and computational efficiency and demonstrates the application of a powerful supervised learning model to streamline biological image processing . We thus envision that our methodology can form the basis for detection algorithms for structures ranging from the molecular to the tissue or organismal level under different experimental methodologies .
C . elegans worms used in this study were maintained and cultured according to standard techniques [52] . Briefly , populations of worms were allowed to reach reproductive maturity and lay eggs on NGM agar media overnight . Age-synchronized worms were then obtained by washing free-moving worms off of the agar plate , allowing the remaining eggs to hatch for one hour and then washing the resulting L1 stage larvae off of the plate . Age-synchronized L1 worms were then transferred onto new NGM plates seeded with OP50 E . coli bacteria as a standard food source and grown until the desired age for imaging . To avoid over-crowding and food depletion , adult worms were transferred onto new plates daily . For starvation experiments , worms were transferred onto fresh NGM plates lacking a bacterial food source the day before imaging . C . elegans strains used in these studies were wild-type N2 worms , QH3833 dpy-4 ( e1166 ) , QL296 drcSi89[pdaf-7::GFP; unc-119 ( + ) ] and QL617 drcSi68[unc-119 ( + ) ; Pins-6::mCherry]II; gjIs140[dpy-20 ( + ) ; gpa-4::GFP] . We use standard soft lithographic techniques to produce polydimethylsiloxane ( PDMS ) imaging devices similar to those previously described [24 , 38] . For automated imaging , worms are washed off of NGM plates using S Basal buffer and introduced via pressure injection into the microfluidic device . Sequential activation of pressure sources driving liquid delivery and on-chip pneumatic valves is then used to drive individual worms within the device for imaging . Images were collected either on a Leica DMI 6000B microscope with a Hamamatsu Orca D2 camera and a 40X oil objective or on an Olympus IX-73 microscope with a Hamamatsu Flash 4 . 0 camera and a 40X oil objective . Relevant specifications and calibration metrics for these set-ups can be found in S1 Table . Although not strictly necessary , for generalizability in cases where the center of focus is adjusted to specific fluorescent targets and do not capture the pharynx well , a sparse three plane z-stack with a 15μm step size is used for bright field image acquisition . To fully capture neuronal cells , a dense z-stack was collected through the body of the worm . For fluorescence imaging of the single neuron pair in QL296 , a 0 . 4μm step size was used over a 60μm thick volume . For fluorescence imaging of multiple neurons pairs in QL617 , a 1μm step size was used over a 100μm thick volume . We use custom MATLAB code to perform all image preprocessing and feature extraction steps and enable the construction and testing of our classification schemes . In preprocessing , the three dimensional information in the acquired z-stacks were either maximum or minimum projected onto a single two-dimensional image for further processing . For bright-field images , a minimum projection with respect to z was utilized to accentuate the appearance of dark objects throughout the stack . Conversely , for fluorescence images , a maximum projection was utilized to accentuate the appearance of bright objects throughout the stack: In order to generate binary particles for classification , we use a local thresholding algorithm that uses information about the mean and variability of pixel intensities within a local region around a pixel: μlocal and σlocal are the means and standard deviations of all pixel values that fall within a square region of width 2R + 1 centered around the pixel of interest xi , yi and k is a parameter specifying the stringency of the threshold . μlocal and σlocal can be derived using standard image filtering with a binary square filter h ( xi , yj ) of width 2R + 1: Using local mean and standard deviation information in the binary decision affords robustness against local background intensity and texture changes . The width of the local region , R , can be roughly selected on the basis of the size scale of the structure of interest . In accordance with the size scales of the pharyngeal structure and individual neurons , we use R = 15μm for detection of the pharyngeal grinder and R = 5μm for fluorescent cell segmentation . The parameter k can be roughly selected by visual inspection of segmentation results . We use k = 0 . 75 for our bright field application and k = 0 . 85 for our fluorescence application . Individual candidate particles in the resulting binary image are defined as groups of nonzero pixels that are connected to each other via any adjacent of diagonal pixel ( 8-connected ) . We note that changes in k can alter the size of segmented particles and the connectivity of segmented particles . Particularly in bright field , where the contrast mechanism lacks specificity , decreases in k can cause particles to merge via small bridges of dark texture . In order to build in some robustness against changes in k and background texture in these scenarios , we perform a form of a morphological opening operation after thresholding to remove small bridges that may arise between otherwise distinct particles . To do this , we perform a morphological erosion with a small circular structuring element followed by a morphological dilation with a smaller structuring element [53] . In order to fully capture both intrinsic and secondary characteristics of biological structures , we calculate distinct sets of features for two layers of classification . The first layer , which delineates structures of interest from other structures on the basis on its intrinsic geometric properties , is generally applicable to particle classification problems and is used for both the bright field and fluorescent structure detection outlined here . Details and equations for the calculation of the 14 features for layer 1 classification can be found in S2 Fig . Secondary characteristics of biological structures describe the context in which structures exist and their relationship to other structures . Due to the large variability in the secondary characteristics of biological structures , a generic set of features is not necessarily attainable or desirable due to concerns for computational efficiency . Rather , secondary features can be derived via a mathematical description of empirical observations of important structural properties . In the case of pharyngeal grinder detection , the secondary features are regional , forming a description of the image context in which the grinder structure resides . The form of the features is based on an empirical understanding of this structural context and full details and equations for the calculation of the 34 features in layer 2 of the bright field classifier can be found in S3 Fig . In the case of cell pair detection , the secondary features are mostly relational , describing how particles from layer 1 of classification may or may not exist as pairs on the basis of both positioning and intensity . Second layer features for cell pair detection can be found in S5 Fig . We do briefly note that we scale all calculated features using a calibration factor , C , derived from specifications of both the optics and sensors that form the imaging system: The use of this calibration system renders the trained classifier relatively invariable to small changes in the imaging set-up via conversion of all features into real units . Calibration factors for all imaging systems and configurations used here can be found in S1 Table . To implement discrete classification steps using support vector machines , we use the LIBSVM library , which is freely available for multiple platforms including MATLAB [37] . For general performance , we train use a Gaussian radial basis function kernel for all of our trained classifiers [48] . To ensure performance of the SVM model for our datasets , we optimize the penalty or margin parameter , CSVM , and the kernel parameter , γ , for each training set using the five-fold cross-validation performance of the classifier as the output metric . For efficient parameter optimization , we start with a rough exponential grid search ( Fig 3B and 3D and S4 Fig ) and refine parameter selection with a finer grid search based on these results . To adjust for the relative proportions of positive and negative candidates in unbalanced training sets ( Fig 3C ) , we also adjust the relative weight , W , of the classes according to their representation in the training set while training [37] . Additionally , we perform a small grid search for the optimal weighting factor to fully optimize the following performance metric . Probability estimates for single and multiple neuron pair identification are derived according to the native LIBSVM algorithm [37] . For visualization of the high dimensionality feature sets ( Fig 3C and 3F ) , we apply Fisher’s linear discriminant analysis [54] . The two projection directions are chosen to be the first two eigen vectors of: SB is a measure of inter-class separation and SW is a measure of intra-class scatter . | New technologies have increased the size and content-richness of biological imaging datasets . As a result , automated image processing is increasingly necessary to extract relevant data in an objective , consistent and time-efficient manner . While image processing tools have been developed for general problems that affect large communities of biologists , the diversity of biological research questions and experimental techniques have left many problems unaddressed . Moreover , there is no clear way in which non-computer scientists can immediately apply a large body of computer vision and image processing techniques to address their specific problems or adapt existing tools to their needs . Here , we address this need by demonstrating an adaptable framework for image processing that is capable of accommodating a large range of biological problems with both high accuracy and computational efficiency . Moreover , we demonstrate the utilization of this framework for disparate problems by solving two specific image processing challenges in the model organism Caenorhabditis elegans . In addition to contributions to the C . elegans community , the solutions developed here provide both useful concepts and adaptable image-processing modules for other biological problems . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Automated Processing of Imaging Data through Multi-tiered Classification of Biological Structures Illustrated Using Caenorhabditis elegans |
Large-scale protein interaction networks ( PINs ) have typically been discerned using affinity purification followed by mass spectrometry ( AP/MS ) and yeast two-hybrid ( Y2H ) techniques . It is generally recognized that Y2H screens detect direct binary interactions while the AP/MS method captures co-complex associations; however , the latter technique is known to yield prevalent false positives arising from a number of effects , including abundance . We describe a novel approach to compute the propensity for two proteins to co-purify in an AP/MS data set , thereby allowing us to assess the detected level of interaction specificity by analyzing the corresponding distribution of interaction scores . We find that two recent AP/MS data sets of yeast contain enrichments of specific , or high-scoring , associations as compared to commensurate random profiles , and that curated , direct physical interactions in two prominent data bases have consistently high scores . Our scored interaction data sets are generally more comprehensive than those of previous studies when compared against four diverse , high-quality reference sets . Furthermore , we find that our scored data sets are more enriched with curated , direct physical associations than Y2H sets . A high-confidence protein interaction network ( PIN ) derived from the AP/MS data is revealed to be highly modular , and we show that this topology is not the result of misrepresenting indirect associations as direct interactions . In fact , we propose that the modularity in Y2H data sets may be underrepresented , as they contain indirect associations that are significantly enriched with false negatives . The AP/MS PIN is also found to contain significant assortative mixing; however , in line with a previous study we confirm that Y2H interaction data show weak disassortativeness , thus revealing more clearly the distinctive natures of the interaction detection methods . We expect that our scored yeast data sets are ideal for further biological discovery and that our scoring system will prove useful for other AP/MS data sets .
Insights into the architectures and mechanisms of cellular processes can be obtained by elucidation of genome-wide protein interaction networks ( PINs ) that describe the physical associations between the component proteins . Such maps , or interactomes , can be exploited to enhance many types of biological discovery including protein function prediction [1] , inference of disease genes [2] , and identification of condition-specific response modules [3] . The yeast Saccharomyces cerevisiae has been routinely employed as a model system for high-throughput studies and PINs have been determined using a number of platforms including yeast two-hybrid ( Y2H ) screens [4]–[6] , affinity purification followed by mass spectrometry ( AP/MS ) [7]–[9] , and protein-fragment complementation assays ( PCA ) [10] . Each approach perceives interactions in a distinct manner . The Y2H and PCA techniques detect direct binary interactions , although the PCA approach does not rely upon expression of a reporter gene as required in Y2H screens , while the AP/MS techniques purify and identify protein complexes . The reliability of each technique has been extensively debated in the literature and comprehensive analyses have resulted in contrasting conclusions [6] , [10]–[12] . However , it is generally accepted that any measure of reliability is not absolute and largely dependent on the nature of a pre-defined gold standard reference set . An additional complexity arises in the analysis , or interpretation , of an AP/MS data set because there is no standard , or well-defined , system to distinguish between the direct and indirect interactions present in a purified complex . The only information available for an individual purification is its composition: a tagged bait protein and associated co-purified prey proteins . Furthermore , the constituent proteins are identified by complex MS methods and different platforms often yield varying compositions for identical purifications [9] , [13] . Another concern is that the compositions of the purifications are influenced by the protein abundances [11] , [14] , [15] - proteins having a higher abundance are more likely to be detected in more purifications and , therefore , inferred to be involved in more interactions after tabulation of all bait-prey pairs [15] . To address these issues , a number of approaches for the analysis of AP/MS data sets have been employed [8] , [9] , [16] , [17] . These techniques have the common goal of discerning protein pairs that are appreciably co-purified relative to some random background . While each method determines scores representing the likelihood of observing two proteins together , the scores are computed using different procedures: Gavin et al . calculate log-ratios of observed co-occurrences relative to expected [8]; Krogan et al . utilize a combination of machine learning algorithms [9]; Collins et al . implement a supervised algorithm derived from Bayesian methods and optimized with empirically-derived parameters [16]; and Hart et al . determine interaction probabilities based on hypergeometric distributions [17] . The qualities of the generated PINs have been found to be superior to comparable data sets constructed by straightforward tabulations of bait-prey interactions [9] , [16] , [17] . These evaluations were generally deduced from direct comparisons against complexes manually curated by the Munich Information Center on Protein Sequences ( MIPS ) [18] . A recent study of high-throughput Y2H data sets explored the characteristic strengths and distributions of functional ( specific ) interactions and non-functional ( non-specific or transient ) interactions in order to assess the extent to which the latter impedes the formation of functional protein complexes [19] . It was conjectured that the overall impact upon biochemical efficiencies had evolved to a tolerable limit . Motivated by the use of randomization techniques as a tool to measure , or discover , enrichments of network motifs [20] and connectivity correlations [21] in complex networks , we developed a shuffling-based approach to assess the levels of interaction specificity detected in AP/MS data sets . This system allows for the computation of pair-wise protein co-occurrence significance ( CS ) scores by comparing experimentally observed numbers with those from randomized realizations . A CS score for two proteins provides a statistical measure of their propensity to co-purify , or interact , in an AP/MS data set . The approach requires no training set or machine learning and is , therefore , applicable to any AP/MS data set for any species regardless of whether any curated information exists or not . It is found that these AP/MS data sets contain significant enrichments of specific , or high-scoring , associations . Additionally , we showed that high-quality direct physical interactions curated in two prominent data bases have significantly high CS scores . Therefore , while the AP/MS data sets contain prevalent non-specific , or transient , associations , our scoring analysis reveals that there is an underlying preference for proteins to form selective , or discriminating , associations . Our resultant scored interaction data sets were further assessed by comparisons against four diverse , high-quality reference data sets , each representing a unique manner of interaction detection , association mechanism ( direct or co-complex ) , and/or curation . For most references , we found that the accuracies of our scored interaction sets were manifestly higher than those of previous studies . Additionally , our scored data sets are the only ones that typically outperformed experimental Y2H interaction sets [4]–[6] . A high-confidence PIN extracted from the AP/MS data of Gavin et al . [8] was revealed to be free of abundance effects while those derived from the data of Krogan et al . [9] contained weak abundance biases . Therefore , it would appear that in high-quality AP/MS data sets , interaction specificity is not coupled with protein abundance . We note that the converse has recently been found to be true of Y2H interaction data sets [19] . The high-confidence PIN derived from the data of Gavin et al . [8] was shown to be highly modular , containing many localized densely-connected regions , and strikingly different to a commensurate random network . We also demonstrated that the observed high modularity is not a result of misinterpreting indirect associations as direct interactions; rather , it is a result of direct physical associations . Furthermore , we suggest that the modularity in Y2H interaction data sets may be underrepresented as indirect associations in these PINs are significantly enriched with manually-curated physical interactions , i . e . , they are likely false negatives . The high-confidence AP/MS PIN shows assortative mixing , meaning that proteins having similar numbers of total interactions prefer to interact with each other . A consequence of assortativity is that high-degree proteins , or hubs , prefer to associate with each other rather than with proteins having very small numbers of total interactions . In agreement with a previous study [21] , we find that a consolidated Y2H PIN shows weak disassortative mixing while a manually-curated set of high-confidence physical binary interactions displays both , and in equal measure , assortative and disassortative mixing . Therefore , high-quality AP/MS data appear assortative while Y2H interaction data appear disassortative . We expect that our scored yeast data sets are ideal for further investigations involving biological discovery and that our procedure will prove useful for the analysis of current and future AP/MS data sets for a variety of species . We have compared our high-quality AP/MS interaction data sets with those from Y2H screens and perceived a number of novel insights regarding their substances and network properties . Certainly , their topologies are contrasting and must reflect their different methods of interaction detection .
A CS score is a measure of the propensity for two proteins to be identified together in purifications , either as bait-prey or prey-prey combinations , relative to what would be expected by chance . They were determined by comparing observed co-occurrences , the number of times two proteins coincided in purifications , with those from random simulations , where the latter were realized by thoroughly shuffling , or exchanging , prey proteins ( see below ) . Therefore , our CS scores are derived from a purely numerical procedure and , unlike previous systems of Krogan et al . [9] and Collins et al . [16] , require no training or reference data sets . Our CS scores are related to the socio-affinity indices of Gavin et al . [8] and the probabilistic scoring scheme of Hart et al . [17] in that they attempt to quantify the propensity for proteins to co-purify . The socio-affinity scoring system [8] uses log-ratios of actual co-occurrences relative to what would be expected based upon protein purification frequencies , while the probabilistic scoring scheme [17] calculates interaction scores based upon hypergeometric distributions . However , both of these methods use expected occurrence baselines determined from total numbers of protein populations or interactions . As such , they do not account for the great variations in bait affinities , i . e . , the observation that some bait proteins purify with very many preys while others purify with very few . Our procedure is distinct in that we determine numbers of expected , or chance , co-occurrences via constrained randomized simulations that preserve the individual purification structures , i . e . , the number of preys . Although simplistic in its nature , our scoring system is advantageous in several ways . First , the method generates co-occurrence distributions for each protein pair and , therefore , is able to gauge the statistical significances of the actual experimentally observed co-occurrences . Second , while the method penalizes proteins having higher frequencies of purification , or abundances , it is able to uniformly distinguish between specific and indiscriminate partnerships . In fact , the method is able to identify instances of negative associations , or protein pairs that have significantly under-represented observed co-occurrences relative to that expected . Third , our randomized simulations preserve the numbers of proteins in the individual purifications and , consequently , utilize the experimentally discerned affinities of the bait proteins . Last , as mentioned above , the procedure is purely numerical and does not require a training or reference data set . Therefore , it is completely devoid of any associated bias and is applicable to any affinity purification data set , regardless of whether any other high-confidence interaction sets exist or not . Our interaction detection based on shuffling ( IDBOS ) procedure is depicted in Figure 1A . For a given affinity purification data set in which individual purifications are specified by a bait protein and co-purifying prey proteins , we counted , for each unique protein pair i and j , the total number of times they co-occurred in the same purification . These observed co-occurrences , oij , do not distinguish between bait-prey or prey-prey combinations . We then constructed randomized , or shuffled , purification sets and computed average shuffled co-occurrences , ōij , and associated standard deviations , σij . The CS score for each protein pair was then determined as the Z-score of the observed co-occurrences: ( 1 ) A shuffled purification set was constructed by shuffling , or exchanging , pairs of prey proteins in a reference data set . A single realization was accomplished by enumerating all prey proteins ( in all purifications ) once and , for each prey protein , exchanging it with another prey protein chosen at random . However , an exchange was subject to the following constraints: ( i ) the two prey proteins must occur in different purifications , and ( ii ) the exchange cannot result in any purification having a protein that appears twice , whether as bait or prey . This construction procedure ensured that the shuffled purification sets were comparable to the experimental data set , whereby the numbers of proteins in the individual purifications were conserved and the global population of each protein remained unchanged . We constructed a million shuffled sets for each affinity purification data set analyzed here . An initial shuffled set was derived directly from the experimental purification data and subsequent shuffled sets were derived from ones immediately previous . When tabulating CS scores of protein pairs , or interactions , derived from an experimental affinity purification data set , we retained only those for observed co-occurrences greater than one , i . e . , oij>1 . We deemed that statistical significances of protein associations having co-occurrences less than two were not as reliable as those having higher co-occurrences . However , we stored mean shuffled co-occurrences and associated standard deviations computed from the million shuffled sets for all possible protein pairs . These were used to gauge the distribution of the tabulated CS scores through the following steps . First , an additional 105 shuffled sets were constructed in the same manner as that described above . Second , for each shuffled set , we determined the Z-scores for protein pairs having a shuffled co-occurrence of greater than one: ( 2 ) where ( >1 ) is the co-occurrence of proteins i and j in the nth shuffled set , and ōij and σij are the mean co-occurrences and standard deviations , respectively , determined from the million shuffled sets as in Equation ( 1 ) . The total shuffled distribution , comprising Z-scores accumulated from the 105 shuffled sets , was used as a baseline to contrast the distribution of CS scores . A standard way to evaluate an interaction data set is to contrast it against a reference set that is considered to be high quality . Commensurate with a previous approach [22] , we have computed accuracy versus coverage , where coverage is the number of coinciding interactions in the evaluated and reference sets and accuracy is the fraction of interactions in the evaluated set that are coincident . When an interaction data set included confidence scores , as in the sets derived in this work and in previous studies [8] , [9] , [16] , [17] , we ranked the interactions by decreasing score and plotted accuracy versus coverage curves over a range of score cutoffs . We used four reference interaction data sets that are each considered to be high quality in some way . However , they are also individually distinct in that each represents a different style of interaction measurement or curation . By evaluating , or contrasting , interaction data sets against these references , we were able to assess their substances from a number of viewpoints . Descriptions of the reference sets follow: We also analyzed yeast PINs determined from a number of high-throughput Y2H screens in order to contrast their contents and network structures against the scored AP/MS data sets . The Y2H data sets studied included the interaction sets of Yu et al . [6] ( CCSB-YI1 ) , Ito et al . [4] ( core subset ) , Uetz et al . [5] , and a union of these sets [6] ( Y2H-union ) . The network structures of protein interaction data sets were analyzed by computing a variety of graph-theoretical properties . The clustering coefficient of a node ( or protein ) i is defined as the fraction of possible edges between neighbors that are present , where a neighbor of node i is any other node that shares an edge with it [25] . The average clustering coefficient of a network was determined by averaging the clustering coefficients of all nodes , where nodes involved in only one interaction are defined here to have a clustering coefficient of zero . The clustering coefficient of a network is an indication of the network's modularity , although it is not a strict measure . The nature of the connectivity in a network was assessed here by determining interaction frequencies between pairs of degrees , i . e . , for two degrees k1 and k2 , we counted the total number of interactions occurring between two nodes where one has degree k1 and the other has degree k2 . Enrichments of interaction frequencies between degrees were measured as Z-scores , where actual numbers were compared to those of commensurate , randomly-rewired , degree-preserving networks ( 103 realizations ) that were constructed using a similar procedure to that of Maslov and Sneppen [21] . To verify our interpretation of the interaction frequencies with regards to the connectivity in a network , we also computed the degree-degree correlation coefficient [26] , [27] , which quantifies the level of interaction between proteins of similar degrees: ( 3 ) where the averaged quantities are determined over all interactions and the denominator is the variance of the node degree k . When nodes of similar degrees prefer to interact in a network , i . e . , their interaction frequencies are significantly enriched resulting in a positive degree-degree correlation coefficient ( r>0 ) , then the network connectivity is said to be assortative – nodes of high degree ( hubs ) prefer to interact with each other while low-degree nodes avoid interacting with hubs . Conversely , when nodes of diverse degrees prefer to interact in a network , leading to a negative correlation coefficient ( r<0 ) , then the connectivity is said to be disassortative – hubs avoid each other and generally prefer to interact with low-degree proteins .
We applied our IDBOS scoring procedure ( Figure 1A and see Materials and Methods ) to the yeast AP/MS experimental data sets of Gavin et al . [8] and Krogan et al . [9] . Gavin et al . used matrix-assisted laser desorption/ionization-time of flight ( MALDI-TOF ) MS to identify proteins present in the purification while Krogan et al . used two MS techniques for protein identifications: MALDI-TOF and liquid chromatography tandem MS ( LCMS ) . Although previous studies have merged the MALDI-TOF and LCMS data sets of Krogan et al . , we chose to keep them separate initially . Therefore , we computed three sets of CS scores for each of the Gavin , Krogan ( MALDI-TOF ) , and Krogan ( LCMS ) AP/MS data sets that formed our IDBOS-Gavin , IDBOS-Krogan ( MALDI ) , and IDBOS-Krogan ( LCMS ) scored interaction data sets , respectively ( Tables S1 , S2 , S3 ) . Only CS scores for protein pairs having total co-occurrences greater than one were retained . As discussed above ( see Materials and Methods ) , the CS score for a protein pair represents the propensity for them to co-purify ( or associate ) relative to a random background derived from simulations that shuffled prey proteins . We illustrate the approach for the two proteins Tub1 ( YML085C ) and Tub2 ( YFL037W ) that are known to form alpha and beta subunits of heterodimers that polymerize to form microtubules . These cytoskeletal filaments participate in a variety of cellular functions , including structural support [28] . The significance of the Tub1–Tub2 associations in the AP/MS data set of Gavin et al . , i . e . , the CS score in the IDBOS-Gavin data set , is shown in Figure 1B . The random profile has a mean co-occurrence of 61 . 2 and a standard deviation of 6 . 0 , indicating that the observed co-occurrence , or frequency of co-purification , at 156 is statistically significant with a Z-score of 15 . 8 . Therefore , we would consider that Tub1 and Tub2 have a high affinity of association . In contrast , previous analyses of the Gavin et al . data set have not concluded that these two proteins have a significant association [8] , [16] . In fact , only the study of Hart et al . [17] infers a significant association for these two proteins; however , some of their scores are computed by multiplying P-values across data sets . Curiously , the interaction between Tub1 and Tub2 has not been identified in any of the high-throughput Y2H or PCA screens [4]–[6] , [10] . Perhaps a more intriguing illustration of our approach is the discerned highly-specific non-interaction , or perceived repulsive association , between the two proteins Ssa1 ( YAL005C ) and Ssa2 ( YLL024C ) . These proteins have an experimentally observed co-occurrence of 65 in the AP/MS data set of Gavin et al . while the random profile has a mean-co-occurrence of 187 . 6 and a standard deviation of 6 . 6; therefore , the resultant CS score is considerably negative at −18 . 7 ( Figure 1C ) . This score implies that not only do these proteins not interact; they would rather not associate , even by chance . The reasons for this inferred repulsive association are not immediately clear . Ssa1 and Ssa2 are cytosolic members of the heat shock protein 70 family that have a number of functions , including serving as molecular chaperones and assisting in protein folding [29] . A possible explanation for their avoidance may be to enhance their protein translocation efficiencies – if they were to come together , even by chance , their individual abilities to function as chaperones may be lost . It is also possible that Ssa1 and Ssa2 interact with diverse sets of proteins , i . e . , Ssa1 may interact strongly with a particular set of proteins whereas Ssa2 may interact with a different group . While there has been much focus recently on elucidating the high-confidence or steadfast interactions in experimental interaction data sets , little effort has been made to identify proteins that strongly avoid each other . It remains to be seen whether this latter type of non-interaction amongst proteins is also fundamental for normal cellular function . Although the random co-occurrence profiles for the Tub1–Tub2 and Ssa1–Ssa2 cases discussed above appear to be normally distributed ( Figures 1B and 1C ) , it should be noted that as the average random co-occurrence for two proteins approaches zero , the corresponding random co-occurrence profile will become less normal and skewed to the right . Therefore , one may query the reliability , or appropriateness , of CS scores in such instances . We have somewhat diminished this concern by only scoring protein associations that have an observed co-occurrence of two or more ( see Materials and Methods ) . However , we recognize that in some instances random co-occurrence profiles will deviate from normality . Nonetheless , as a starting point for more advanced ( and possibly computationally inefficient ) approaches , we analyzed the performance of the current procedure . The number of potential protein pairs in an AP/MS data set is very large , in the millions for the three analyzed in this work . As such , it is possible for pairs to have significant scores for their co-occurrences purely by chance . To investigate this likelihood we contrasted the distributions of the CS scores in the three IDBOS data sets against shuffled , or random , score distributions accumulated from 105 commensurate shuffled sets ( see Materials and Methods ) . We found that the experimental distributions have longer tails in the high-score region ( Figure 1D ) , indicating that they are enriched with discriminating protein associations . These results are encouraging in that they reveal , in a unique way , perceptible levels of specificity in the associations detected by the AP/MS experiments . Furthermore , all three random distributions are nearly identical , indicating that we are using consistent random baselines in our approach . We note that of the three experimental distributions , the IDBOS-Gavin data set has the most pronounced enrichment in the high-score region , possibly suggesting differences in the qualities of the experimental data . This issue is discussed in more detail later . Careful examinations of the randomized Z-score distributions indicate that they deviate slightly from normality in that they are slightly skewed to the right . This is most likely a result of only scoring interactions that have a co-occurrence of two or greater in any of the experimental or the additional 105 randomized data sets , i . e . , for a given data set , whether experimental or one of the additional randomized , Z-scores were only determined for protein pairs that had co-occurrences of two or greater in that data set ( see Materials and Methods ) . Therefore , the experimental and random score distributions are slightly skewed to the right . Even so , when contrasted against the random score distributions , the experimental distributions are noticeably enriched in the high-score region . As a first step to analyzing the reliability of our scoring scheme , we gauged the scored interactions in the IDBOS-Gavin data set by mapping them on to curated interactions that represent high-confidence associations identified in small-scale , or low-throughput , experiments . For a given curated data set , we tabulated their IDBOS-Gavin scores , i . e . , we accumulated IDBOS-Gavin scores for interactions that occurred in both the curated data set and our IDBOS-Gavin scored set . If the curated set contains steadfast interactions and our procedure is able to identify them as being statistically over-represented in the AP/MS data set of Gavin et al . , then the accumulated score distribution should reflect this . Indeed , we discovered that interactions in two prominent curated sets have distinctively high CS scores in the data set of Gavin et al . ( Figure 1E ) . The first curated set is a collection of interactions between proteins occurring in the same MIPS annotated complex ( see Materials and Methods ) and this data shows two peaks near CS scores of five and twenty . The distribution about five may be due to the nature of the interaction tabulation . We inferred that all proteins occurring in the same MIPS complex interact; however , most likely many of these pairs do not have a direct physical association . The second curated set is a collection of manually-curated physical interactions reported twice or more in the SGD-BioGRID repositories ( see Materials and Methods ) . This set of interactions ( SMBC2 ) has a CS score distribution that is also well separated from the total experimental and shuffled distributions and , like the MIPS data , exhibits a peak near twenty . Therefore , we concluded that our IDBOS scoring scheme was able to reliably distinguish between the specific and non-specific associations detected in the AP/MS experiments . To further evaluate the IDBOS procedure we compared its performance against the previously described scoring systems of Collins et al . [16] and Hart et al . [17] by contrasting each against a variety of reference interaction sets . Both systems of Collins et al . and Hart et al . have been shown [16] , [17] to out-perform the high-confidence PINs derived in the original AP/MS studies [8] , [9] . Collins et al . provide purification enrichment ( PE ) scores computed independently for the AP/MS data sets of Gavin et al . [8] and Krogan et al . [9]; however , they analyze the latter by combining the original MALDI-TOF and LCMS purifications into one data set . Hart et al . [17] only provide scores determined by multiplying individual results across the Gavin et al . [8] , Krogan et al . [9] , and Ho et al . [7] data sets . Since consolidated data sets generally show greater accuracy than individual ones [16] , [17] , we felt that comparison of IDBOS-Gavin and Collins-Gavin interaction data against the combined data of Hart et al . [17] advantaged the latter . Accuracy versus coverage curves using four diverse reference sets are shown in Figures 2A–D ( see Materials and Methods for fuller descriptions of the references and evaluation procedure ) . The first two references represent high-quality direct physical interactions that were either curated binary gold standard ( BGS ) [6] ( Figure 2A ) or detected in a large-scale experiment ( PCA ) [10] ( Figure 2B ) . In each instance , we found that IDBOS-Gavin scored data performed better than the Collins-Gavin and Hart data sets . Similar results were obtained for the third reference ( Figure 2C ) , which consists of manually-curated physical interactions detected in small-scale experiments ( SBMC2 ) [23] , [24] . These results suggest that our method was more adept at discerning the direct associations from the indirect that are present in the purifications . The fourth reference is a collection of interactions between proteins co-occurring in MIPS curated complexes identified in low-throughput experiments . All three scoring schemes show very high overlaps ( Figure 2D ) and this is probably not unexpected . By assuming that all proteins comprising a complex are interacting , we are not distinguishing between the direct and indirect associations . However , the results are encouraging for the IDBOS and Hart et al . [17] approaches as neither relies upon external data , while the method of Collins et al . [16] employed empirical parameters that were optimized using MIPS complexes . Very similar results were observed when analyzing the IDBOS- and Collins-scored data of Krogan et al . ( Figure S1 ) . While our technique for the analysis of AP/MS data sets compares favorably with previous methods , it is of interest to contrast our scored interaction data sets against those from high-throughput Y2H screens [4]–[6] . It has recently been surmised that AP/MS methods are best at detecting co-complex associations while Y2H screens are better at detecting binary interactions when compared against the BGS set [6] . When using this BGS set as a reference , we found that the Y2H interaction sets show better relative accuracies than the Collins-Gavin and Hart data sets . However , our IDBOS-Gavin data set performed at a slightly higher level than the Y2H interaction sets ( Figure 2A ) , although the differences are small . Nonetheless , the result further affirmed that the IDBOS procedure discerned direct physical associations in the AP/MS purification data . The IDBOS-Gavin set performed markedly better than the Y2H data sets for the other references ( Figures 2B–D ) . The results are not unexpected when using the MIPS reference , but noteworthy for the others as they represent distinct types of high-quality direct binary interactions . Although the IDBOS-Krogan data is of slightly poorer quality than the IDBOS-Gavin data , the comparisons against the Y2H interaction sets yielded comparable results ( Figure S1 ) . We determined score cutoffs for each IDBOS data set by comparisons of their experimental and random score distributions ( see Materials and Methods ) shown in Figure 1D . For a given score threshold ζ , we can compute the fractions of protein pairs in the commensurate random and experimental distributions that have a higher score as fR ( Z>ζ ) and fE ( CS>ζ ) , respectively . Therefore , we approximated the false-discovery rate as the ratio of these fractions , i . e . , PFP ( ζ ) = fR/fE . We used a false-discovery rate of 5% to compute score cutoffs for the IDBOS-Gavin ( ζ0 . 05 = 5 . 95 ) , IDBOS-Krogan ( MALDI ) ( ζ0 . 05 = 8 . 26 ) , and IDBOS-Krogan ( LCMS ) ( ζ0 . 05 = 12 . 92 ) data sets . Corresponding high-confidence PINs were compiled by including only interactions having higher CS scores than the respective cutoffs . The number of proteins/interactions in the IDBOS-Gavin , IDBOS-Krogan ( MALDI ) , and IDBOS-Krogan ( LCMS ) PINs were 1274/7879 , 1061/3398 , and 1719/3640 , respectively . The IDBOS-Gavin PIN has the largest number of interactions of the three , which demonstrated the superior enrichment of high CS scores in the AP/MS data set of Gavin et al . [8] . The IDBOS-Krogan ( LCMS ) PIN is the sparsest , as judged by the average number of interactions , or degree , of the constituent proteins , implying that the LCMS data of Krogan et al . [9] has the lowest enrichment of significant association scores . Certainly , these observations are mirrored by the order of the computed score cutoffs given above . From the results presented so far , one might conclude that of the three AP/MS data sets investigated here , the set of Gavin et al . [8] showed the highest specificity of protein associations: ( i ) it had the most considerable enrichment of high CS scores ( Figure 1D ) and , consequently , it yielded the most interactions from use of a 5% false-discovery-rate filter; and ( ii ) the IDBOS-Gavin and Collins-Gavin scored data sets generally showed superior performance over the comparable scored sets derived from the data of Krogan et al . [9] ( Figure 2 and Figure S1 ) . We investigated this premise further by analysis of protein abundance trends in the high-confidence PINs derived in this work and by Collins et al . [16] . It has previously been demonstrated that proteins having higher cellular abundances tend to be involved in more interactions , or have higher degrees , in AP/MS experimental data sets; such an abundance-degree relationship is not present in PINs determined from Y2H screens [11] , [14] , [15] . Abundance effects were assessed using an approach similar to that of von Merring et al . [11] , whereby proteins in a PIN were sorted into classes according to their abundances . We utilized the recent abundance measurements of Newman et al . determined from flow cytometry [30]; however , similar results were observed when using abundances measured by western blot analysis [31] ( data not shown ) . We found that the IDBOS-Gavin PIN is free of any abundance effects while the IDBOS-Krogan ( LCMS ) PIN shows a weak bias in the high-abundance/high-degree region ( Figure 3A ) . Equivalent results were obtained when we analyzed the high-confidence networks of Collins et al . [16] ( Figure 3B ) , which were each constructed using the score cutoff of 3 . 19 used for their merged data . Like our IDBOS-Gavin PIN , the high-confidence Collins-Gavin network shows no significant abundance effects . We could only construct a merged Collins-Krogan ( MALDI+LCMS ) PIN from their available data and this network shows the largest high-abundance/high-degree bias . The observation that only the high-confidence PINs derived from the results of Gavin et al . [8] are free of any abundance bias is consequential . This finding , together with those discussed earlier , imply that the AP/MS experiment of Gavin et al . [8] detected more specific protein associations than that of Krogan et al . [9] . For the latter study , the score-enrichment and abundance analyses described above indicate that the MALDI-TOF method identified more specific associations than the LCMS technique . However , we do not wish to make firm conclusions regarding the two identification methods . There are other important factors that we have not considered , not least that the LCMS method is purported to be more successful in identifying small and lower-abundance proteins [9] , [13] . Such an advantage might certainly lead to a perceived lower-specificity , at least by the analysis methods used here , simply because more unique proteins may be detected . Although the primary focus of the present article is the description and analysis of the IDBOS scoring procedure for AP/MS data , it is useful to examine the network structures of the derived high-confidence PINs . Since our evaluations suggest , but certainly not affirm , that the AP/MS data of Gavin et al . [8] contains more specific protein associations than the data sets of Krogan et al . [9] , we opt to present network analyses of the high-confidence IDBOS-Gavin PIN described above; however , the IDBOS-Krogan PINs show very similar characteristics . The IDBOS-Gavin PIN is depicted in Figure 4A and its modular nature is immediately apparent . We want to make it clear that in this work we have strictly not quantified the levels of modularity in any network . Rather , we have inferred modular natures , or lack of , via a number of graph-theoretical analyses and illustrations . While a refined two-dimensional portrayal of a network can reveal the inherent modularity , it often also disperses modules that are incorporated in the giant component . Nonetheless , it is clear from Figure 4A that the IDBOS-Gavin PIN contains many localized highly-clustered regions as well as numerous disjoined complexes . The IDBOS-Gavin PIN is strikingly different to a commensurate randomly-rewired , degree-preserving network ( constructed using a similar procedure to that of Maslov and Sneppen [21] ) , which shows no modularity or disjoined regions ( Figure 4B ) . Interaction data sets generated from the raw AP/MS data of Gavin et al . [8] , using the spoke ( bait-prey tabulation ) and matrix ( bait-prey and prey-prey tabulation ) models , appear very similar to the random in that they exhibit very little modularity and appear uniformly dense ( Figure S2 ) . While the number of disjoined components in a network , relative to that of a commensurate random network , is not a strict measure of modularity , it does provide insight into the level of interaction localization . The IDBOS-Gavin PIN has 90 disjoined components compared to the expected 1 . 95 ( SD = 0 . 8 ) for the random equivalent based on 1000 realizations ( Figure 4C ) . This substantial 46-fold increase clearly indicates preferential protein complexation . In contrast , the Y2H interaction networks show no significant enrichments of disjoined components with observed/expected ratios of close to one ( Figure 4C ) . Therefore , the IDBOS-Gavin PIN shows a much higher level of selective complexation than the Y2H data sets . While this is to be expected due to the nature of the AP/MS method , the results imply that the IDBOS scoring procedure was able to identify individual complexes occurring in the purification data . Another indicator of the level of modularity in a network is the average clustering coefficient of a network which is literally a measure of edge clustering around the nodes or proteins [25] . We averaged clustering coefficients over proteins in a PIN having degrees greater than one . The IDBOS-Gavin PIN has an average clustering coefficient of 0 . 74 and this is much higher than those for the Y2H interaction data sets ( Figure 4D ) . This 19-fold ratio of observed relative to commensurate random suggests a significant enrichment of clustering . Of the Y2H interaction sets , the Uetz et al . [5] data set has the highest ratio of observed/expected of 14 while the core PIN of Ito et al . [4] has the lowest of approximately one . Therefore , the Y2H PIN of Uetz et al . [5] also shows a significant clustering enrichment . Figure 4D shows the average clustering coefficients of proteins by degree for the IDBOS-Gavin PIN and two realizations of a commensurate random network . It is clear that the clustering tendency of a protein in the IDBOS-Gavin PIN is essentially independent of its degree , only dropping slightly at very high degrees , and is substantially higher than the random . The clustering profile in the IDBOS-Gavin PIN is manifestly different from the power-law profile of hierarchical networks previously proposed to model biological networks having power-law-like degree distributions [32] , [33] . Although the IDBOS-Gavin PIN is also characterized by a power-law-like degree distribution ( see High-Confidence AP/MS PINs Show Assortative Mixing ) , it is clear that this network does not have a hierarchical structure ( Figures 4A and D ) and that the IDBOS scoring procedure is discerning an inherent modular nature for the preferential protein interactions in the AP/MS purifications . It has previously been concluded that , generally , Y2H interaction sets consist of high-quality direct binary associations while AP/MS data sets contain complexes composed of direct and preponderant indirect associations [6] . Therefore , it is possible that our scoring system assigns artificially high scores to pairs of proteins occurring in the same complex , but that are not directly physically interacting . We assessed the scope of these misrepresented indirect associations in our high-confidence IDBOS-Gavin PIN by contrasting , via accuracy versus coverage curves , the weakest links in the modules against the manually curated BGS set of high-confidence physical binary interactions that represent direct protein associations rather than indirect ones [6] . Modules , or highly interconnected regions in a network , can be considered to contain enrichments of triangles in which three nodes are completely interconnected . The IDBOS-Gavin PIN contains 43 , 054 triangles , 17 times more than that in a commensurate random network ( averaged over 1000 realizations ) . The weakest link in a triangle is the interaction having the lowest CS score . As such , the weakest links in the IDBOS-Gavin PIN are good candidates for possible indirect associations . We compiled all the weakest links , and their corresponding CS scores , in the IDBOS-Gavin PIN and evaluated this interaction subset against the BGS set ( Figure 4E ) . While this weakest-link IDBOS-Gavin set performs slightly worse than the complete IDBOS-Gavin scored data ( Figure 2A ) , it is of very similar quality to the Y2H data sets ( Figure 4E ) . Therefore , the weakest links in the IDBOS-Gavin PIN most likely represent direct interactions , indicating that the observed modularity is not an artifact arising from misrepresenting indirect associations as direct interactions . We next turned our attention to the undetected interactions in the experiments . For Y2H data sets , they denote protein pairs that did not restore a transcription factor activating expression of a reporter gene , while in our analysis of AP/MS data they represent non-specific , or low-scoring , protein associations in the purifications . A false negative is here defined as an undetected interaction that is curated as a direct physical interaction in a reference set . The BGS and SBMC2 curated data sets were considered to be appropriate references ( see Materials and Methods ) . Good candidates for false negatives are undetected associations between two proteins who share an interaction partner , i . e . , indirect associations arising from cases of A–C–B , where two proteins A and B are not found to associate but both are evinced to interact with protein C . The fraction of these indirect associations that are false negatives ( actual ) was compared with the fraction of all undetected interactions that are false negatives ( expected ) . Enrichments were computed as ratios of actual/expected . Enrichments for the IDBOS-Gavin and Y2H PINs were greater than three ( Figures 5A–B ) ; therefore , the results suggest that in all these data sets indirect associations are more likely to be false negatives , at least as categorized by the BGS and SMBC2 references . The enrichment is least for the IDBOS-Gavin PIN but substantial for the Y2H interaction sets with the data of Uetz et al . [5] showing the largest proportion of possible missed interaction detections . These findings imply that the high modularity observed in the IDBOS-Gavin PIN was not a result of misrepresenting indirect associations as direct interactions and , in fact , indicate that the modularity would be enhanced if curated high-quality binary interaction data was included . The results affirm that the IDBOS scoring procedure is able to adequately distinguish between the direct and indirect associations in the purification data . We also found that the modularity in Y2H interaction data sets may be underrepresented , particularly the data of Uetz et al . [5] , as the constituent indirect associations were significantly enriched with false negatives . It must be stressed that these inferences were largely based on the assumption that the BGS [6] and SBMC2 [24] data sets comprise veritable direct binary physical protein interactions . We note that the BGS data has recently been utilized to demonstrate that the qualities of high-throughput Y2H data sets are substantially better than those of high-throughput AP/MS data sets [6] . Having established that the observed modularity in the high-confidence IDBOS-Gavin PIN is likely a result of direct interactions , we probed the network features further . As noted above , the IDBOS-Gavin PIN has a power-law-like degree distribution that is substantially different from that of a completely random Erdös-Rényi graph having the same number of nodes and edges ( Figure 6A ) . The observed non-random degree distribution is not surprising , but welcome , since it is well established that many real-world networks , including biological , have power-law-like degree distributions [26] , [33] , [34] . With respect to biological networks and PINs , previous studies have found that they are disassortative [21] , [26] , meaning that interactions tend to occur between two nodes , or proteins , that have very different degrees , i . e . , hubs , or proteins having very many interactions , prefer to connect to proteins having very few interactions . A consequence of disassortativeness is that hubs avoid interacting with each other and prefer to spread out in a PIN rather than clump together centrally . We investigated the connectivity in the IDBOS-Gavin PIN by computing interaction frequencies for pairs of degrees . The significances of the frequencies , computed as Z-scores illustrated in Figure 6B , were evaluated by comparison against frequency distributions resulting from 1000 realizations of commensurate , degree-preserving random networks . The diagonal nature of the degree-degree frequency distribution is immediately apparent . High-degree proteins prefer to interact with each other while low-degree proteins avoid interacting with hubs . In fact , the IDBOS-Gavin PIN appears highly assortative - interactions tend to occur between proteins having very similar degrees . To confirm this property , we evaluated the degree-degree correlation coefficient ( −1≤r≤1 ) , whereby a negative value indicates disassortativeness and a positive value signifies assortativeness [26] , [27] . As expected , the IDBOS-Gavin PIN has a considerably positive correlation coefficient of 0 . 62 , confirming its inherent assortative nature . For comparison , a commensurate degree-preserving random network has an average correlation coefficient of −0 . 02 ( SD = 0 . 01 ) for 1000 realizations; this value is slightly negative due to the exclusion of self interactions . The previous finding of disassortativity [21] was based on a study of the Y2H interaction data of Ito et al . [4] . Significances of degree-degree interaction frequencies for the Y2H-union data set [6] are shown in Figure 6C and it is clear that this network contains weak disassortative mixing ( r = −0 . 08 ) - hubs generally prefer to interact with low-degree proteins and there is only a slight diagonal propensity . Therefore , we confirm the previous finding [21] that Y2H interaction data appears disassortative while high-quality AP/MS interaction data constitutes significant assortative mixing . These findings are in line with the observations noted above , whereby the modularity in the IDBOS-Gavin PIN is likely due to direct interactions while the modularity in Y2H data sets may be underestimated due to missed interaction detections . This inference is reflected in the significances of the degree-degree interaction frequencies in the manually-curated BGS set [6] shown in Figure 6D , where significant simultaneous disassortative and assortative elements result in an ‘X’ pattern . In fact , the degree-degree correlation coefficient for this interaction data was essentially zero ( r = 0 . 004 ) , indicating that the disassortative and assortative mixing effects are nearly identical .
We have developed a statistical approach to measure the affinity for two proteins to co-purify in an AP/MS data set . The method is not based on machine-learning techniques and , therefore , requires no external reference data set . As such , it is applicable to any current and future AP/MS data set regardless of how much curated information is available . Our scoring mechanism is distinct from previous approaches in that it utilizes random baselines derived by thoroughly shuffling , or exchanging , prey proteins . Therefore , the approach preserves the numbers of proteins in the individual purifications and takes into account the experimentally discerned affinities of the bait proteins . The procedure was applied to two recent yeast AP/MS studies [8] , [9] and it was shown that the derived scored interaction data sets were enriched with specific , or discriminating , protein associations as compared to random profiles . It was also demonstrated that known high-quality direct physical interactions had significantly high scores . The scored interaction data sets were further evaluated by comparisons against four diverse high-quality reference data sets and it was generally found that our scoring system performed superior to previous scoring schemes [16] , [17] . Additionally , our scored interaction data sets were the only ones that almost consistently outperformed experimental Y2H interaction sets [4]–[6] , including when contrasted against the curated BGS set which represents high-confidence direct physical binary associations [6] . Although it is generally accepted that AP/MS experiments detect preponderant non-specific ( transient ) protein interactions , our analyses reveal an underlying specificity for protein associations , i . e . , a subtle preference for proteins to form functional interactions . While ours and previous studies [8] , [9] , [16] , [17] have implied such a specificity by showing that high-scoring associations generally appear in manually curated reference sets , we have further demonstrated that the experimental score distributions are distinct from commensurate random profiles . The random profiles for three different AP/MS data sets are almost identical revealing a consistency in our scoring approach . Additionally , the experimental score distributions have enhanced tails in the high-score region , thereby demonstrating enrichments of interaction specificity in each experimental data set . The interplay between non-functional and functional interactions was recently explored using Y2H interaction data and it was conjectured that the impact of non-functional interactions upon biochemical efficiencies of specific complexes was near the tolerable limit [19] . Since our analyses of AP/MS data provide specificity profiles , we hope that our scored interaction data sets may reveal further insights into the non-functional/functional interaction dynamics occurring in the cell . From the scored data sets we derived , using 5% false-discovery rates , corresponding high-confidence PINs . We selected that derived from the AP/MS data of Gavin et al . [8] for further network study after inferring that it contained the highest specificity . We stress that our determination of specificity in the AP/MS data sets was based on our score-enrichment and abundance analyses and did not consider other mitigating factors . Therefore , we are reluctant to make firm conclusions regarding the data sets of Gavin et al . [8] and Krogan et al . [9] . Our high-confidence PIN derived from the data of Gavin et al . [8] was shown to be highly modular and strikingly distinct to a commensurate random degree-preserving network . Additionally , we demonstrated that the high modularity was not a consequence of misinterpreting indirect associations as direct interactions . We propose that the lack of modularity in Y2H PINs is the result of enrichments of false negatives due to undetected interactions between indirectly associating proteins . In line with these findings , the network structures of our high-quality AP/MS and Y2H PINs were found to be significantly different - our AP/MS PIN shows strong assortative mixing while Y2H PINs show weak disassortative mixing . A consequence of assortative mixing in AP/MS data sets is that high-degree proteins ( hubs ) prefer to interact with other high-degree proteins; however , the disassortative mixing in Y2H PINs means that hub proteins avoid each other and instead connect to low-degree proteins . As a result of the network connectivity differences , our high-quality AP/MS data set appears more modular than Y2H interaction sets . However , the curated BGS set shows both , and in equal measure , assortative and disassortative mixing , suggesting that both elements are actually present in comprehensive cellular interaction networks . It remains to be seen whether the enriched levels of specificity observed in the yeast AP/MS data sets also exist in AP/MS data for other organisms , particularly those that do not have multiple compartments . The modular nature of the specificity discovered here for the yeast AP/MS data indicates a clear biological propensity for the formation of individually functioning complexes . Maximum insights into the nature of this selective clustering will be gained by mapping biological properties of the proteins , such as function and compartment locality , upon the scored interaction data . While we have carried out such analyses , these results will be presented and discussed at a later time . Previous studies of AP/MS data suggest that high-confidence interactions most likely occur between proteins having the same function and locality [16] and we can confirm that the modules involve proteins of similar function ( results not shown ) . Therefore , the observed assortative mixing by degree also exists for biological function . Comprehensive analysis of the mixing patterns by function and compartment in high-quality AP/MS and Y2H PINs should yield further insights into the natures of interaction detection of both platforms . We anticipate that our scored yeast data sets will be valuable for further biological discovery and that our technique will be useful for the analysis of current and future AP/MS data sets for a variety of species . | To understand and model cellular processes , we require accurate descriptions of the interactions occurring between constituent proteins . Large-scale protein interaction maps have typically been measured in two distinct ways . The first detects direct pair-wise associations by testing only two proteins at a time for an interaction . The second detects large groups of proteins that have conglomerated or purified together . With regard to the latter , it is difficult to deduce which pairs of proteins are physically interacting in the purification data , and interaction maps generally appear random and unstructured . We have developed a novel computational method to analyze the purification data ( from the second method ) and identify which proteins are directly interacting . The resultant protein interaction map is highly modular , meaning that the proteins organize themselves into localized , densely connected regions that likely represent individually functioning units . We also analyzed interaction maps of the first method and propose that their lack of modularity is a consequence of missing interactions that are undetected for unclear reasons . This study provides insights into the differences between the two interaction detection methods as well as the nature of biological organization . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"genetics",
"and",
"genomics/bioinformatics",
"computational",
"biology/systems",
"biology"
] | 2009 | A Novel Scoring Approach for Protein Co-Purification Data Reveals High Interaction Specificity |
A structure-based model of myosin motor is built in the same spirit of our early work for kinesin-1 and Ncd towards physical understanding of its mechanochemical cycle . We find a structural adaptation of the motor head domain in post-powerstroke state that signals faster ADP release from it compared to the same from the motor head in the pre-powerstroke state . For dimeric myosin , an additional forward strain on the trailing head , originating from the postponed powerstroke state of the leading head in the waiting state of myosin , further increases the rate of ADP release . This coordination between the two heads is the essence of the processivity of the cycle . Our model provides a structural description of the powerstroke step of the cycle as an allosteric transition of the converter domain in response to the Pi release . Additionally , the variation in structural elements peripheral to catalytic motor domain is the deciding factor behind diverse directionalities of myosin motors ( myosin V & VI ) . Finally , we observe that there are general rules for functional molecular motors across the different families . Allosteric structural adaptation of the catalytic motor head in different nucleotide states is crucial for mechanochemistry . Strain-mediated coordination between motor heads is essential for processivity and the variation of peripheral structural elements is essential for their diverse functionalities .
Motors belonging to myosin superfamily are associated with a host of important cellular functions , including muscle contraction , cytokinesis , chemotaxis , targeted vesicle and organelle transport [1 , 2 , 3 , 4] . These motors perform mechanical work by producing movement on the actin filament powered by ATP hydrolysis [5 , 6 , 7] . While in some cases the myosin motors perform a single cycle of ATP dependent force generation and releases from the actin filament , there are examples where myosin motors perform multiple cycles of stepping on actin prior to its detachment [6 , 8] . The classic example of such processive motor is myosin V which carries cargo inside the cell towards positive end of the actin filament [8] . Myosin VI , however , carries cargo towards negative end of actin filament [9] . The unidirectional movement of these motors is a result of the coupling between nucleotide dependent conformational changes and actin binding/unbinding cycle [7] . Recent experimental studies have made significant progresses in understanding several features of the mechanochemical cycle of myosin . Kinetic study of single headed myosin has suggested that the rate limiting step in the whole cycle is the ADP release [10] . Kinetic experiments have also probed the weak binding states of myosin V [11] . Trybus et . al . has concluded from their kinetic experiments that monomeric myosin V is processive [12] . Structural investigations of myosin motors in different nucleotide bound state and also in the inactive state have improved our understanding significantly [13 , 14 , 15] . Several experiments have explored the head-to-head coordination between two heads of myosin motors at dimer level [16 , 17] . Recently , kinetic experiments have investigated the asymmetry in the ADP release rate between two heads and discussed about the effect of strain [18 , 19 , 20 , 21] . On the other hand using theoretical and computational techniques , the dynamics of motor processivity , [22] coupling between different parts of the motor domain , [23] transition between different states , [24]flexibility and collective vibrations , [25]coarse-gained modeling of mechanochemical cycle , [26 , 27] design principles for motility , [28] electrostatic origin of directionality [29] have been explored . Despite significant successes by these theoretical and experimental studies , there is still a need for a detailed structure-based comprehensive understanding of the unidirectional movement of these motors and the allosteric coordination between the heads for the processivity . Biology appears to have developed principles that are used for the specific needs of each of these motor families . Motor function in general needs allosteric adaptation of their active units to different nucleotide binding states that are mediated by competing interactions between strain interactions in the control element and in the motor binding . These general rules were first observed in our earlier work for the kinesin superfamily . A kinesin model was built based on our protein folding ideas of a strong energetic bias towards the native structure ( folding funnel ) of the unbound motor in solution complemented by competing interactions arising from nucleotide binding/hydrolysis/release and binding/release from the filament to explore the diverse functionalities in the kinesin superfamily [30 , 31 , 32] . Our results suggest a set of underlying principles governing the functional cycle of molecular motors in this family . The structural adaptation of the catalytic motor head in different nucleotide states and binding/release from microtubule is quite similar across the family . The nucleotide dependent allosteric transition of the peripheral structural elements like the neck-linker in kinesin -1 ( moves towards positive end on microtubule ) and the neck-helix junction in Ncd ( moves towards negative end on microtubule ) were found to be crucial for their directionality [30 , 31 , 32] . Additionally , a strain mediated coordination between the two motor heads in terms of nucleotide binding/release was found to be essential for processivity [27] . In a recent study by Zhang et . al . has dissected the whole mechanochemical cycle of kinesin-1 into three stages , including neck-linker docking , anisotropic translational diffusion of the trailing head and binding of the trailing head to αβ tubulin [33] . In the present work , we show that the same principles apply to the myosin functional cycle . Similarly to the kinesin superfamily , a structure-based model of myosin was built using available structural data that was complemented by adding nucleotide binding/hydrolysis/release and actin binding/release to investigate the structural aspects of the mechanochemical cycle . Using these same principles we could identify the mechanisms governing the functional cycle of the molecular motors in the myosin family . We find a structural adaptation of the trailing head under strain that provides crucial coordination between the two heads of myosin in terms of ADP release . This is essential for processivity . A nucleotide dependent allosteric transition of the converter domain is found to be needed for directionality . The opposite directionality in myosin V and VI can be explained in terms of the different peripheral structural elements . As the model has been built from the available crystal structure , we could provide the structural origin of these important conclusions along with allosteric communications between different parts of myosin in structural terms . By doing so , we are able to determine of the molecular details of governing mechanism of myosin . Indeed , this is one of the main highlights of the current study . Earlier modeling studies did not include this level of structural details . These results for myosin and kinesin strongly suggest that there are general rules/principles that govern the motor functionality across different superfamilies .
Recent experimental and structural data have revealed the different intermediates and the sequence of events of the stepping cycle of myosin motors [5 , 7 , 13] . Myosin motors , especially myosin V , can accomplish stepping with a single head ( Fig 1A ) [12] . The cycle starts with the ATP-bound myosin head ( state i ) which does not have actin binding ability . This state has a pre-powerstroke lever arm conformation pointing towards the right . After hydrolysis , the head with the ADP and the inorganic phosphate ( Pi ) ( ADP +Pi state ) binds to the actin filament ( state ii ) while still keeping its pre-powerstroke lever arm conformation . In the next step , the lever arm changes its conformation from the pre to the post-powerstroke conformation ( from right-directed to left-directed ) simultaneously with the release of the inorganic phosphate ( Pi ) while still bound to the actin ( state iii ) . This event is usually termed as powerstroke . Then , the bound head releases ADP to provide an actin bound empty head state ( state iv ) . This empty head , now , binds to ATP and subsequently releases from actin . Once released from the actin , the post-powerstroke conformation of the lever arm , generally , converts to the pre-powerstroke lever arm state ( state v ) . This step is termed repriming event . As it is clearly evident , state v is exactly same as state i with an additional stepping towards the left . Therefore , the cycle can repeat with the same sequence of events . We emphasize the two most important steps of the cycle that are responsible for the directionality as well as the progression of the cycle . In the powerstroke step ( ii → iii ) , the release of Pi changes the lever arm conformation of the bound head towards a definite direction ( towards positive end of the actin for Myosin V ) . This step crucially determines the directionality of the motor . In the next step ( iii → iv ) , the ADP release should be faster in this conformation to allow for a faster access for the ATP to bind the head which is important for the speeding up of the cycle . As described for the single myosin head , the motor detaches from the actin after the completion of each step . On the other hand , the double-headed myosin ( myosin V and myosin VI ) performs several consecutive cycles of unidirectional movement without being separated from the actin [8 , 9] . Such a processive movement requires coordination between the two heads for optimal functioning . In Fig 1B , the mechanochemical cycle for the double-headed myosin is presented . It starts when the head 1 is bound to the ATP and head 2 is bound to the ADP ( state i ) . The conformation of the lever arm of the head 1 is in a pre-powerstroke conformation , and it is not bound to actin as in the case of the single-headed myosin . Head 2 , on the other hand , is bound to the actin with post-powerstroke lever arm conformation . Next , ATP hydrolysis and subsequent binding of the head 1 while the head 2 is still bound to actin leads to a two-head bound state ( state ii ) . As we show later that this motor takes steps towards the left , the head 1 with ADP + Pi is referred as leading head with pre-powerstroke lever arm conformation . The ADP bound head 2 , however , is in post-powerstroke lever arm conformation and called as trailing head . The head 1 ( leading head ) releases Pi to provide a state with both heads bound to the ADP and to actin ( state iii ) . As discussed in the earlier section , the ADP bound leading head performs the powerstroke step to achieve its stable post-powerstroke lever arm conformation ( the stable lever arm conformation of head 1 is shown in green ) . However , the powerstroke step is delayed for the head 1 because head 2 ( the trailing head ) is still bound to the actin . This state is referred as powerstroke-postponed waiting state . Now , the release of ADP from the head 2 ( trailing head ) while still bound to actin results in state iv . Head 2 ( trailing head ) subsequently binds to the ATP and detaches from the actin . Now , once the head 2 is not bound to actin , head 1 ( leading head ) performs its waited powerstroke step . The head 2 also performs its repriming event towards the pre-powerstroke lever arm conformation ( state v ) . State v is exactly the same as state i with an additional step towards the left . Therefore , in state v , the identity of the leading and trailing head is interchanged and thus the cycle continues . In addition to the powerstroke step which is important for the directionality , the ADP release step is crucial for the processivity ( iii → iv ) . In state iii , if the leading head ( head 1 ) releases ADP faster than the trailing head ( head 2 ) , then the cycle has now a very different fate . The subsequent binding of ATP to the leading head ( head 1 ) releases it from actin and then eventually the trailing head ( head 2 ) also will be released from actin [19] . The resulting state is a completely unbound myosin dimer and thus the processivity is certainly hampered . However , such kind of termination can only happen if the ATP concentration is fairly high compared to its physiological concentration [20] . Therefore , the coordination between the two heads becomes really important in state iii where the ADP release from the head 2 with post-powerstroke lever arm conformation ( trailing head ) should be much faster than that from the head 1 with pre-powerstroke lever arm conformation ( leading head ) . In addition , the leading head in state iii provides an extra strain ( due to its postponed-powerstroke state ) on the trailing head that in turn can favor faster ADP release [18–21] . Here , in this article , we present the molecular origin of this important ADP release and powerstroke step using a structure-based modeling . There are two key conformations of the myosin motor in the mechanochemical cycle described above; the pre-powerstroke ( leading head ) and the post-powerstroke ( trailing head ) conformations when bound to actin filament . In the pre-powerstroke state , two conformations have been identified for myosin VI with two nucleotide bound states; ADP . Pi and ADP bound [34 , 35] . While these two conformations have significant changes in terms of actin binding affinity , the converter domain position with respect to motor head domain does not show any significant differences . The structural data of the post-powerstroke conformation is available for myosin VI in the literature [36] . Three-dimensional structures of these two conformations are shown in Fig 2A . Here we show the two important structural parts of myosin motor: ( 1 ) the catalytic motor head domain ( MH ) which binds to the actin filament and nucleotides , ( 2 ) the converter domain ( that connects to the lever arm ) which changes its conformation during the powerstroke to provide the directionality . It is evident from the figure that the major structural difference between these two conformations occurs in the converter domain . While the converter domain points towards the right in the pre-powerstroke conformation , it points towards the left in the post-powerstroke one . It is also worthwhile to note here that both myosins V and VI ( although they move towards opposite directions on actin ) have similar structural topology for these two conformations . The MH domain of these two motors has overlapping three-dimensional structures . Therefore , understanding the mechanochemistry of one motor is sufficient to extract the overall mechanism governing them . The difference in directionalities comes from differences in the peripheral structural elements that will be discussed later . In Fig 2B , the residue-residue native contact maps of these two conformations are shown . In the upper triangle of this figure , the contact map for the post-powerstroke conformation ( red ) is pasted upon the pre-powerstroke map ( blue ) . Such a representation reveals the unique pre-powerstroke conformational contacts that are mostly between the converter and MH domains . These contacts are also shown as lines in the pre-powerstroke conformation in Fig 2A . The unique contacts for the post-powerstroke conformations are similarly extracted and are shown as lines in the post-powerstroke conformation . It is important to note that these two sets of contacts describe the essential differences between these two conformations . We develop a structure-based model to analyze the non-native fluctuation in the MH domain of the trailing head compared to leading head in the ADP bound state when bound to actin [37 , 38 , 39] . In this model , we derive the Hamiltonian for the MH domain based on the pre-powerstroke crystal structure when bound to ADP [35] . The converter domain conformation and the converter-MH contacts of the pre-powerstroke crystal structure are used for the leading head simulation . To investigate non-native structural adaptation of the MH domain of the trailing head , we use the converter domain conformation and the converter-MH contacts of the post-powerstroke crystal structure with MH domain Hamiltonian is kept unchanged . The actin binding interface is extracted from the structural superposition with the actin bound myosin II structure ( PDB 1M8Q ) [40] . As ADP binds in the interface region between big ( red ) and small ( green ) subunit ( as shown in Fig 3A ) of the MH domain , we first calculate the root mean square deviation ( RMSD ) of the small subunit after least square fitting of the big subunit relative to the pre-powerstroke ( leading head ) MH conformation for both the cases . In Fig 3C , the RMSD distributions ( P ( RMSD ) ) for the leading and trailing heads are shown . The distribution of the lading head has a peak around 0 . 28 nm whereas the value for the trailing head is around 0 . 32 nm with a tail extending up to 0 . 8 nm . Thus , the change in the conformation of the converter domain has a considerable effect on the structural adaptation of the MH domain . This allosteric adaptation of the trailing head may result in a faster release of ADP that is crucial for the motor processivity ( double-headed myosin ) or for the speeding up of the detachment from the actin after the powerstroke step ( single-head myosin ) . In Fig 3B , we show the conserved region of ADP binding pocket in terms of P loop ( blue ) , switch I ( red ) and switch II ( green ) . We have calculated the distance between those conserved regions from both the simulations to explore specific structural changes . The results are shown in Fig 3D–3F . We find that the distribution of distances between switch I and P loop ( P ( dSWI-Ploop ) ) does not show any difference when comparing trailing and leading head simulation ( Fig 3D ) . The distribution of distances between switch II and P loop ( P ( dSWII-Ploop ) ) shifts towards larger distances for the trailing head compared to leading head slightly ( Fig 3E ) . Interestingly , the distribution for the distances between switch I and switch II ( P ( dSWI-SWII ) ) shows a significant shift towards larger distances for the trailing head compared to leading head simulation ( Fig 3E ) . We have also measured the RMSD of these structural motifs from both the simulations and found a similar conclusion ( shown in S1 Fig ) . Therefore , the nucleotide-binding region of the trailing head adapts a different structure upon the change in the conformation of the converter domain . This might lead to a faster ADP release from the trailing head compared to the leading head . Our result of faster ADP release form the head with post-powerstroke conformation compared to pre-powerstroke conformation is in agreement with previous experimental finding [21] . As discussed earlier , this coordination between the two heads ( state iii ) is necessary to maintain the processivity . Faster ADP release in turn dictates a faster ATP binding to the trailing head , which after binding to ATP detaches from the actin to maintain the sequence of events correct for optimal functioning . This adaptation to a non-native like structure around the ADP binding pocket in response to a strain ( change in the conformation of the converter domain ) is commonly referred as cracking [41 , 42 , 43] . One of the interesting suggestions about the mechanochemical cycle of a double-headed myosin is that the trailing head experiences a forward strain in the waiting state due the postponed-powerstroke nature of the leading head after Pi release ( state iii in the double-headed cycle ) [18–21] . This strain can regulate the faster ADP release of the trailing head which speeds up the whole cycle since ADP release step is rate limiting [10] . In Fig 4A , we show a schematic representation of the strain build up for the trailing head after Pi release from the leading head ( left head ) . While the strain should depend on many parameters like strength of the powerstroke , geometry of the waiting region , etc . , here we model the strain as the forward pulling force parallel to the actin axis acting on the converter domain . The allosteric response of the strain on the MH domain is calculated by monitoring distance between the small subunit and the big subunit of the MH domain . The distributions of the distances from different simulations with varying forces are shown in Fig 4B . As the strain ( force ) increases ( from 0 to 5 kJ mol-1 nm-1 ) , we find a gradual shifting of the distance distribution towards larger values . The peak positions of the distribution are at 3 . 75 nm , 4 . 0 nm and 4 . 2 nm for 0 , 3 and 5 kJ mol-1 nm-1pulling forces , respectively . The larger distance between these two domains at higher strain signals faster ADP release from the trailing head and that essentially speeds up the cycle . One should also note here that F-actin plays a very important role in the mechanochemical cycle of myosin . The above strain on the trailing head is developing because of the fact that the two heads of myosin are tightly bound to the two consecutive binding sites . Without this arrangement or without this binding surface , it is easy to guess that such a strain will not be possible to develop in the waiting state of motor . This is the first computational study on myosin that describes how strain can regulate the speed of the cycle and processivity of myosin . Next , we investigate the powerstroke step that is the central part of this process . During this step , myosin changes its conformation from the pre-powerstroke to the post-powerstroke state . Also the converter domain ( that is connected to the lever arm ) changes its structure ( as shown in Fig 5A ) , eventually altering the direction of the lever arm . This change in the direction of the lever arm provides directionality to the myosin motor with a displacement of cargo by ~36 nm ( for myosin VI ) in the direction of its motion . Experiments suggest that the release of Pi is important for this step [7] . Here , we build structure-based models with the MH Hamiltonian derived from the pre-powerstroke crystal structure bound to ADP and Pi [34] . A dual basin model for the converter domain and converter-MH contacts is derived from the above structure and post-powerstroke structure . In our model , we also keep the interactions between the MH and actin as earlier . Such a model allows the inter-conversion between the two structural states in myosin . In Fig 5A , we also show the important regions in MH domain which are in contact to the Pi in terms of the P-loop ( blue ) , switch I ( red ) and switch II ( green ) . First , we simulate the model including those Pi mediated contacts and the results are shown in Fig 5B . The RMSD ( calculated with respect to pre-powerstroke crystal structure ) probability distribution shows a bimodal behavior . The distribution for the ensemble of the pre-powerstroke conformations is peaked around 0 . 26 nm and for the post-powerstroke ensemble conformations around 0 . 7 nm . With the Pi mediated contact present in the system , the pre-powerstroke conformations are stable . We then remove the Pi mediated contacts from the MH domain . The RMSD probability distribution of the structural ensemble for this simulation is shown in Fig 5C . It is evident that now the post-powerstroke conformation ensemble is largely populated . Therefore , the structural adaptation due to the removal of Pi from the MH domain induces the transition from the pre to post powerstroke conformation; this is essentially the powerstroke mechanism . Here , the structural change in the converter region is observed as a consequence of a change occurring in the MH domain . These two structural elements , however , are distant in the three dimensional structure of myosin . This makes this transition as an example of allostery in protein motion . Thus , our simple model captures the essential coordination between the two heads and also explains the powerstroke mechanism in the myosin motors as a consequence of the allosteric structural adaptation . Similar models are used earlier to explore the mechanochemistry of kinesin family of motors ( kinesin-1 and Ncd ) [30 , 31 , 32] . The structural adaptation due to the change in nucleotide states or microtubule binding was the key for their functions . Therefore , it is safe to state that the allosteric structural adaptation due to local perturbation regulates the nucleotide binding/unbinding , filament binding/unbinding or the conformation change of peripheral structural elements . These features are crucial for the function of molecular motors . While myosin V moves towards positive end of the actin , myosin VI moves in the opposite direction . The MH domains of these two motors are structurally very similar with the conserved nucleotide and actin-binding domain . Although , myosin VI has a small insert in the MH domain , it does not influence the directionality [44] . The converter domains for both motors are also structurally very similar . These observations lead to the conclusion that the opposite directionality must be a consequence of the variation of other peripheral structural elements . Indeed , a unique insert present in myosin VI near its lever arm is found to be responsible for directing the lever arm in the opposite direction compared to myosin V [45] . A structural superposition and actin bound state of these two motors are presented in S2 Fig . Therefore , these two motors essentially use the same physical principle for their function with the modification of the peripheral structural elements to accomplish opposite directionality . Indeed , Tsiavaliaris et al have also shown that a forward moving myosin can be engineered to a backward moving myosin by inserting an insert which reveres direction of lever arm extension [46] . We should also mention that a similar strategy of changing the peripheral structural elements to achieve different directionalities is used by the kinesin superfamily of motors ( e . g . kinesin-1 and Ncd ) . Motor proteins of widely different superfamilies follow the same general working principles . Catalytic motor head domains of different motor proteins of a particular superfamily perform a similar sequence of tasks . The variation of structural elements peripheral to catalytic domains determine different directionalities of their movement . In this work , we reveal that the motors belonging to the myosin family follow general functional rules . The allosteric structural adaptation of the catalytic motor head in response to the different conformations of the converter domain creates asymmetry in the ADP release rates between the two heads . This asymmetry gets enhanced under a forward strain on the trailing head imposed by a postponed powerstroke conformation of the leading head . This coordination between the heads is essential in controlling the processivity of the cycle . An allosteric structural transition of the converter domain is observed upon the Pi release from the catalytic head domain and thus determining the structural aspects of powerstroke mechanism . Finally , we explain the diverse functionalities of myosin V and VI ( former walks towards positive end and later walks towards negative end of actin ) in terms of the variation in structural elements outside of the catalytic domain . These results parallel our earlier results that motors belonging to the kinesin family ( kinesin-1 and Ncd ) follow the same physical principles to accomplish their mechanochemical cycle and diverse functionalities . Note that all the models have built from the available crystal structure and therefore , we could provide detailed structural aspects of these important conclusions . Additionally , the allosteric communications between different parts of motor proteins are also understood in structural terms . By doing so , we are able to determine the molecular details of governing mechanism of myosin . Earlier modeling studies did not include this level of structural details . This collection of results suggests general underlying principles governing the functionality of molecular motors across different superfamilies .
The structure of actin bound myosin conformations after structural alignment is used to derive the structure-based model where amino acids are represented by single beads at the location of the C-α atom [37 , 46 , 48] . The complete Hamiltonian has the following form: H ( {r→i} ) =HpreMH+HpreConv+HpostConv+HpreConv−MH+HpostConv−MH+HMH−Actin ( 1 ) For every simulation , an initial structure is relaxed under the SB Hamiltonian followed by Langevin dynamics simulations at the low-friction limit at T = 300 K to sample the equilibrium structural ensemble ( this limit speeds sampling even if friction may be larger in the real system ) . The equation of motion for the Langevin dynamics used for integration is mr→¨i=−ζr→˙i−∂r→H ( {r→i} ) +Γ→i ( t ) ( 6 ) where ζ is the friction coefficient , −∂r→H ( {r→i} ) is the conformational force . Γ→i ( t ) is the random force satisfying ⟨Γi→ ( t ) ⋅Γj→ ( t' ) ⟩= ( 6ςkBT/h ) δij ( t−t' ) where integration time h is discretized . In this dynamics we chose ζ = 0 . 05 τL−1 and h = 0 . 0025 τL with τL= ( mσ2/εh ) 12 . Low friction was chosen for the purpose of effective conformational space sampling [49] . | Molecular motors are perhaps the most important proteins present in the cell . The importance specifically lies with the fact that these proteins use the chemical energy source ( such as ATP ) of the cell to generate mechanical work and perform a wide range of functionalities . In this article , we generalize the idea of using structure-based models to explore the mechanochemistry of myosin molecular motors in structural terms . We find that a structural adaptation of the motor head domain in post-powerstroke state signals faster ADP release from the trailing head to maintain its processivity while directionality arises from a careful design of peripheral structural elements . These results along with our earlier results on other motors provide a general rule for motor activity . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Methods"
] | [
"cell",
"motility",
"kinesins",
"crystal",
"structure",
"actin",
"filaments",
"engineering",
"and",
"technology",
"condensed",
"matter",
"physics",
"engines",
"molecular",
"motors",
"actin",
"motors",
"crystallography",
"motor",
"proteins",
"contractile",
"proteins",
"s... | 2016 | Strain Mediated Adaptation Is Key for Myosin Mechanochemistry: Discovering General Rules for Motor Activity |
The aim of this study was to determine , through a genome-wide association study ( GWAS ) , the genetic components contributing to different clinical sub-phenotypes of systemic sclerosis ( SSc ) . We considered limited ( lcSSc ) and diffuse ( dcSSc ) cutaneous involvement , and the relationships with presence of the SSc-specific auto-antibodies , anti-centromere ( ACA ) , and anti-topoisomerase I ( ATA ) . Four GWAS cohorts , comprising 2 , 296 SSc patients and 5 , 171 healthy controls , were meta-analyzed looking for associations in the selected subgroups . Eighteen polymorphisms were further tested in nine independent cohorts comprising an additional 3 , 175 SSc patients and 4 , 971 controls . Conditional analysis for associated SNPs in the HLA region was performed to explore their independent association in antibody subgroups . Overall analysis showed that non-HLA polymorphism rs11642873 in IRF8 gene to be associated at GWAS level with lcSSc ( P = 2 . 32×10−12 , OR = 0 . 75 ) . Also , rs12540874 in GRB10 gene ( P = 1 . 27 × 10−6 , OR = 1 . 15 ) and rs11047102 in SOX5 gene ( P = 1 . 39×10−7 , OR = 1 . 36 ) showed a suggestive association with lcSSc and ACA subgroups respectively . In the HLA region , we observed highly associated allelic combinations in the HLA-DQB1 locus with ACA ( P = 1 . 79×10−61 , OR = 2 . 48 ) , in the HLA-DPA1/B1 loci with ATA ( P = 4 . 57×10−76 , OR = 8 . 84 ) , and in NOTCH4 with ACA P = 8 . 84×10−21 , OR = 0 . 55 ) and ATA ( P = 1 . 14×10−8 , OR = 0 . 54 ) . We have identified three new non-HLA genes ( IRF8 , GRB10 , and SOX5 ) associated with SSc clinical and auto-antibody subgroups . Within the HLA region , HLA-DQB1 , HLA-DPA1/B1 , and NOTCH4 associations with SSc are likely confined to specific auto-antibodies . These data emphasize the differential genetic components of subphenotypes of SSc .
Genetic factors play an essential role in scleroderma or systemic sclerosis ( SSc ) etiology as in most complex autoimmune diseases [1] . Multiple reports of well powered candidate gene association and replication studies , together with the first genome-wide association study ( GWAS ) in this disease have led to the establishment of the Major histocompatibility complex ( MHC ) , STAT4 , IRF5 , BLK , BANK1 , TNFSF4 and CD247 as SSc susceptibility genes [2]–[15] . SSc is a clinically heterogeneous disease with a wide range of clinical manifestations , ranging from mild skin fibrosis with minimal internal organ disease to severe skin and organ involvement , reflecting the three main pathological events that characterize this disease: endothelial damage , fibrosis , and autoimmune dysregulation [16] . SSc patients are classified into two clinical subgroups based on the extent of skin involvement , limited SSc ( lcSSc ) and diffuse SSc ( dcSSc ) that are associated with different clinical complications and prognoses [17] . Another SSc hallmark is the presence of disease specific and usually mutually exclusive auto-antibodies that correlate both with the extent of skin involvement and the various disease manifestations , such as pulmonary fibrosis and renal crisis [18] . The most common are DNA topoisomerase I ( ATA ) , and anti-centromere antibodies ( CENP A and/or B proteins ) [19] . Each of these auto-antibodies is a marker for relatively distinct clinical subgroups of SSc , with anti-centromere typically associated with limited cutaneous disease , uncommon pulmonary fibrosis , late-onset pulmonary hypertension but generally an overall good prognosis , while ATA is a marker for diffuse skin disease and clinically significant pulmonary fibrosis with a resultant poorer prognosis . It has been observed that certain SSc clinical features and the presence of disease specific auto-antibodies vary in different countries and ethnicities [20] . This fact supports the likelihood that genetic factors may influence the different clinical features of the disease and auto-antibody subsets [19] . Furthermore , the affected members within multicase SSc families tend to be concordant for SSc-specific auto-antibodies and HLA haplotypes , thus , providing further evidence for a genetic basis for auto-antibody expression in SSc [21] . Moreover , several studies have reported that certain SSc genetic risk factors correlate with specific clinical subsets of the disease or SSc-related auto-antibodies [4] , . In this study , we aimed to identify novel genetic factors associated with different SSc clinical and auto-antibody subsets through a stratified re-analysis of results from a previous GWAS from our group and validation in a large replication study .
In the lcSSc subtype , seven non-HLA novel loci were identified as susceptibility markers in the GWAS data ( Table S1 and Figure S1 ) . Two out of the seven genetic markers showed evidence of association in the replication cohorts: rs11642873 near the IRF8 gene ( lcSSc P = 2 . 32×10−12 , OR = 0 . 75 [0 . 69–0 . 81] ) at the GWAS level of significance and rs12540874 in the GRB10 gene ( lcSSc P = 1 . 27×10−6 , OR = 1 . 15 [1 . 09–1 . 22] ) at the suggestive level of significance ( Figure 1 , Table 1 and Table S1 ) . Regarding the dcSSc subtype , five non-HLA loci were found to be associated in the GWAS cohorts ( Table S2 and Figure S2 ) . Upon analyzing these five SNPs in the replication cohorts we could only replicate the association of rs11171747 in the RPL41/ESYT1 locus ( overall dcSSc P = 5 . 99×10−8 , OR = 1 . 23 [1 . 14–1 . 33] ) ( Figure 1 , Table 1 and Table S2 ) . However , the association found in this locus was heterogeneous among cohorts ( Breslow-Day P = 5 . 32×10−9 ) . The observed associations in the ACA positive subgroup and lcSSc were difficult to differentiate because of substantial overlap between these two disease subgroups . In the GWAS cohorts , SNPs in IL12RB2 and RUNX1 genes were identified as novel non-HLA loci associated with SSc patients positive for ACA antibodies ( Table S3 and Figure S3 ) . However , none of these associations could be confirmed at the replication stage . Interestingly , the SNP rs11047102 of the SOX5 gene , which was selected for replication due to its association with the lcSSc subgroup in the GWAS data , showed suggestive evidence of association with the ACA subgroup ( P = 1 . 39×10−7 , OR = 1 . 36 [1 . 21–1 . 52] ) ( Figure 1 , Table 1 and Table S3 ) . In the ATA positive subgroup , four new susceptibility loci were identified in the GWAS data ( Table S4 and Figure S4 ) , none of which were confirmed in the replication phase . Since the ATA subgroup of patients has the smallest sample size , the lack of replication in any of the non-HLA locus may be due to a lower statistical power ( Table S5 ) . The associations found in the HLA region in the GWAS data set showed clear differences between SSc subgroups ( Figure 1 , Figure 2 , and Table 2 ) . The observed effects in the lcSSc and dcSSc subtype were similar to that of the overlapping group of patients with ACA and ATA respectively , but less significantly . Therefore , we focused the analysis on antibody subgroups only . We observed independent genetic associations in the ACA positive subgroup in the HLA region ( Table 2 and Figure 1 , Table S6 ) . The stronger independent signal was identified in the HLA-DQB1 gene of HLA class II: SNPs rs6457617 ( ACA+ P = 1 . 99×10−36 , OR = 0 . 48 [0 . 42–0 . 54] ) and rs9275390 ( ACA+ P = 2 . 62×10−54 , OR = 2 . 38 [2 . 13–2 . 67] ) . The TC allele combination ( both risk alleles ) showed a high association in the ACA positive subgroup ( ACA+ P = 7 . 81×10−61 , OR = 2 . 48 [2 . 22–2 . 77] ) , being present in 45 . 3% of the ACA positive patients compared to 25 . 1% of the controls ( Table 3 ) . Regarding the ATA positive subgroup , we also observed evidence of independent association in the HLA region ( Table 2 and Figure 1 , Table S7 ) . We found three associations in the HLA class II region: rs3129882 in HLA-DRA ( ATA+ P = 1 . 89×10−27 , OR = 2 . 17 [1 . 88–2 . 50] ) , rs3129763 in the HLA-DQA1/DRB1 loci ( ATA+ P = 1 . 47×10−11 , OR = 1 . 65 [1 . 42–1 . 91] ) and four associated SNPs in the HLA-DPA1/DPB1 region ( highest association at rs987870 , ATA+ P = 2 . 42×10−20 , OR = 2 . 09 [1 . 78–2 . 45] ) . The combination of three risk alleles in the DPA1/DPB1 locus , CAC ( ATA+ P = 1 . 27×10−76 , OR = 8 . 84 [6 . 72–11 . 63] ) of the SNPs rs987870 , rs3135021 and rs6901221 respectively was present in 10 . 6% of the ATA positive SSc patients compared to only 1 . 3% of the controls ( Table 3 ) . In addition , in the HLA class III region , the NOTCH4 gene was associated with the presence of ACA ( rs443198 , ACA+ P = 8 . 84×10−21 , OR = 0 . 55 [0 . 49–0 . 63] ) and ATA ( rs9296015 , ATA+ P = 1 . 14×10−8 , OR = 0 . 54 [0 . 44–0 . 67] ) , independently of the HLA class II associations ( Table 2 and Tables S6 , S7 ) . Interestingly , SNP rs9296015 had an opposite effect size in ACA and ATA subgroup , being exclusively associated in the ATA subgroup . These two SNPs were not in LD in Caucasian populations either from the HapMap project ( r2 = 0 . 05 in CEU and r2 = 0 . 03 in TSI ) or our cohorts ( r2 = 0 . 1 in the combined cohorts , r2 = 0 . 11 in Spanish , r2 = 0 . 00 in German , r2 = 0 . 00 in Dutch and r2 = 0 . 01 in US ) , pointing to independent associations in the NOTCH4 gene with both ACA and ATA positive subgroups . All the associations ORs found in the HLA region were consistent among the four GWAS cohorts ( Tables S8 , S9 ) . We wanted to investigate previously reported associations with subphenotypes or overall disease , such as CD247 , TNFSF4 , STAT4 , BANK1 , IRF5 and BLK in the present study's GWAS cohorts , to further establish them as SSc ( or its subphenotypes ) susceptibility loci . Table S10 shows the analysis of the SNPs in the previously mentioned genes which were present in our GWAS combined panel . As expected , association previously found in these six genes was replicated . Interestingly associations previously described to be confined to one of the SSc subgroups were also replicated as in the cases of TNFSF4 and lcSSc ( lcSSc P = 7 . 70×10−4 , OR = 1 . 18 [1 . 03–1 . 31] ) , STAT4 and lcSSc ( lcSSc P = 7 . 70×10−8 , OR = 1 . 31 [1 . 19–1 . 48] ) , BANK1 and dcSSc ( dcSSc P = 0 . 0103 , OR = 0 . 85 [0 . 75–0 . 96] ) and BLK and ACA+ ( ACA+ P = 1 . 45×10−4 , OR = 1 . 27 [1 . 12–1 . 44] ) . Furthermore association of CD247 with SSc was more strongly represented in the lcSSc subgroup than the others ( lcSSc P = 2 . 66×10−6 , OR = 0 . 81 [0 . 75–0 . 89] ) , although evidence of association was also found in the other subgroups . Similarly , the association found in IRF5 was stronger in lcSSc ( lcSSc P = 1 . 64×10−10 , OR = 1 . 50 [1 . 32–1 . 69] ) , although association was also found in the dcSSc , ACA+ and ATA+ subgroups .
Systemic sclerosis ( SSc ) is a rare , severe , complex and heterogeneous rheumatic disease . Multiple lines of evidence suggest that genetic factors may underlie not only SSc susceptibility but also the predisposition to develop specific clinical phenotypes such as lcSSc , dcSSc subtypes and the presence of SSc-specific auto-antibodies . The discovery of genetic variants associated with specific clinical manifestations of the disease will lead to new insights regarding pathogenesis and may open novel avenues of therapy that can be targeted to specific subsets . The aim of this study was to assess the genetic component involved in four different SSc clinical and auto-antibody subphenotypes through an analysis of our previous genome-wide association study ( GWAS ) data stratified for these disease subphenotypes , together with a large , new replication study . We have identified an association of the NOTCH4 gene with both ACA and ATA positive subgroups independent of the HLA associations . This gene is located in the MHC and encodes a transmembrane protein which plays a role in a variety of developmental processes by controlling cell fate decisions . Interestingly , NOTCH4 has been implicated in the pathways by which TGF-β induces pulmonary fibrosis [24] , one of the most severe clinical manifestations of SSc [25] , [26] . The Notch signaling pathway also controls key functions in vascular smooth muscle and endothelial cells which may be particularly relevant to the microvascular damage seen in SSc [27] . Genetic variants in NOTCH4 also have been previously associated , independently from HLA genes or alleles , with other autoimmune disorders like diabetes type 1 [28] , rheumatoid arthritis [29] and alopecia areata [30] , [31] . Additionally , through the analysis of the largest SSc case/control cohort reported to date we identified three new susceptibility loci ( IRF8 , SOX5 and GRB10 ) , outside the HLA/MHC region , implicated in genetic predisposition to different SSc subphenotypes , in addition to other suggestive loci . Type I and II interferons ( IFN ) are well known immunomodulators which can also regulate collagen production . Furthermore , they are believed to play a key role in the pathogenesis of SSc and other autoimmune diseases [32]–[34] . Interestingly , we found a strong association of the IRF8 gene with the lcSSc subtype and the ACA positive subgroup . IRF8 modulates TLR signaling and may contribute to the crosstalk between IFN-γ and TLR signal pathways , thus acting as a link between innate and adaptive immune responses [35] . IRF8 also has been demonstrated to be a key factor in B cell lineage specification , commitment and differentiation [36] . In addition , IRF8 has been associated with another autoimmune disease , multiple sclerosis [37] , although the SNP associated with multiple sclerosis ( rs17445836 ) was not present in our study . Nevertheless , both variants are in medium LD in the CEU population of the HapMap project ( r2 = 0 . 51 ) and both associations have a protective OR for the minor allele; pointing to a dependence in the associations found in these two diseases . The most prominent SSc specific auto-antibodies , ACA and ATA , are associated with the lcSSc and dcSSc clinical subsets , respectively [19] . The lcSSc subtype greatly overlaps with the ACA positive subgroup of patients ( almost all ACA positive patients belonged to the lcSSc subtype ) . Similarly , the dcSSc subtype overlaps with the ATA positive group of patients . Therefore , it is difficult to determine whether some of the observed associations specifically belonged to one of the four subgroups . Such is the case of the association found with the SOX5 gene . In the GWAS data , SOX5 was associated with lcSSc as well as with the ACA positive subgroup , although the association with the lcSSc subtype was stronger than that in the ACA positive subgroup . Upon completion of the replication study with the resultant increase in statistical power , we were able to determine that the SOX5 gene was indeed a risk factor for the ACA positive group at the genome wide significance level , but not for lcSSc . The SOX5 gene encodes a member of the SOX ( SRY-related HMG-box ) family of transcription factors involved in the regulation of embryonic development , in the determination of cell fate , as well as in chondrogenesis [38] . Conversely SOX5 , together with SOX6 and SOX9 , can induce many cellular types ( including melanocytes and bone marrow stem cells ) into the chondrogenic pathway , leading to expression of COL2A1 and the formation of cartilage [38] , [39] . As stated above , IFN type I and II are inhibitors of collagen production and chondrogenesis; more precisely IFN-γ ( type II IFN ) inhibits the COL2A1 gene which is one of the main downstream genes in the chondrogenesis pathway [40] . Taken all together , IRF8 ( part of the interferon pathway and induced by IFN-γ [41] ) and SOX5 may be affecting the formation of the extra-cellular matrix through COL2A1 in the skin and other organs of SSc patients . We also identified an association of the GRB10 gene with the lcSSc subtype; GRB10 codes for an adaptor protein known to interact with a number of tyrosine kinase receptors and signaling molecules and has a potential role in apoptosis regulation [42] . In dcSSc patients , the only observed genome wide significant association was with the RPL41/ESYT1 locus , although this association was heterogeneous among the investigated populations , probably due to lower statistical power in this smaller group . Three genes are relevant to this locus: RPL41 , a ribosomal protein not considered to be related to the immune system; ZC3H10 , a zinc finger protein related to tumour growth; and ESYT1 , a synaptotagmin-like protein of unknown function . Although none of these genes has a suggestive role in the pathogenesis of SSc a priori , further studies are needed to investigate this intriguing finding . Since most genes in the HLA region are implicated in the regulation of the immune system , it is not surprising that the HLA-association with SSc is primarily related to auto-antibody expression . We found different patterns of independent association for the two major SSc auto-antibody subgroups across the HLA class II region . Both genetic markers located in the HLA-DQB1 locus were associated with the presence of ACA auto-antibodies in SSc patients . The allelic combination of these SNPs tags the described association of HLA-DQB1*0501 with the ACA positive subgroup of the disease [22] , [43] . The associations within the HLA region in the ATA positive subgroup are more complex: SNP rs3129763 ( located near HLA-DRB1 ) tags the association of HLA-DRB1*1104 , which has been described to be associated with the whole disease [22] . Furthermore , the haplotype in the HLA-DPB1 region described in Table 3 , tags the HLA-DPB1*1301 also previously described [3] , [22] . Interestingly , the remaining independent association observed , rs3129882 , is found within the HLA-DRA gene , which is much less polymorphic than the other HLA genes already mentioned; nevertheless , the association found in this SNP is tagging through the extensive LD structure of the MHC region the association of some aminoacidic positions in the nearby HLA-DQB1 gene , which has not been previously reported to be associated with the ATA positive subgroup of SSc . In summary , taking advantage of our GWAS data and a large replication cohort , we have identified three new non-HLA loci associated with subphenotypes of SSc: GRB10 , IRF8 , and SOX5 . In addition , we shed light on HLA associations with this disease , establishing different patterns of independent association in the ACA and ATA positive subgroups . Our findings provide evidence for genetic heterogeneity underlying the clinical and especially autoantibody subtypes of SSc . These findings may prompt reconsideration of the current classification of SSc patients; provide insight into pathogenetic pathways differing among subphenotypes , especially specific auto-antibody subgroups , and lead to novel therapeutic targets for this devastating autoimmune disease .
For the GWAS analysis , a total of 2 , 296 Caucasian SSc patients and 5 , 171 Caucasian healthy controls were recruited through an international collaborative effort in the United States of America ( USA ) , Spain , Germany and The Netherlands . The North American cases ( initial n = 1 , 678; after applying quality control criteria , n = 1 , 486; 179 men , 1 , 307 women; mean age = 54 . 5 ( median , 55 . 0 ) ; SD = 12 . 9 ) were recruited from May , 2001 to December , 2008 from three U . S . sources: the Scleroderma Family Registry and DNA Repository and the Center of Research Translation in Scleroderma at The University of Texas ( UT ) Health Science Center-Houston , The Johns Hopkins University Medical Center and the Fred Hutchinson Cancer Research Center , each enrolling patients from a US-wide catchment area . The initial European SSc cases came from previously established nationally representative collections of 380 Spanish , 288 German and 190 Dutch patients with SSc . As control populations , healthy unrelated individuals of Spanish ( initial n = 414 ) , German ( initial n = 678 ) and Dutch ( initial n = 643 ) origin were included in the study as well as 3478 controls from across the US collected as non-cancer controls for GWAS studies of breast and prostate cancers in the Cancer Genetic Markers of Susceptibility ( CGEMS ) studies [44] , [45] ( http://cgems . cancer . gov/data_access . html ) . In the second replication phase , a large independent replication cohort , consisting of 3 , 175 SSc patients and 4 , 971 healthy controls of Caucasian ancestry , were collected from Belgium , Spain , The Netherlands , Germany , Italy , Norway , Sweden , UK and the USA . Details on the investigated populations are provided in the Table S11 . All cases met the American College of Rheumatology preliminary criteria for the classification of SSc [46] . Furthermore , patients were classified according to the extent of skin involvement into limited ( lcSSc ) or diffuse ( dcSSc ) forms [17] , [47] . In addition , the presence of SSc specific auto-antibodies , anti-topoisomerase I ( ATA , Anti-Scl70 ) and anti-centromere ( ACA ) was assessed by passive immunodiffusion against calf thymus extract ( Inova Diagnostics , San Diego , CA , USA ) and indirect immunoflourescence of HEp-2 cells ( Antibodies Inc , Davis , CA , USA ) , respectively , in a total of 5 , 229 and 5 , 238 SSc patients respectively . Auto-antibodies to RNA Polymerase III are also considered to be characteristic of SSc , but testing for this antibody is not widely available and since results were not known in almost two-thirds of our cases , this analysis was not done [18] , [19] . The distribution of SSc patients among these disease subsets is summarized in Table S11 . Collection of blood samples and clinical information from case and control subjects was undertaken with informed consent and relevant ethical review board approval from each contributing centre in accordance with the tenets of the Declaration of Helsinki . Most of the individuals included in this study , GWAS and replication cohorts , have been analyzed in a previous study [15] but novel genotypes were generated in the replication cohorts for phenotype associated SNPs found in the GWAS , expanding the scope of the study . Our goal was to examine any novel genetic association specific for each subset rather than overall disease . Although partial overlapping exists between lcSSc and ACA+ subgroups , and dcSSc and ATA+ subgroups; we wanted to assess whether association found in overlapped groups belonged to a subtype or an auto-antibody positive group . With that purpose we selected SNPs from the GWAS data based on the following criteria: This resulted in the selection of 18 non-HLA SNPs ( 7 for lcSSc , 5 for dcSSc , 2 for ACA+ , and 4 for ATA+ ) as shown in Tables S1 , S2 , S3 , S4 , corresponding to lcSSc , dcSSc , ACA and ATA positive patients respectively . The GWAS genotyping of the SSc cases and controls was performed as follows: the Spanish SSc cases and controls together with Dutch and German SSc cases was performed at the Department of Medical Genetics of the University Medical Center Utrecht ( The Netherlands ) using the commercial release Illumina HumanCNV370K BeadChip , which contains 300 , 000 standard SNPs with an additional 52 , 167 markers designed to specifically target nearly 14 , 000 copy number variant regions of the genome , for a total of over 370 , 000 markers . Genotype data for Dutch and German controls were obtained from the Illumina Human 550K BeadChip available from a previous study . The SSc case group from the United States was genotyped at Boas Center for Genomics and Human Genetics , Feinstein Institute for Medical Research , North Shore Long Island Jewish Health System using the Illumina Human610-Quad BeadChip . CGEMS and Illumina iControlDB controls were genotyped on the Illumina Hap550K-BeadChip . SNPs selected for the replication phase were genotyped in the replication cohorts using Applied Biosystems' TaqMan SNP assays on ABI Prism 7900 HT real-time thermocyclers . Markers with call rates of 95% or less were excluded , as were markers whose allele distributions deviated strongly from Hardy-Weinberg ( HW ) equilibrium in controls ( P<10−3 ) . Imputation was performed in the GWAS cohorts in order to gain genome coverage for the SNP selection . Imputation was performed with IMPUTE software 1 . 00 as previously described [48] , using as reference panels the CEU and TSI HapMap populations . However , SNP imputation did not show any new independent SNP associated at P<10−5 in the four subphenotypes considered . The imputed GWAS data in the four subphenotypes is shown in Figure S5 . Data in the SSc GWAS cohorts was filtered as follows: Using Plink , we identified and excluded pairs of genetically related subjects or duplicates and excluded the genetic-pair members with lower call rates . To identify individuals who might have non–western European ancestry , we merged our case and control data with the data from the HapMap Project ( 60 western European ( CEU ) , 60 Nigerian ( YRI ) , 90 Japanese ( JPT ) and 90 Han Chinese ( CHB ) samples ) . We used principal component analysis as implemented in HelixTree ( see Text S2 ) , plotting the first two principal components for each individual . All individuals who did not cluster with the main CEU cluster ( defined as deviating more than 4 standard deviations from the cluster centroids ) were excluded from subsequent analyses . Additionally , we excluded individuals with low call rates ( 11 individuals from the US group , 24 from the Spanish , 1 from the German and 1 from the Dutch ) , relatedness ( 50 from the US group , 2 from the Spanish , 1 from the German and 1 from the Dutch ) , non-European ancestry ( 42 from the US group , 5 from the Spanish , 6 from the German and 4 from the Dutch ) and inconsistent gender ( 83 from the US group , 2 from the Spanish , 2 from the German and 2 from the Dutch ) . Then we filtered for SNP quality , removing SNPs with a genotyping success call rate < 98% and those showing MAF < 1% . Deviation of the genotype frequencies in the controls from those expected under Hardy-Weinberg equilibrium was assessed by a χ2 test or Fisher's exact test when an expected cell count was < 5 . SNPs strongly deviating from Hardy-Weinberg equilibrium ( P<10−5 ) were eliminated from the study . For the combined analysis of the four datasets , the same quality controls per individual and per SNP were applied with the exception of the Hardy-Weinberg equilibrium ( HWE ) requirement . The genotyping success call rate on the merged dataset after all these quality filters were applied was 99 . 83% in the GWAS cohorts . The replication cohorts were filtered as follows: all individuals with a SNP success call rate below 0 . 95 were excluded , SNPs with a per individual success call rate below 0 . 95 were excluded , SNPs with a HWE comparison P value below 0 . 001 in controls were excluded and SNPs with a MAF below 0 . 01 were also excluded . As a result , 18 SNPs selected for replication all were in HWE ( P value > 0 . 001 ) and the overall genotype successful call rate was 96 . 61% and all SNPs individually had a successful call rate greater than 95% . We performed power calculations for GWAS and replication cohorts for the whole dataset and the clinical/auto-antibodies subphenotypes according to Skol et al . [49] ( Table S5 ) . The significance level for these calculations was set at 5×10-8 . χ2 tests were performed for allelic model for significant differences between cases and controls . Derived P values for the replication cohorts were not adjusted . All nine replication cohorts were jointly analyzed conducting Cochran-Mantel-Haenszel ( CMH ) tests to control for population differences . A threshold meta-analysis P value of <0 . 05 for the replication phase was considered significant . We also conducted CMH meta-analysis of all the nine replication cohorts and the four cohorts previously included in the GWAS , considering a P value lower than 5×10−8 as significant . Furthermore , P values in the range 5×10−8 to 5×10−6 were considered as suggestive associations . In all tests , odds ratios ( OR ) were calculated according to Woolf's method . We also applied Breslow-Day ( BD ) tests for all meta-analyses to check for heterogeneity in association among the investigated populations , and all associations with a P<0 . 05 in BD analysis were considered heterogeneous . Due to the partial overlapping of the lcSSc and dcSSc subgroups with ACA+ and ATA+ subgroups , respectively , we wanted to test whether an association found in both overlapping groups belonged to one or the other specifically . With that purpose , all the associations in the present study claimed to belong to a group were tested for association in the correlated group ( e . g . ACA associations were tested in lcSSc and vice versa ) to look for the best P value . In addition , ACA and ATA hits were tested in lcSSc-ACA- and dcSSc-ATA- , respectively , to ensure group specific associations . Also , lcSSc and dcSSc were tested in ACA+-non-lcSSc and ATA+-non-dcSSc with the same purpose . To determine independent associations in the HLA region , conditional logistic regression was carried out for all associated SNPs in the complete SSc group and the ACA and ATA positive subgroups . This analysis was carried out as implemented in Plink software , conditioning each SNP association to each of the other significantly associated ( P<5×10−7 ) SNPs in the corresponding LD block , controlling for the presence of the four populations as covariates . All SNPs which remained significant after conditioning were considered independent associations . All haplotype analysis was performed using Haploview software , defining the blocks by confidence intervals [50] . We only analyzed haplotypes or allelic combinations with frequencies of 1% and above . Statistical analyses were undertaken using R ( v2 . 6 ) , Stata ( v8 ) , Plink ( v1 . 07 ) [51] and HelixTree's SNP & Variation Suite ( v7 . 3 . 0 ) software ( see Text S2 ) . Plink software: http://pngu . mgh . harvard . edu/purcell/plink/ SVS HelixTree software: http://www . goldenhelix . com/SNP_Variation/HelixTree/index . html Stata software: http://www . stata . com/ R Statistical Package: http://www . r-project . org/ Haploview: http://www . broadinstitute . org/scientific-community/science/programs/medical-and-population-genetics/haploview/haploview | Scleroderma or systemic sclerosis is a complex autoimmune disease affecting one individual of every 100 , 000 in Caucasian populations . Even though current genetic studies have led to better understanding of the pathogenesis of the disease , much remains unknown . Scleroderma is a heterogeneous disease , which can be subdivided according to different criteria , such as the involvement of organs and the presence of specific autoantibodies . Such subgroups present more homogeneous genetic groups , and some genetic associations with these manifestations have already been described . Through reanalysis of a genome-wide association study data , we identify three novel genes containing genetic variations which predispose to subphenotypes of the disease ( IRF8 , GRB10 , and SOX5 ) . Also , we better characterize the patterns of associated loci found in the HLA region . Together , our findings lead to a better understanding of the genetic component of scleroderma . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"clinical",
"immunology",
"genetics",
"immunology",
"biology",
"genetics",
"and",
"genomics"
] | 2011 | Identification of Novel Genetic Markers Associated with Clinical Phenotypes of Systemic Sclerosis through a Genome-Wide Association Strategy |
Tripartite Motif ( TRIM ) ubiquitin ligases act in the innate immune response against viruses . One of the best characterized members of this family , TRIM5α , serves as a potent retroviral restriction factor with activity against HIV . Here , we characterize what are likely to be the youngest TRIM genes in the human genome . For instance , we have identified 11 TRIM genes that are specific to humans and African apes ( chimpanzees , bonobos , and gorillas ) and another 7 that are human-specific . Many of these young genes have never been described , and their identification brings the total number of known human TRIM genes to approximately 100 . These genes were acquired through segmental duplications , most of which originated from a single locus on chromosome 11 . Another polymorphic duplication of this locus has resulted in these genes being copy number variable within the human population , with a Han Chinese woman identified as having 12 additional copies of these TRIM genes compared to other individuals screened in this study . Recently , this locus was annotated as one of 34 “hotspot” regions that are also copy number variable in the genomes of chimpanzees and rhesus macaques . Most of the young TRIM genes originating from this locus are expressed , spliced , and contain signatures of positive natural selection in regions known to determine virus recognition in TRIM5α . However , we find that they do not restrict the same retroviruses as TRIM5α , consistent with the high degree of divergence observed in the regions that control target specificity . We propose that this recombinationally volatile locus serves as a reservoir from which new TRIM genes arise through segmental duplication , allowing primates to continually acquire new antiviral genes that can be selected to target new and evolving pathogens .
The TRIM protein family constitutes a newly appreciated group of innate immune effectors involved in the response to viral infection [1]–[3] . TRIM5α , one of the best studied members of this family , is a pattern-recognition receptor for mammalian retroviruses including HIV [4] , [5] . TRIM5α assembles into a hexameric lattice on the surface of retroviral cores as they enter the cytoplasm of a newly infected cell [6] . This interaction stimulates premature capsid disassembly [7] , [8] and the formation of unanchored K63-linked polyubiquitin chains that trigger the production of chemokines and cytokines including interferon [4] , [9] . The TRIM5 genetic locus has profound penetrance in determining viral titers in SIV ( simian immunodeficiency virus ) infected macaques [10] . TRIM5 also serves as a significant genetic barrier to the transmission of retroviruses between primate species [5] , [10]–[13] . Other TRIM proteins have been linked to infection by different families of viruses altogether . TRIM25 interacts with the influenza protein NS1 [14] , [15] and also activates the inflammatory response through the production of unanchored K63-linked polyubiquitin chains [16] . TRIM23 interacts with human cytomegalovirus [17] , TRIM56 with pestivirus [18] , while TRIM19/PML confers resistance to a broad range of DNA and RNA viruses [19] . In fact , more than one third of the approximately 70 known human TRIM genes have been shown to be transcriptionally upregulated in response to interferons [20] . Although the mechanistic details behind how TRIM proteins perform their antiviral roles remain elusive in most cases , their profound relevance to viral infection is made clear by the many viral antagonists that have been shown to target them . For example , influenza , herpes simplex virus-1 , human cytomegalovirus , and adenovirus are all known to encode proteins that interact with , or alter the activity of , human TRIM proteins [15] , [17] , [19] . By definition , TRIM genes encode proteins with a conserved domain order: a RING zinc-coordinating domain , one or two zinc-coordinating B-boxes , followed by a coiled-coil domain ( Figure 1A ) [21] . These three domains constitute the “tripartite motif” that gives this family its name . Most TRIM genes also encode a variable C-terminal domain , and in many of them , this is a B30 . 2 domain [2] . The B30 . 2 is composed of a series of β-strands folded into a globular β-sandwich structure [22] . Different metazoan genomes contain different complements of TRIM genes . For example , Drosophila melanogaster has seven TRIM genes and Caenorhabditis elegans has eighteen [2] . In a previous comparison of the TRIM gene complements found in the human and mouse genomes , most were found to be strict 1∶1 orthologs [2] . This suggests that the majority of human TRIM genes are ancient , having originated more than 90 million years ago when human and mouse shared a last common ancestor . However , that study identified one phylogenetic clade of TRIM genes specific to the mouse genome , and two clades specific to the human genome . The clade of TRIM genes specific to the mouse genome ( TRIM12/TRIM30 and related genes ) was subsequently shown to be an expanded set of TRIM5 paralogs [23] . Based on this , we wished to describe the two phylogenetic clades of TRIM genes ( TRIM50/73/74 and TRIM43/48/49/64 ) which are specific to the human genome ( Figure 1B ) . We also wished to determine whether these genes have been maintained by neutral drift or by selection , potentially imposed by evolutionarily recent viral infections . In the process of characterizing these young human TRIM genes , we identified many additional , previously unidentified human TRIM genes to which they are closely related , bringing the total number of known human TRIM genes to approximately 100 . We show that these novel genes have arisen from recent , and in some cases even human-specific , segmental duplication events . Specifically , we find that one locus on chromosome 11 , containing nine tandemly situated TRIM genes , has spawned at least two separate segmental duplications of itself during the evolution of great apes , as well as having produced at least one other segmental duplication that is still polymorphic in the human population . This locus is therefore evolutionarily dynamic as well as copy number variable within the human population . In a fascinating example of trans-species copy number variation , this locus was recently annotated as one of 34 “hotspot” regions that are also copy number variable in the genomes of chimpanzees and rhesus macaques [24] . Trans-species copy number variation remains largely unstudied , and the evolutionary forces behind it remain unknown [25] . We propose that this locus is selected to remain recombinationally volatile so that it can serve as a reservoir from which new primate TRIM genes constantly arise . Theoretically , increased gene dosage of innate immunity genes , conveyed by increased copy number , could in itself provide protection against viral infection and disease progression . However , many of these genes are evolving under positive selection like other primate genes known to encode antiviral molecules [26]–[35] . Therefore , these duplicated genes also appear to be rapidly diversifying in function , possibly to expand the spectrum of antiviral affinities in response to new and evolving viruses .
In a previous comparison of the TRIM genes found in the mouse and human genomes , several human-specific genes were noted ( Figure 1B ) [2] . Although these genes could have arisen anytime during the last 90 million years since human and mouse last shared a common ancestor , we were interested to know whether any of them have arisen during Catarrhini speciation ( Figure 1C ) . This group constitutes our closest evolutionary kin , primates that have most likely faced pathogens similar to those that humans encounter . To address the evolutionary origins of these human-specific TRIM genes , we took advantage of the genome projects of several Catarrhini species , including chimpanzee and orangutan ( both great apes ) , human , and rhesus macaque ( an Old World monkey ) . The first clade of human TRIM genes absent in the mouse genome contains TRIM50 , TRIM73 , and TRIM74 ( Figure 1B ) . To investigate when these genes arose , orthologous sequences were identified in the other Catarrhini genomes and a phylogeny was constructed ( Figure 1D ) . The most closely related human outgroup sequence , TRIM72 , was also included . TRIM72 forms a clear orthogroup containing one gene from each species , with all nodes being consistent with speciation events ( boxed in green ) . However , the TRIM50/TRIM73/TRIM74 clade has been more dynamic ( boxed in yellow ) . This branching pattern is consistent with an ancestral TRIM50 that experienced two duplication events , each indicated by a star on the phylogeny . The first duplication , giving rise to TRIM73 , occurred after the split between great apes and Old World monkeys , but before our last common ancestor with orangutan . It involved only the exons encoding the first three protein domains . The second duplication event occurred in the human lineage , less than 7 million years ago , giving rise to TRIM74 . Consistent with two duplication events , TRIM50 , TRIM73 , and TRIM74 reside near each other on three segmental duplications on human chromosome 7 [36] . Spliced transcripts have been identified for all three genes , and while TRIM50 has been demonstrated to act as an E3 ubiquitin ligase , the biological functions of TRIM73 and TRIM74 remain uncharacterized [36] . Thus , this small TRIM clade has gained gene copies though segmental duplications that have occurred during recent primate evolution . One gene , TRIM74 , is even specific to humans . The second clade of human TRIM genes absent in the mouse genome contains TRIM43 , TRIM48 , TRIM49 , and TRIM64 ( Figure 1B ) . Using reciprocal best hit analysis , we failed to detect strict 1∶1 orthologs between human and the other primates being analyzed . In fact , reciprocal searches of the human genome with primate orthologs continually returned large numbers of mostly un-annotated human genes . In all , we identified a group of 31 human TRIM genes that form a single monophyletic clade to the exclusion of all other TRIM genes in the human genome ( Figure 2A ) . The clade includes seven TRIM genes previously assigned standard TRIM names ( including the four used as queries , bold type ) and 24 uncharacterized paralogs . Uncharacterized genes were given temporary names reflecting their phylogenetic subclade ( i . e . A1 and A2 are two genes in the ‘A’ subclade shown in Figure 2A ) , but actual locus identifiers for each gene are given in Table 1 . Of these 31 genes , 20 have full-length open reading frames that are predicted to encode tripartite motifs either with or without a C-terminal B30 . 2 domain ( Figures S1 , S2 ) . Using RT-PCR on testes mRNA , we were able to identify processed transcripts for 11 of these 20 full-length genes , and when combined with cDNA reads available in Genbank , 14 out of 20 genes have evidence for expression and splicing ( Figure S2 ) . Thus , we have identified a large set of previously undiscovered human TRIM genes , most of which appear to have protein coding potential . The 31 human TRIM genes in this clade are located at three genomic loci , one near the centromere of chromosome 2 , one spanning the centromere of chromosome 11 , and one on the arm of chromosome 11 at 11q14 . 3 ( Figure 2B ) . When the genes in this schematic were color-coded to reflect the subclades in Figure 2A , it became clear that they arose through a series of segmental duplications . For example , the cluster of TRIM genes located on the chromosome 11 arm contains a mirror-image tandem inversion of 7 TRIM genes ( denoted by two orange bars in Figure 2B ) . A second duplication event is evident in the region directly adjacent to the chromosome 11 centromere , where a stretch of 6 TRIM genes appears to be an inverted copy of part of the sequence located on the chromosome 11 arm ( denoted by green bars in Figure 2B ) . The genes located near the chromosome 2 centromere cluster phylogenetically with those found on chromosome 11 , suggesting that this may be yet another segmental copy , although the gene order is sufficiently degraded that we cannot draw any clear conclusions . For discussion purposes , we denote the regions containing these apparent duplications on chromosome 11 as segment 1 , segment 2 , and segment 3 , as illustrated in Figure 3 . The duplicated chromosomal regions bearing these three segments are large , but we have focused only on the portion that contains TRIM genes . A careful inspection of the chimpanzee , orangutan , and rhesus macaque genomes in these regions was then performed ( Figure 3 ) . Segment 1 is found in all of these primate genomes , while segment 2 is found in the chimpanzee and human genomes only . In support of the young age of segment 2 , segments 1 and 2 in the human genome are 96% identical along their length ( calculated for 302 kilobase; Table S1 ) . The identification of segment 2 in the genomes of orangutan and rhesus macaque is somewhat complicated by a large chromosomal inversion that has been reported in the region of the chromosome 11 centromere , which arose in the common ancestor of human , chimpanzee , and gorilla [37] . However , this segmental duplication has been previously analyzed by FISH , using BAC-derived probes that anneal in both segment 1 and segment 2 [37] . In that study , all primate species investigated showed a hybridization signal on the chromosome 11 arm at the location of segment 1 . A second hybridization signal at the chromosome 11 centromere ( segment 2 ) was observed in the genomes of human , chimpanzee , and gorilla , but not of orangutan , rhesus macaque , or any other primate tested , congruent with our conclusions made through comparative genomics . Therefore , segment 2 is a segmental duplication of segment 1 that is specific to humans , chimpanzees , gorillas , and presumably the final species of this monophyletic clade of African apes , bonobos . This is in agreement with a previous age estimate of 14 million years , based on sequence divergence , for the segmental duplication containing segment 2 [38] . Interestingly , segment 3 is found only in the human genome ( Figure 3 ) . The very recent acquisition of this segment as another copy of segment 1 is supported by the observation that segments 1 and 3 are 99 . 6% identical along their length ( calculated for 170 kilobase; Table S1 ) . In intra-chromosomal comparisons of all segmental duplications on chromosome 11 , only 0 . 2 MB was found to have this level of identity [39] , consistent with segment 3 being one of the newest segmental duplications on the entire chromosome . It is curious to note that genes A1 and A2 , located in tail-to-tail fashion , are 100% identical along their length ( over 6 kilobase ) in the human genome , but only 96–98% identical in the genomes of orangutan or chimpanzee ( Figure 3 ) . A gene conversion event between A1 and A2 in the human genome may have accompanied or seeded the tandem inversion of segment 1 to create segment 3 . In summary , we have identified species- and human-specific TRIM genes on chromosome 11 . The chromosomal region bearing segment 1 has been highly dynamic during the evolution of humans and African apes , seeding at least 2 segmental duplications in the last 18 million years since their last common ancestor with orangutan . There appears to be a gross dichotomy in the timing of TRIM gene acquisition by the human genome because , while the majority of human TRIM genes are ancient and arose >90 million years ago , the rest of them ( approximately 20% of the TRIM genes in our genome ) have arisen in very recent time , during the evolution of the great apes . Because the locus containing segment 1 has spawned multiple segmental duplications in recent primate history , we wished to determine whether there are also newer segmental duplications of this region that might be polymorphic in the human population . Genomic regions containing polymorphic segmental duplications or deletions greater than 1 kilobase in size are called ‘copy number variable’ ( CNV ) regions [40] . To characterize the population genetics of this locus , we employed the multiplex ligation-dependent probe amplification ( MLPA ) assay [41] . This is a PCR-based assay that utilizes a fragment analyzer to quantify the amount of product generated from different target regions in a genome ( Figure 4A ) . Eighteen probe pairs were designed to tile across the length of segment 1 ( Figure 4B and Table S4 ) . Each probe in each pair anneals to 23–63 bases of genomic DNA , and is a perfect match to a target sequence in segment 1 . Because of the high degree of similarity between segment 1 and segment 3 , probes will also anneal to the cognate locus in segment 3 with perfect complementarily . However , probe pairs were carefully situated such that they have multiple mismatches to the corresponding sequence in segment 2 , or to sequence anywhere else in the human genome ( see materials and methods ) . Thus , each probe is expected to have four binding sites in a diploid genome because there will be two copies of each segment 1 and segment 3 target . The one exception is the probe pair “M-uniq , ” which sits in a unique stretch of sequence between segments 1 and 3 and will have only two binding sites in a diploid genome . Control probe pairs that recognize standard single-copy genes distributed throughout the genome were also included ( see materials and methods ) . Initially , 50 genomic DNA samples from individuals from around the world were analyzed . For each probe pair , the quantity of products produced from each sample was normalized to the quantity produced from a reference genome , as is standard for this assay . As the reference , we used the genome of a Caucasian male from Utah ( NA10851 ) that has been previously used as the reference genome in several studies of CNV regions [40] , [42]–[44] . The normalized fragment values for 12 representative genomes are graphed in Figure 4C , and values for all 50 surveyed genomes are presented in Table S5 . We identified only one CNV , in the genome of a Han Chinese female ( NA18573 ) . In this individual , the 16 probe pairs spanning from GAP1-1 to M-uniq all yielded approximately 1 . 5 times the quantity of fragments as the reference genome ( average enhancement across all 16 probe pairs = 1 . 6 ) . This was verified in eight independent experiments ( not shown ) . Therefore , this Han Chinese individual has two additional binding sites for each of these probe pairs , and one additional binding site for the M-uniq probe pair . This pattern is consistent with the segmental duplication scenario diagrammed in Figure 4D . A signal at this locus was previously detected in this same individual as part of a whole-genome array-based study that identified 1 , 447 human CNV regions [40] . Thus , we have identified a human individual of Chinese decent that has 43 TRIM genes belonging to this dynamic phylogenetic group instead of the 31 TRIM genes which most human individuals have . Based on this finding , we then screened 22 additional Han Chinese samples , but did not find another instance of this segmental duplication ( Table S5 ) . In all , 72 human genomes were analyzed by MLPA ( human samples are listed in Table S6 ) . The polymorphism thus seems to be rare , as it was detected in only 1/72 human individuals . However , CNVs have also been detected at this locus , using array-based platforms , in the genomes of several other Asians , including 2 Japanese , 2 Koreans , and 1 additional Chinese individual ( Figure S5 ) [40] , [42] , [44] . There is one report of a CNV at this locus in the genome of a Yoruban from Africa [43] . Perhaps most interestingly of all , the region containing segment 1 is also copy number variable in chimpanzees and rhesus macaques [24] . In summary , the region containing segment 1 has been highly dynamic both during primate speciation , and also in current human and primate populations . All of the findings described so far can potentially be explained as random events occurring in a dynamic genome . As segmental duplications arise , they may go to fixation through neutral drift even if there is no selection acting for or against them . A hallmark of genes that are being retained by neutral drift is that they accumulate equal rates of non-synonymous and synonymous mutations . Such genes have a characteristic signature of dN/dS = 1 , where dN is the number of non-synonymous mutations per non-synonymous site , and dS is the number of synonymous mutations per synonymous site . In contrast , most functional genes accumulate non-synonymous mutations at a rate far slower than synonymous mutations ( dN/dS<<1 ) due to the evolutionary constraint at play [45] . A third mode of evolution , recurrent positive selection , has influenced several TRIM genes in primate genomes , including Pyrin/TRIM20 [46] , TRIM5 [28] , [32] , and TRIM22 [27] . Genes or gene regions subject to such a selective regime accumulate a characteristic signature of dN/dS>1 [47] . We analyzed the evolutionary pressures that have shaped these young TRIM genes at the sequence level in order to determine whether they have been neutrally or selectively retained . Usually , evolutionary datasets of orthologous sequences are used for such analyses , but because these genes are so new and dynamic , deep species sets of strictly orthologous sequences cannot be easily obtained . Instead , we looked at the patterns of nucleotide substitution that have occurred during the diversification of these genes by comparing human paralogs , all of which can be traced to a common ancestral gene ( asterisk in Figure 2A ) . Of the 31 TRIM genes in the dynamic clade being investigated , 15 are predicted to encode proteins with the full TRIM-B30 . 2 structure ( Figure S2 ) . However , of these , two very recently diverged gene pairs ( A1/A2 and C1/C2 ) are still identical along the length of their coding sequence ( Table S2 ) , leaving 13 unique sequences which can be analyzed . Importantly , analysis of sequence evolution requires an accurate phylogenetic representation of the genes being analyzed [48] . One problem with understanding the phylogenetic relationship of paralogs from a single genome is the fact that gene conversion may have occurred . To detect phylogenetic incongruencies in our alignment , indicative of such events , we used the GARD program [49] as described in the materials and methods . Only one phylogenetic breakpoint was identified ( p<0 . 05 ) , located between the RING and B-box2 encoding domains of the first protein-coding exon ( Figure 5A ) . The alignment of the 13 TRIM genes was subsequently divided at this point and the trees produced by each half are shown in Figure 5C . Only two branches differ between the trees ( highlighted in red ) , suggesting that gene conversion has not been extensive . The tree for each half is highly supported , regardless of the phylogenetic method utilized ( Figure 5C ) , or whether just sites at the third positions of codons are utilized ( data not shown ) . With these trees , each half of the multiple alignment was analyzed separately using the codeml package in PAML ( see materials and methods ) . We analyzed each half of the alignment separately , under variable models of selection and codon usage ( Table S3 ) . All models yielded strong support for positive selection acting on both halves of the gene ( p<0 . 05 ) . Each of the two trees has one poorly supported node ( highlighted in green; Figure 5C ) . We confirmed that support for positive selection remains strong when each of these nodes is collapsed ( p<0 . 05; Table S3 ) . Therefore , there is convincing evidence that these new TRIM genes have not been retained by neutral drift alone . dN/dS values were calculated for each branch on the two trees . On these trees , there are many branches where dN/dS>1 ( text highlighted red; Figure 5C ) . In fact , the dN/dS values along some of the gene lineages are remarkable . For example , there have been 17 non-synonymous and 0 synonymous mutations that have accumulated during the divergence of the B5 gene ( 6 non-synonymous in the first half and 11 in the second half ) . The identical genes C1 and C2 have accumulated 25 non-synonymous changes and 0 synonymous changes since they shared a last common ancestor with the other genes of the ‘C’ clade , a stunningly intense episode of positive selection . Collapsing of the poorly supported node in each tree only marginally affects the results ( Figure S4 ) . Such extreme evolutionary patterns are unusual , but have been documented previously in other viral restriction factors due to the intense evolutionary arms races in which these genes are engaged [26] , [28] , [33] , [34] , [50] . These analyses can identify specific codon positions , and corresponding amino acid residues , that have repeatedly been subject to positive natural selection . Ten rapidly evolving codons were identified ( Table S3 ) , as illustrated with tick marks on the protein schematic in Figure 5A . Five of these fall in or near the RING domain ( residues F48 , V50 , E54 , E60 , H69 in TRIM49/F1 coordinates ) , three in the coiled-coil domain ( R166 , C167 , R222 ) , and two in the B30 . 2 domain ( Y320 , A323 ) . Using secondary structure prediction and alignments to other TRIM proteins , we determined that rapidly evolving residues Y320 and A323 fall in a small loop ( 11–16 aa long ) that lies between the second and third β-strands of the B30 . 2 domain ( Figure S3 ) . This surface-exposed “variable loop 1” ( Figure 5B ) has been rather well characterized , at least in the case of TRIM5α . In this protein , this loop is known to be a major determinant of recognition for retroviral capsids , and presumably constitutes the major binding interface with retroviruses . Sequence variation in this loop of TRIM5α accounts largely for the species-specific viral restriction patterns observed in various primate and mammalian species [28] , [51]–[53] . It is hypothesized that the TRIM5 gene has been engaged in an arms race with retroviruses throughout the diversification of mammals , and that natural selection has driven rapid sequence evolution of this loop for improved recognition of constantly changing retroviruses [47] , [54] . Accordingly , there are multiple sites of positive selection in the variable loop 1 region of TRIM5α from primates [28] , cows [27] , and rabbits and hares [55] , all of which restrict mammalian retroviruses . It is intriguing that the young TRIM genes identified here should have two codons evolving under positive selection in the B30 . 2 domain , with both of them falling in this small surface loop known to interact with retroviral capsids . Three rapidly evolving residues were also identified in the coiled-coil domain ( Figure 5A ) . In TRIM5α , the coiled-coil domain is the second domain that participates in retroviral target specificity [51] , [56] . The rapidly evolving residues identified here are in regions shown to be critical in defining virus-specificity in TRIM5α [56] , and known to contain codons evolving under positive selection in the TRIM5 gene ( Figure S3 ) . In summary , positive selection has acted on these young TRIM genes in regions analogous to the major determinants of retroviral specificity in TRIM5α , suggesting that the novel genes could be retroviral restriction factors . Based on the evolutionary signatures observed , we next tested whether these genes might encode retroviral restriction factors . We chose some of the genes with the highest branch-specific dN/dS values for functional testing , ones which could also be amplified in their complete form from human mRNA samples . These candidates ( A1 , B1 , B5 , F1/TRIM49 , F2 , and F3 ) are indicated with stars in Figure 5C . Except for F1/TRIM49 , none of these genes have ever been previously studied . We tested the ability of these genes to restrict cellular entry of three different mammalian retroviruses which are known to be restricted by human and/or rhesus macaque TRIM5α: feline immunodeficiency virus ( FIV ) , HIV , and murine leukemia virus ( MLV ) . Interestingly , the young human TRIM genes did not restrict these retroviruses ( Figure S6 ) . If these genes do have anti-retroviral activity , these data suggest that their specificity is different than that of TRIM5α , consistent with the high degree of divergence observed in the regions that control target specificity . Perhaps these TRIM genes were honed to target retroviruses that are now extinct [57] , [58] . Alternately , these TRIM genes may target other virus families altogether . For instance , TRIM22 has experienced positive selection in these same retroviral targeting motifs , and while this gene may be relevant to retroviral infection [59] , it also has activity against hepatitis B [60] and picornaviruses [61] .
Here we identify and characterize what are likely to be the youngest TRIM genes in the human genome . While the ∼100 human TRIM genes are for the most part ancient , we now show that a substantial number of them ( approximately 20% , not counting additional TRIM gene copies that are polymorphic in the human population ) have arisen in recent evolutionary time , during the speciation of the great apes . Many of these genes have full-length open reading frames and produce spliced transcripts . Almost all have arisen from segmental duplications that can be traced to a single locus on the arm of chromosome 11 . We propose that the segment 1 region on the arm of chromosome 11 is a TRIM gene “factory , ” producing copies of the genes that it contains by spawning segmental duplications around the genome . Increased dosage of these genes may , in itself , be adaptive . Further , depending on the chromosomal context of new segmental duplications , the genes that they contain may be expressed in different tissues or at different developmental stages . However , we also find that positive selection is rapidly shaping the sequence of these genes , such that new copies may quickly become specialized for new functions or specificities . It will be important to determine what role , if any , these young TRIM genes play in innate immunity , particularly because these genes are copy number variable in human and primate populations . Several insights into the evolutionary dynamics of large gene families can be gained from this study . TRIM genes are found throughout the human genome and the segmental duplications described here may illustrate one mechanism by which this family has expanded over time . Because all gene duplicates start out as polymorphisms segregating in populations , the observed copy number variation suggests that this gene family is still growing . Several lines of evidence support the idea that the fixation of new TRIM paralogs in the human genome has , at least in part , been an adaptive rather than a neutral process . First , TRIM5 [28] , TRIM22 [27] , Pyrin/TRIM20 [46] , and many of the TRIM genes described herein , have all evolved under positive selection , accumulating unexpectedly high numbers of non-synonymous mutations . New genes are especially prone to positive selection , probably because redundant gene copies provide templates for the evolution of new functions [24] , [62] . Second , some TRIM genes have been highly dynamic in terms of species-specific gene gain and loss during mammalian evolution . For instance , cows and rodents possess independent , species-specific expansions of tandemly situated TRIM5 paralogs , while dogs and cats have independently lost the function of this gene [27] , [54] , [63] . Likewise , the TRIM genes in segment 1 on chromosome 11 are also highly dynamic , having spawned at least two segmental duplications in African apes and another that is now polymorphic in the human population . Immunity genes are , in general , overrepresented amongst mammalian gene families that show rapid gene gain and loss , suggesting that these events are often adaptive [64] . Third , the acquisition of the young TRIM genes has not come without a cost to the human genome . Unequal crossing-over and aberrant homologous recombination between the tandem segments that contain TRIM50 , TRIM73 , and TRIM74 causes Williams-Beuren syndrome in 1/7 , 500 to 1/25 , 000 newborns [65] . This fitness consequence might be expected to be offset by a fitness advantage , otherwise these regions would be selectively lost from the genome . Historically , studies of human genetic variation have focused almost exclusively on single nucleotide polymorphism ( SNP ) differences between individuals . Recently , it has become apparent that large DNA segments are also commonly polymorphic between individuals , resulting from recent segmental duplications and deletions . CNV regions can be associated with disease , usually related to the altered gene dosage that they convey [66] . Another negative consequence of CNV duplications is that blocks of nearly identical sequence interspersed in a genome can create a volatile landscape for recombination . However , positive attributes can also be imagined for CNV regions , such as benefits that might be gained from increased dosage of certain genes . Such a fitness advantage has been suggested for the salivary amylase gene ( AMY1 ) , which is found in higher copy number in populations with higher starch diets [67] . Here we propose that CNV regions can also be a positive , adaptive force in genomes by driving the generation and diversification of gene families important to human immunity . For all of these reasons , studies of CNV regions are important for understanding individual disease susceptibility .
Refseq annotated human coding sequences of genes of interest were downloaded from Genbank . Analysis of human – chimpanzee – orangutan – rhesus macaque orthogroups was performed by reciprocal-best hit analysis performed in the UCSC genome database [68] . Briefly , each human gene was used as a BLAT query [69] against the genome projects of the other species investigated . The top hits from these queries were then used to reciprocally query the human genome . All related sequences were then compiled and subjected to phylogenetic analysis . cDNA or full-gene sequences were aligned using MUSCLE as implemented in MEGA5 [70] . Alternate trees ( neighbor joining , maximum likelihood , and maximum parsimony ) were constructed within MEGA , with gapped positions excluded . Tree nodes were critically evaluated by performing 1 , 000 bootstrap replicates . The physical locations of the TRIM genes in the human ( hg19 ) , chimpanzee ( panTro2 ) , orangutan ( ponAbe2 ) , and rhesus macaque ( rheMac2 ) genome assemblies were determined by manual inspection in the UCSC genome browser [68] . It was not possible to determine the structure of the chromosomal 11 loci in the marmoset genome assembly ( calJac3 ) , due to poor sequence quality . Primers were designed to recognize novel TRIM genes ( primer sequences are given in Table S7 ) . Each primer set was designed to span at least one intron so that products resulting from processed transcripts could be differentiated from those potentially resulting from contaminating genomic DNA . SuperScript first-strand synthesis system for RT-PCR ( Invitrogen ) was used to synthesize cDNA from human testis total RNA ( Clontech , 636533 ) . PCR Supermix ( Invitrogen ) or Ex Taq polymerase ( Takara ) was used to amplify from cDNA . Individual PCR amplicons were cloned into vectors using the TOPO-TA Cloning kit ( Invitrogen ) . For each sample , at least ten different colonies were randomly selected and were sequenced . These sequences have been submitted as records to Genbank ( JF968445-JF968463 ) . A sequence alignment was created from TRIM genes that have a full-length open reading frame ( RING through B30 . 2 ) . The tree length of this dataset is approximately 5 [73] . First , this alignment was checked for the signatures of gene conversion . If gene conversion of one paralog by another has occurred along the entire length of the two genes , this will not present a problem because a gene phylogeny will correctly reflect the fact that these two genes now have a very recent common ancestor and have been diverging only since the gene conversion event ( although the record of previous evolutionary adaptations will have been erased in the converted gene ) . Problems occur when a gene conversion event has affected only part of a gene , as each gene half will then have a different location on the phylogenetic tree and no single tree will accurately represent the evolutionary history of the entire gene . To detect such events , the alignment was checked for phylogenetic incongruencies with the GARD program [49] implemented in Datamonkey [71] . Once the breakpoint had been identified , the tree structure for each half of the alignment was checked with multiple bootstrapping algorithms using MEGA5 as described above . Using the two halves of the alignment and the corresponding trees , maximum likelihood analysis was performed with codeml in the PAML 3 . 14 . 1 software package [72] . The multiple alignments were fitted to the NSsites models M1a , M7 , M8a ( null models ) and M2a , M8 ( positive selection models ) . Simulations were run with alternate models of codon frequencies ( f3x4 and f61 ) , and with multiple seed values for dN/dS ( ω ) . Likelihood ratio tests were performed to assess whether positive selection models provide a significantly better fit to the data than null models . In situations where the null model could be rejected ( p<0 . 05 ) , posterior probabilities were assigned to individual codons belonging to the class of codons with dN/dS>1 with the Naive Empirical Bayes ( NEB ) algorithm implemented in codeml . The free ratio model ( model 1 , one dN/dS per branch ) was also run in codeml to assess branch-specific values of dN/dS . HA-tagged versions of human and rhesus TRIM5 in the LPCX retroviral vector were obtained from the National Institutes of Health AIDS Research and Reference Reagent Program . TRIM A1 , B1 , F1 , F2 , and F3 open reading frames were amplified from human cDNA using primers shown in Table S7 . We were unable to amplify B5 in its full length form , so in this case we fused the B30 . 2 domain of this gene to the tripartite domains of rhesus TRIM5 . HA tags were fused to the C-terminus of each gene using PCR and these products were cloned into the LPCX retroviral vector ( Clontech ) . Retroviruses containing these vectors were packaged in 293T cells by co-transfecting them along with the NB-MLV packing plasmid pCS2-mGP [74] and pC-VSV-G ( provided by Hyeryun Choe ) . Supernatants were collected and used to infect CRFK cells purchased from American Type Culture Collection ( ATCC ) and grown in DMEM supplemented with 10% FBS . After 24 hours , media containing 8 µg/ml puromycin was added to select for transduced cells . Expression of TRIM proteins was detected by Western blot of 30 µg total protein using an anti-HA antibody ( 3F10 , Roche , catalog 1867431 ) . Viruses for single-cycle infection assays were packaged in 293T cells by co-transfection of plasmids encoding viral proteins and VSV-G , along with a transfer vector , as follows: N-MLV ( pCIG3-N [75] , pC-VSV-G , pLXCG:GFP ) , HIV-1 ( pMDLg/pRRE , pRSV-Rev , pMD2 . G , pRRLSIN . cPPT . PGK-GFP . WPRE; all available on Addgene ) , FIV ( pFP93 [76] , pC-VSV-G , pGINSIN:GFP [77] ) . After 48 hours , supernatant containing viruses was harvested , filtered , and frozen . For infection assays , CRFK stable cell lines were plated at a concentration of 5×104 cells/well in a 12-well plate and infected with N-MLV , FIV , or HIV-1 the following day . Two days post-infection , cells were analyzed by flow cytometry for expression of GFP . We utilized the SALSA MLPA kit P200 Human DNA reference-1 and associated protocol ( MRC-Holland , Amsterdam , The Netherlands ) . This kit includes the human control probes utilized . Our custom probe set was designed to contain eighteen pairs of MLPA probes spanning segment 1 ( Table S4 ) . These probes also perfectly match paralogous regions in segment 3 , due to the fact that these segments are nearly identical , but are designed to contain at least two mismatches to all other paralogous sequences located on chromosomes 2 or 11 ( or anywhere else in the human genome , as determined by BLAT searching on the UCSC human genome browser ) . Probes were positioned both in genes ( 8 probe pairs ) and in intergenic regions ( 10 probe pairs ) . The average distance between probe pairs is 9304 bp . PCR primers supplied with this kit were fluorescently labeled with FAM , and FAM-labeled fragments obtained from each experiment were analyzed with an Applied Biosystems 3730 DNA analyzer . Peak spectra were checked for quality in two ways . First , the spectra were analyzed with the ABI software Peak Scanner ( v . 1 . 0 ) to evaluate the fragment size quality using a size standard that was included during fragment analysis ( 500ROX , ABI ) . If the quality flag indicated “pass , ” the samples including their fragment size information were exported as a combined table . Second , the MRC-Holland software Coffalyzer ( v . 8 ) was used to evaluate the signals of the control probes supplied with the MLPA kit . Controls are designed to confirm sufficient amounts of template DNA and completion of DNA denaturation and ligation steps . Finally , GeneMarker ( v . 1 . 7 ) software was used to normalize and analyze MLPA experiments that passed both of these quality control steps . Advanced population normalization was used and MLPA analysis settings were as follows: MLPA ratio ( analysis method ) , adjustment by control probes , and quantification by peak height . After normalization of fragment data to the reference genome ( sample NA10851 from Utah ) , duplications and deletions were defined as probes that gave a signal intensity of >1 . 35 ( duplication ) or <0 . 65 ( deletion ) that of the reference genome . Because the samples analyzed were a mixture of male and female samples , control probes on the X and Y chromosomes were used to show that these enrichment and depletion thresholds are robust in predicting gain and loss of control targets located on sex chromosomes ( Table S5 ) . Because false signals may be caused by unknown SNPs at the target locus or elsewhere , signals of enrichment or depletion seen only with a single probe pair were disregarded . | A fundamental question in biology is how the immune system is able to inactivate the enormous number of pathogens that it faces . The vast majority of pathogens are quickly neutralized by the innate immune system , a large network of defenses to which approximately 1/30 of the human genome is devoted . Because pathogens are always evolving , these innate immunity genes must be able to acquire new specificities . Here we illustrate a novel mechanism of evolution that has been employed by the large family of TRIM innate immunity genes . We have found a cluster of tandemly arranged TRIM genes on chromosome 11 that serves as a “reservoir” from which new TRIM genes constantly arise . We show that this gene cluster is prone to spawning duplications of itself , allowing primate genomes to continually acquire new TRIM gene copies that can presumably be selected to combat present and new pathogens . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"biology",
"genomics",
"evolutionary",
"biology",
"genetics",
"and",
"genomics"
] | 2011 | Identification of a Genomic Reservoir for New TRIM Genes in Primate Genomes |
Plasmodium falciparum employs antigenic variation to evade the human immune response by switching the expression of different variant surface antigens encoded by the var gene family . Epigenetic mechanisms including histone modifications and sub-nuclear compartmentalization contribute to transcriptional regulation in the malaria parasite , in particular to control antigenic variation . Another mechanism of epigenetic control is the exchange of canonical histones with alternative variants to generate functionally specialized chromatin domains . Here we demonstrate that the alternative histone PfH2A . Z is associated with the epigenetic regulation of var genes . In many eukaryotic organisms the histone variant H2A . Z mediates an open chromatin structure at promoters and facilitates diverse levels of regulation , including transcriptional activation . Throughout the asexual , intraerythrocytic lifecycle of P . falciparum we found that the P . falciparum ortholog of H2A . Z ( PfH2A . Z ) colocalizes with histone modifications that are characteristic of transcriptionally-permissive euchromatin , but not with markers of heterochromatin . Consistent with this finding , antibodies to PfH2A . Z co-precipitate the permissive modification H3K4me3 . By chromatin-immunoprecipitation we show that PfH2A . Z is enriched in nucleosomes around the transcription start site ( TSS ) in both transcriptionally active and silent stage-specific genes . In var genes , however , PfH2A . Z is enriched at the TSS only during active transcription in ring stage parasites . Thus , in contrast to other genes , temporal var gene regulation involves histone variant exchange at promoter nucleosomes . Sir2 histone deacetylases are important for var gene silencing and their yeast ortholog antagonises H2A . Z function in subtelomeric yeast genes . In immature P . falciparum parasites lacking Sir2A or Sir2B high var transcription levels correlate with enrichment of PfH2A . Z at the TSS . As Sir2A knock out parasites mature the var genes are silenced , but PfH2A . Z remains enriched at the TSS of var genes; in contrast , PfH2A . Z is lost from the TSS of de-repressed var genes in mature Sir2B knock out parasites . This result indicates that PfH2A . Z occupancy at the active var promoter is antagonized by PfSir2A during the intraerythrocytic life cycle . We conclude that PfH2A . Z contributes to the nucleosome architecture at promoters and is regulated dynamically in active var genes .
Plasmodium falciparum causes the majority of malaria-induced morbidity and mortality , resulting in approximately 860 , 000 deaths each year [1] . Plasmodium parasites have a complicated life cycle , during which they differentiate into several morphologically distinct asexual and sexual stages in the human host and the mosquito vector . Disease occurs during the repeated cycles of invasion and asexual replication of the parasite inside human erythrocytes . A central mechanism of malaria pathogenesis is the ability of the infected erythrocytes ( IE ) to sequester at vascular sites by cytoadherence to host receptors . Through this process the parasite avoids clearance by splenic macrophages and contributes to severe malaria complications such as cerebral and placental malaria [2] . Sequestration is mediated by the P . falciparum erythrocyte membrane protein 1 ( PfEMP1 ) variant antigens , which are expressed on the IE surface and are encoded by approximately 60 genes of the var multigene family [3] , [4] , [5] , [6] . Only a single var gene is expressed at a time [7] , [8] and switching between expression of different PfEMP1 variants alters both the cytoadherence phenotype and the antigenic profile of the IE , resulting in antigenic variation [9] . These processes are critical to immune evasion of P . falciparum and chronic infection [10] , [11] , [12] . During its asexual , intaerythrocytic life cycle , P . falciparum employs stringent regulatory mechanisms to achieve stage-specific gene expression [13] , [14] . The multiple layers of regulation that mediate morphological and physiological adaptations , include specific transcription factors and repressors [15] , [16] , [17] , [18] , translational repression [19] , [20] , post-translational protein modifications ( reviewed [21] ) and epigenetic mechanisms [22] , [23] . Recent studies assessing the global structure of chromatin during the asexual intra-erythrocytic developmental cycle ( IDC ) have shown that chromatin undergoes massive reorganization in P . falciparum , emphasizing the significance of epigenetic control in the parasite [24] . Epigenetic gene regulation confers a heritable state of gene expression and is typically mediated by changes in chromatin structure without a change in DNA sequence . The fundamental units of chromatin are nucleosomes and are formed by ∼146 bp of DNA wrapping around an octamer of histones . The canonical nucleosome components are two H2A/H2B dimers and an H3/H4 tetramer . In most eukaruyotes an additional subunit , H1 , links nucleosomes , but this subunit is missing in Plasmodium species . Chromatin can exist as compact , silent heterochromatin and as open , transcriptionally competent euchromatin . These different physical states are essentially determined by the composition and distribution of nucleosomes , post-translational modification of histones , presence of chromatin associated trans-factors and by the covalent modification of DNA , although this latter mechanism has not yet been shown in Plasmodium . The Plasmodium genome encodes a broad set of common chromatin remodelling and modifying factors , including many that are novel and that may be employed in unique epigenetic mechanisms [25] , [26] , [27] . The organisation of P . falciparum chromatin has unique features that differ markedly from other eukaryotes . P . falciparum intergenic regions , including promoters , display a global nucleosome depletion [24] , [28] . Consequently , and unlike in other eukaryotes , nucleosome occupancy at P . falciparum promoters does not correlate inversely with steady-state mRNA levels [29] , [30] , [31] . Genome-wide analysis of histone modifications has shown that the blood stage genome of P . falciparum exists in an unusually euchromatic state conferred by the euchromatin marks trimethylated lysine 4 of histone 3 ( H3K4me3 ) and acetylated lysine 9 of histone 3 ( H3K9ac ) [22] , [32] . In contrast to the situation in yeast and other eukaryotes , the presence of neither H3K4me3 nor H3K9ac enrichment seems to correlate with gene activity throughout most of the intraerythrocytic cycle , except for late stage schizont parasites when these marks are enriched at the 5′ ends of active genes [22] . P . falciparum heterochromatin , defined by the histone post-translational modification H3K9me3 and its cognate trans-factor heterochromatin protein 1 ( HP1 ) , is restricted to subtelomeric and several chromosome internal domains in P . falciparum that contain gene families including the majority of var genes [22] , [33] , [34] , [35] , [36] . The var genes present both in subtelomeric and central chromosomal positions form clusters at the nuclear periphery [37] , [38] , [39] and expression of a var gene appears to require it to leave the heterochromatic cluster and relocate to a specific perinuclear expression site [35] , [38] , [40] , [41] . The variegated , monoallelic expression of var genes is controlled by epigenetic mechanisms [42] . To become activated in ring stages , a var gene must acquire the histone marks H3K4me3 and H3K9ac in its promoter [34] , [35] . The var promoter is the only cis element required for monoallelic exclusive transcription [8] but promoter activity of the var gene intron is important for var gene silencing [43] , [44] , [45] , [46] and additional cis sequence elements contribute to the rate of var gene switching [47] . A critical role of histone acetylation in var gene regulation was proven by the upregulation of numerous var genes and the loss of monoallelic var gene expression that occurred when either of two Sir2 histone deacetylase genes were disrupted [41] , [48] . Alternative histones can replace canonical histones through ATP-dependent deposition to create structurally and functionally specialized chromatin domains [49] . While the importance of histone modifications to P . falciparum gene regulation is apparent , the role of alternative histones has not yet been investigated . H2A . Z is an H2A variant , which is essential for viability in most organisms apart from yeast and has been highly conserved through evolution [50] , [51] , [52] , [53] . H2A . Z has been implicated in the regulation of very diverse processes such as heterochromatin formation , chromosome stability and segregation , proliferation and transcriptional activation or repression [49] . Its role in transcriptional regulation has been clearly established across species , but the underlying mechanisms are not well understood and reported effects of H2A . Z deposition are contradictory . Genome wide analyses consistently revealed that H2A . Z is enriched in nucleosomes near the transcriptional start site ( TSS ) in RNA polymerase II promoters [54] , [55] . In human cells and in the protist parasite Toxoplasma gondii , H2A . Z enrichment at promoters ( and enhancers in humans ) correlates with gene activity [54] , [56] , whereas in yeast H2A . Z was found to occupy the promoters of active as well as poised genes from where it is lost with active transcription [55] , [57] . In this study , we provide the first characterization of the alternative histone H2A . Z in P . falciparum . Using immunofluoresecence analysis ( IFA ) , co-immunoprecipitation ( Co-IP ) and chromatin immunoprecipitation ( ChIP ) in conjunction with quantitative real time PCR ( qPCR ) , we show that PfH2A . Z is enriched in the promoter of a set of developmentally regulated genes in the euchromatin compartment . In these genes promoter occupancy of PfH2A . Z does not correlate with transcription levels , suggesting that presence of PfH2A . Z is important for providing a transcriptionally competent chromatin structure at the promoter but does not directly relate to promoter activity . In var genes by contrast , PfH2A . Z promoter occupancy is strongly associated with transcriptional activity , as PfH2AZ is periodically enriched in the active var gene and is depleted in silent var genes , which is consistent with our IFA evidence that H2A . Z is absent from the subtelomeric heterochromatin compartment . This balance is distorted in cells in which the histone deacetylase Sir2A has been disrupted . Thus , these data suggest that var gene silencing requires expulsion of PfH2A . Z from the var promoter and that this involves histone deacetylation .
The P . falciparum H2A . Z variant ( PfH2A . Z ) is encoded by the gene PFC0920w [58] . Sequence alignments show that PfH2A . Z shares 56 . 4% amino acid identity with the S . cerevisiae H2A . Z variant Htz1 and 67 . 3% identity with the human and mouse H2A . Z proteins ( Figure S1 ) , whereas conservation between other organisms was reported to be 70%–90% [59] . The major divergences affect the extended and highly charged N-terminus of PfH2A . Z , which is characterized by an accumulation of lysine residues and a three-fold repetition of the peptide sequence GGKV ( position 9–20 ) . Mass spectrometric evidence indicates that the N-terminus of PfH2A . Z can be acetylated in seven lysine residues , which would partially neutralize the positive charge [60] . In loop 1 , two threonine residues implicated in mediating H2A . Z/H2A . Z self-interactions and prohibiting dimerization with H2A [61] are changed to isoleucine and serine ( PfH2A . Z Ile 65 and Ser 66 ) , respectively , and a conserved positively charged residue is substituted by aspartic acid ( PfH2A . Z Asp 68 ) introducing a negative charge in this region . In the C-terminal docking domain all amino acids critical for interaction with the H3/H4 dimer are conserved [61] . To analyse PfH2A . Z , antiserum against recombinant full length PfH2A . Z was generated . The serum specifically recognizes PfH2A . Z but not H2A and cross-reacts with human H2A . Z , which has a different N terminal acetylation pattern , indicating reactivity of the antibody with the non-acetylated , conserved C-terminal domains ( Figure 1A ) . We further showed that anti-PfH2A . Z immunoprecipitates both acetylated and non-acetylated forms of PfH2A . Z because antibodies specific for an acetylated peptide present in the PfH2A . Z and H4 N-termini [61] labels immunoprecipitated PfH2A . Z by Western Blot ( Figure 1B ) . We found that PfH2A . Z is present throughout the asexual life cycle ( Figure 1C & D ) . In comparison to the canonical histone H3 , PfH2A . Z abundance increases significantly in the schizont stage ( Figure 1C ) . This is consistent with an increase in PfH2A . Z mRNA observed in schizonts [58] . We confirmed that PfH2A . Z is localized in the nucleus in all asexual stages by immunofluorescence analyses ( IFA ) and confocal microscopy . An area stained with the DNA dye DAPI but devoid of PfH2A . Z labelling was consistently observed , indicating that PfH2A . Z is enriched towards one side of the nucleus ( Figure 1D ) . 3D reconstruction verified this polarized localization of PfH2A . Z ( Video S1 ) and transgenic P . falciparum ectopically expressing PfH2A . Z-GFP fusion proteins corroborated the sub-nuclear distribution of PfH2A . Z ( Figure S2 ) . A similar cap-like pattern has previously been reported for the euchromatic histone mark H3K4me3 [62] . H2A . Z has been shown to contribute to diverse biological processes associated with different chromatin compartments , such as gene activation and poising [63] , [64] , [65] , [66] , chromosome segregation [67] , [68] and heterochromatin structure [69] , [70] , [71] , [72] , [73] . To investigate the chromatin association of PfH2A . Z we performed co-localization experiments with well-characterized chromatin marks . Double staining showed good overlap between PfH2A . Z and the euchromatin marks H3K4me3 and H3K9ac ( Figure 2A ) , which are enriched across the P . falciparum genome [22] . In contrast , PfH2A . Z staining was distinct from the subtelomeric heterochromatin marks H3K9me3 and HP1 ( Figure 2A ) . Consistent with these results , immunoelectron microscopy indicated that PfH2A . Z was not restricted to the nuclear periphery , where inactive subtelomeric and internal var genes cluster . However , its distribution appeared concentrated in certain subnuclear compartments , frequently at the border of the electron lucent and electron dense nuclear material that is presumed to represent euchromatin and heterochromatin , respectively ( Figure 2B ) [38] . In mammalian cells it was shown that H2A . Z-containing nucleosomes preferentially carry the euchromatic mark H3K4me3 [73] , [74] . Consistent with this , we found by co-immunoprecipitation experiments with anti-PfH2A . Z that mononucleosomes containing PfH2A . Z are highly enriched in H3K4me3 ( Figure 2C ) . Together , these data support that PfH2A . Z is functionally linked to euchromatin but is largely depleted from subtelomeric heterochromatin . To identify the genomic target sites of PfH2A . Z and investigate how the presence of PfH2A . Z correlates with transcriptional activity , we performed chromatin immunoprecipitation ( ChIP ) followed by quantitative PCR in ring stage parasites , trophozoites and schizonts . H2A . Z is enriched in nucleosomes surrounding the transcription start site ( TSS ) in other organisms , therefore at least two quantitative ( q ) PCR reactions were performed for each gene to amplify regions near the predicted TSS upstream of the start codon ( ups ) as well as in the open reading frame ( orf ) . Our results show that PfH2A . Z is enriched in all three stages in the promoter regions of candidate genes that are differentially regulated throughout the life cycle ( Figure 3A , S3A ) . In contrast to PfH2A . Z , H2A showed equal distribution in the promoter and the open reading frame of the investigated genes ( Figure 3B ) . In rings , promoter enrichment of PfH2A . Z was highest in genes that are constitutively expressed ( e . g . HSP70 and casein kinase ) or induced during asexual intra-erythrocytic differentiation ( e . g . schizont genes MSP2 and Eba175 ) , but also clearly apparent in silent genes ( e . g . sporozoite genes CSP and SSP2 ) . In trophozoites and schizonts , PfH2A . Z enrichment in the promoter showed similar levels across all genes . To determine whether the level of PfH2A . Z promoter occupancy in the examined genes correlated with gene expression , mRNA levels were quantified by q-RT-PCR ( Figure S4 ) . The ups/orf ratio of PfH2A . Z enrichment was determined and plotted against the relative expression levels . No significant correlation between PfH2A . Z promoter occupancy and transcription level could be observed at any stage ( Spearman correlation , p>0 . 5 ) . To further investigate the relationship between PfH2A . Z and euchromatic and heterochromatic histone marks , ChIP was performed in parallel for PfH2A . Z as well as H3K4me3 , H3K9ac and H3K9me3 . PfH2A . Z and the two euchromatic marks , H3K4me3 , H3K9ac all showed significant enrichment in the ups region when compared to the orf in ring and schizont stage parasites , in contrast the levels of the heterochromatic mark H3K9me3 were the same in ups and orf ( Figure S5 ) . This is consistent with previous work on the histone marks [22] and supports our finding that PfH2A . Z and H3K4me3 are present in the same nucleosomes ( Figure 2C ) . Furthermore , we found that the PfH2A . Z enrichment level in the upstream region of genes positively correlates with both euchromatic marks ( P<0 . 0001 ) , whereas a negative correlation was evident between PfH2A . Z and H3K9me3 ( p<0 . 0001 ) ( Figure S6 ) . Together , these results demonstrate that PfH2A . Z is enriched near the TSS in genes , independently of their transcriptional activity , and that PfH2A . Z enrichment near the TSS correlates with enrichment of the euchromatin marks H3K4me3 and H3K9ac . The P . falciparum histone deacetylases Sir2A and Sir2B silence subtelomeric and central var genes [41] , [48] , [75] and in S . cerevisiae H2A . Z antagonises subtelomeric gene silencing by Sir2 [53] . Therefore we investigated whether PfH2A . Z was involved in the transcriptional control of var genes . We harvested chromatin and RNA from a parasite culture that had been selected for the expression of a single var gene encoding VAR2CSA by panning on chondroitin sulphate A ( CSA ) . To map the position of PfH2A . Z along the var2csa gene , we designed seven qPCR reactions spanning this var locus . The TSS of var2csa has previously been mapped to −1475 bp upstream of the start codon in the FCR3 parasite line [34] and is predicted to be located at approximately −1200 bp in 3D7 ( PlasmoDB ) . Three primer pairs amplified regions in the non-coding upstream region ( -1500 bp , −1000 bp , −575 bp ) , and four primer pairs targeted areas along the coding region . ChIP analysis showed that in ring stage parasites , when var2csa transcription peaks ( Figure S4 ) [76] , PfH2A . Z is strongly enriched in the areas flanking the predicted TSS ( −1500 and −1000 bp ) , but not further downstream ( −575 bp ) or within the open reading frame ( ATG , DBL3 , DBL6 ) ( Figure 4A , S3B ) . As the parasites progress through the trophozoite to the schizont stage var transcription declines and PfH2A . Z enrichment around the TSS decreases . No enrichment of PfH2A . Z was detectable around the TSS of the var2csa gene at schizont stage . This stage-specific deposition of PfH2A . Z in the var gene promoter contrasts with the continuous presence of PfH2A . Z in the promoters of the limited number of other genes we analysed ( Figure 3A ) indicating that PfH2A . Z deposition differs in its temporal regulation in var genes . To directly compare PfH2A . Z occupancy of an active and inactive var promoter we performed ChIP experiments on unselected 3D7 parasites that do not transcribe the var2csa gene ( var2csa OFF ) , and var2csa expressing parasites ( var2csa ON ) at ring stage . Transcription levels were verified by q-RT-PCR ( Figure S4 ) . In contrast to the strong enrichment of PfH2A . Z in the active var promoter , the alternative histone was clearly not enriched when var2csa was not transcribed . These results demonstrate that PfH2A . Z promoter occupancy in var2csa strongly correlates with transcription ( Figure 4B ) . To verify our observation that PfH2A . Z occupancy is restricted to the active var TSS in a second var gene , 3D7 parasites were selected on ICAM1 and the gene PFL0020w was identified as the dominant transcript by q-RT-PCR and Northern Blot analysis ( Figure S7 A , B ) . ChIP analysis across the gene confirmed increased PfH2A . Z occupancy near the PFL0020w TSS in ICAM1 selected parasites as compared to non-selected parasites at ring stage . In line with our results with var2csa , no enrichment was evident in schizonts ( Figure S7 C , D ) . Var genes possess a second promoter , which is situated in the conserved intron . From this promoter , truncated sense and antisense transcripts are synthesized which are thought to contribute to heterochromatin structure and var gene silencing [77] . To investigate a possible association of PfH2A . Z with the var intron promoter , we designed primer pairs targeting the var introns and repeated the ChIP experiment in var2csa expressing and non-expressing ring stage parasites ( Figure 4C ) . Consistent with the previous observations , PfH2A . Z was enriched around the TSS of the active var2csa gene ( var2csa ON ) , but not the TSS of inactive var2csa ( var2csa OFF ) nor the TSS of the other two silent var genes PF08_0141 ( var41 ) and PFL0020w ( var20 ) . In contrast to the upstream region , all three var genes were moderately enriched in PfH2A . Z in the introns ( Figure 4C ) . This was also evident in the ICAM selected parasite line ( Figure S7 C&D ) . Analysis of intron/orf pairs from ten different var genes confirmed that the enrichment of PfH2A . Z in the intron was statistically significant ( Figure 4D ) . Because H2A . Z may function as a barrier to prevent the spread of Sir2-mediated silencing in yeast [53] we further investigated the relationship between Sir2 and PfH2A . Z in the control of var gene expression in P . falciparum . ChIP and expression analyses were performed on ring and schizont stage parasites in which Sir2A or Sir2B had been disrupted ( 3D7Δsir2A and 3D7Δsir2B ) [41] , [48] . By q-RT-PCR we first monitored the expression profiles of all var genes in ring stages of both 3D7Δsir2 parasite lines ( Figure S8 ) . With the aim to understand how each Sir2 paralogue influences PfH2A . Z deposition and how this correlates with var transcription , we selected five var genes that were highly expressed and five var genes that were lowly expressed in 3D7Δsir2A parasites for further analysis , all of which had previously been shown to be regulated by Sir2A [41] , [48] . We used the same strategy to choose ten var genes previously shown to be regulated by Sir2B [48] for analysis in 3D7Δsir2B . We then analysed PfH2A . Z deposition by ChIP and qPCR in upstream and coding regions of these var genes in knock out and wild type parasites . The ups/orf ratios were determined and compared between 3D7Δsir2A or 3D7Δsir2B and 3D7 parasites , respectively ( Figure 5 ) . Although ups/orf ratios were generally quite low in ring stages , PfH2A . Z enrichment at var promoters was significantly greater in the highly expressed var genes in the 3D7Δsir2A and B lines than in the 3D7 control ( Mann-Whitney , p = 0 . 0232 ( Δsir2A ) and p = 0 . 0152 ( Δsir2B ) ) ( Figure 5A & C ) . In schizonts , no enrichment of PfH2A . Z upstream of active Sir2B regulated genes was detected ( Figure 5D ) . This result is consistent with a role of PfH2A . Z in active transcription of var genes . However , interestingly PfH2A . Z occupancy in the upstream region of var genes was maintained and even elevated at schizont stage in 3D7Δsir2A parasites at significantly higher levels than in 3D7 ( Mann-Whitney , p = 0 . 0159 ) ( Figure 5B ) , although var gene expression is down-regulated in mature 3D7Δsir2 parasites ( data not shown ) [48] . This result indicates a link between Sir2A and loss of PfH2A . Z and implicates PfH2A . Z with maintenance of the integrity of the heterochromatin/euchromatin boundary at Sir2A regulated loci . To further investigate the atypical PfH2A . Z enrichment in Sir2A regulated var genes in 3D7Δsir2A schizonts we analysed by ChIP the relationship between enrichment of PfH2A . Z and H3K9me3 or H3K4me3 , respectively ( Figure S9 ) . Similar to actively transcribed var genes in immature wildtype parasites ( Figure S6 ) , we found a positive correlation between PfH2A . Z and H3K4me3 enrichment ( Spearman correlation , p = 0 . 0005 ) and a . negative correlation between PfH2A . Z and H3K9me3 ( Spearman correlation , p = 0 . 0182 ) , which was consistent with the previously described depletion of H3K9me3 in var genes upregulated in 3D7Δsir2A parasites [35] .
H2A . Z is essential in many eukaryotes including Trypanosoma brucei , Tetrahymena thermophilus , Drosophila , Xenopus , and vertebrates [50] , [52] , [59] , [78] , [79] and is one of the structurally most conserved histones throughout evolution [80] . However , Plasmodium H2A . Z differs significantly from its orthologues , particularly in its extended N-terminus which contains seven lysine residues that can be acetylated [58] , [60] , as opposed to five in humans and four in yeast . In T . thermophilus the N-terminus of H2A . Z has been implicated in directly interacting with the DNA [81] . Acetylation of at least one lysine residue is essential for viability , and probably acts by reducing the positive charge and thereby weakening H2A . Z–DNA interactions [82] . Interestingly , T . thermophilus H2A . Z encodes a repeated GGK motif similar to the one observed in PfH2A . Z , suggesting that a similar mechanism may apply for regulating PfH2A . Z-DNA interactions in the malaria parasite . By western blot we observed increased PfH2A . Z abundance in the schizont stage relative to H3 . Interestingly , PfH2A . Z accumulation coincides with expression of the ATP-dependent chromatin remodeling factor PfSwr1 ( unpublished data ) , orthologues of which mediate the post-replicative H2A . Z incorporation into nucleosomes in yeast and humans [83] , [84] , [85] . The Plasmodium genome becomes densely packed with nucleosomes in the late schizont stages [24] , and the concurrent increase in PfH2A . Z may reflect a rising requirement for the alternative histone in the intergenic regions . H2A . Z facilitates intra-molecular folding of nucleosomal arrays into a 30 nm fibre [86] , so it may play a role in chromatin condensation in the mature schizonts . Enrichment of H2A . Z in RNA polymerase II promoters has been conserved through the evolution of eukaryotes as diverse as yeast , humans and the protist parasites T . brucei and T . gondii [29] , [53] , [54] , [55] , [56] , [57] , [78] , [87] , [88] . Our ChIP analysis showed that P . falciparum conforms to this pattern , with an enrichment of PfH2A . Z , but not H2A , in the upstream regions of genes ( Figures 3 , 4 ) . Consistent with this finding , we co-precipitated PfH2A . Z and H3K4me3 ( Figure 2 ) , which is enriched in 5′-upstream regions of several organisms including P . falciparum [22] , [54] , [89] . Our study also indicates that the pattern of PfH2A . Z occupancy at euchromatic gene promoters remains relatively stable throughout the IDC ( Figure 3 ) , in contrast to the reported fluctuations in levels of the transcriptionally permissive histone modifications H3K4me3 and H3K9ac [22] . Further substantiating an association of PfH2A . Z with promoter architecture , we showed that PfH2A . Z occupation in upstream regions correlates with H3K4me3 and H3K9ac enrichment ( Figures S5 & 6 ) . With the exception of var genes , PfH2A . Z enrichment did not correlate with mRNA levels in our experiments . A global enrichment of H2A . Z at active and inactive promoters has also been observed in yeast [55] , whereas some other studies reported a negative correlation with transcription and proposed that H2A . Z poises inducible silent genes for activation and is subsequently evicted during transcription [29] , [57] , [88] . In humans , H2A . Z is either present in active gene promoters in differentiated cells [54] , [90] or marks poised gene promoters in hematopoietic stem cells [91] . These conflicting results demonstrate that the role of H2A . Z in transcription is very complex and the underlying mechanisms remain enigmatic . In both yeast and humans H2A . Z assists in RNA polymerase II recruitment [90] , [92] . Components of the preinitiation complex are pre-assembled in some erythrocytic stage P . falciparum promoters regardless of gene activity [93] , similar to the PfH2A . Z enrichment at the TSS shown here . Thus PfH2A . Z may contribute to the open chromatin structure necessary for preinitiation complex formation . But how could PfH2A . Z modulate gene activity despite global occupancy at promoters ? Mass spectrometric evidence that PfH2A . Z can be heavily acetylated at multiple lysine residues in the N-terminus suggests that the neutralization of positive charges by lysine acetylation may facilitate an open chromatin structure making the DNA more accessible [58] , [60] . In our experiments , total PfH2A . Z occupancy was monitored because our antiserum did not differentiate between acetylated and non-acetylated forms of PfH2A . Z ( Figure 1 ) . Acetylation may thus provide a functional switch necessary to promote transcription , as has been suggested for yeast and humans [64] , [66] , [94] , [95] . Histone acetylation normally promotes gene activation , whereas sumoylation and ubiquitination are post-translational modification associated with recruitment of histone deacetylase complexes and transcriptional repression [96] , [97] . PfH2A . Z has six potential sumoylation sites [98] and may also be ubiquitinated in the C-terminal domain , although this has not yet been shown . In mammalian cells , mono-ubiquitinated H2A . Z is enriched in the facultative heterochromatin that constitutes the inactive X-chromosome [73] . Thus , acetylation , ubiquitination and sumoylation represent interesting candidate regulators of PfH2A . Z function at promoters . In Arabidopsis thaliana and yeast it has recently been shown that H2A . Z mediates a thermo-sensory response and facilitates differentiation processes by regulating transcription [99] . This is thought to be due to reduced DNA wrapping of H2A . Z containing nucleosomes at higher temperatures , resulting in a relaxed chromatin structure that permits of transcription . It is tempting to speculate that PfH2A . Z may function as a similar physical switch to control gene expression in response to temperature change , for example during fever or as P . falciparum is transmitted between its two hosts . Apart from PfH2A . Z , P . falciparum encodes two other histone variants , H2Bv and H3 . 3 [58] , which have previously been implicated in transcriptional regulation . H2Bv is a variant unique to protist parasites and has so far only been characterized in T . brucei and T . gondii [56] , [78] , [87] . In both parasites H2Bv pairs with H2A . Z . While H2A . Z/H2Bv nucleosomes are enriched in the promoters of active genes in T . gondii [56] , they showed a global association with PolII transcription start sites in T . brucei [87] . Nucleosomes containing H2A . Z and H3 . 3 have been reported to be less stable than nucleosomes containing H2A . Z and canonical H3 in humans [100] , and nucleosomes composed of both variant histones are present in active promoters , enhancers and insulators . It was postulated that instability of H2A . Z/H3 . 3 composite nucleosomes at the TSS might facilitate access to transcription factors [101] . H3 . 3 differs only by three amino acids from canonical H3 so the antibodies directed against H3 in this study precipitated both variants . The functional cooperation of alternative histones in P . falciparum gene regulation will be an interesting field of future research . In many organisms , including Plasmodium , subtelomeric genes are subject to special mechanisms of gene regulation . This occurs via the unique subtelomeric heterochromatin which provides a specialized architecture to control variegated expression of gene family members ( reviewed in [102] , [103] ) . A link between H2A . Z and subtelomeric gene regulation has been established in S . cerevisiae as disruption of the H2A . Z orthologue htz1 results in the down regulation of many subtelomeric genes [53] . Consistent with its proposed role as an anti-silencing factor for subtelomeric genes in yeast , we found PfH2A . Z to be enriched in the subtelomeric var2csa and pfl0020w promoters when these genes were active , but not when they were silent ( Figures 4 , S7 ) . This was supported by significant enrichment of PfH2A . Z in active var genes in the two Sir2 KO lines ( Figure 5 ) . Thus , PfH2A . Z is a novel component contributing to the promoter architecture of active var genes together with the euchromatin factors H3K4me3 and H3K9ac [34] . In contrast to the other genes we examined , var genes exhibited significant temporal modulation in PfH2A . Z enrichment at the promoter . Loss of PfH2A . Z from the var promoter is observed from the trophozoite stage on , which is when DNA replication begins , and thus is consistent with S-phase dependent silencing of var genes [43] . While canonical histones are deposited into chromatin at the replication fork , deposition of histone variants can occur post-replication by ATP-dependent enzyme complexes . This allows dynamic changes to the chromatin structure at the promoter during differentiation ( reviewed in [49] ) . Based on this knowledge we speculate that PfH2A . Z is lost from the var promoter during DNA replication , possibly due to limited stability of PfH2A . Z containing nucleosomes , and is deposited at the var promoter in rings . The loss of PfH2A . Z from the active var gene promoter upon temporal silencing suggests that , unlike in subtelomeric genes in yeast [57] , [104] , PfH2AZ is not involved in epigenetic memory of var genes . Our ChIP and qPCR experiments along the entire var gene revealed that PfH2A . Z enrichment peaks at positions directly surrounding the predicted TSS of the active var promoter . This pattern is consistent with yeast where H2A . Z is highly enriched in the two nucleosomes surrounding a nucleosome free-region at the TSS , and with humans where it is restricted to the two nucleosomes in the −1 and +1 position as well as at the TSS itself [55] , [101] . The distinctive pattern of PfH2A . Z enrichment we observed within the var intron corroborates the existence of a promoter-like chromatin structure at this site . The var gene intron contains a bi-directional promoter from which sense and antisense transcripts are synthesized in both active and inactive var genes during late erythrocytic stages , indicating that this mechanism might function in the control of transcriptional timing rather than monoallelic expression [38] , [77] . The impact of the regulatory activity of the var gene intron upon local chromatin structure has been previously indicated through nucleosome depletion within the intron [24] , [28] . While PfH2A . Z occupancy at the var upstream promoter could only be detected when var2csa was actively transcribed , enrichment at the conserved intron promoter occured in both active and inactive var genes . This is consistent with the previously observed lack of correlation between activity of the var intron promoter and of the upstream var promoter [38] , [77] . In yeast , H2A . Z acts as a boundary element that antagonizes the spread of Sir2-mediated heterochromatin into euchromatic areas [53] , [105] . This is consistent with our findings by IFA , CoIP and ChIP revealing that PfH2A . Z is associated with regions of euchromatin but is absent from the heterochromatin compartment that is characterized by H3K9me3 and HP1 deposition and contains silent var genes ( Figures 2 , 3 , 4 ) . Interestingly , our electron microscopy analysis suggests enrichment of PfH2A . Z at the border between electron light and electron dense nuclear material ( Figure 2 ) , which has previously been interpreted to represent eu- and heterochromatin , respectively [38] . This raises the prospect that PfH2A . Z may also form a barrier to the spread of Sir2-mediated heterochromatin . This is supported by our finding that in 3D7Δsir2A parasites PfH2A . Z is enriched in the upstream region of var genes that are highly transcribed not only when the var genes are active in ring stages but also when they are silent in mature schizonts ( Figure 5 & [41] , [48] ) . In contrast PfH2A . Z is enriched in the upstream region of highly expressed var genes in 3D7Δsir2B parasites only while the var genes are active in the ring stages . Sir2A-regulated var genes occupy the heterochromatin/euchromatin boundary at both chromosome internal and subtelomeric clusters; in contrast Sir2B regulates the most telomere proximal upsB type var genes that are separated from the rest of the chromosome by the Sir2A regulated subtelomeric upsA var genes [41] , [48] . This raises the intriguing possibility that antagonism between Sir2A and PfH2A . Z is involved in maintaining the heterochromatin/euchromatin boundary but Sir2B is not . Future experiments testing the enrichment of PfH2A . Z at boundary sites using ChIP will address this question . Possibly PfSir2A is not only involved in maintaining var gene silencing in heterochromatin by removing activating histone acetylations such as H3K9ac [35] , but also assists in the temporary expulsion of PfH2A . Z from the active var promoter in mature parasites ( Model in Figure 6 ) . This could occur indirectly through recruitment of the ATP-dependent chromatin remodeling machinery responsible for histone variant exchange , or directly through Sir2-mediated deacetylation of PfH2A . Z . Sir2 plays such a direct role in regulating H2A . Z levels in human myocytes where over-expressed Sir2 deacetylates H2A . Z which in turn leads to ubiquitination and proteasome-dependent degradation of H2A . Z [106] . Epigenetic gene regulation contributes significantly to the transcriptional control of fundamental mechanisms such as differentiation and antigenic variation of the malaria parasite ( reviewed in [107] ) . Here , we identify the alternative histone PfH2A . Z as a component of nucleosomes in the promoters of euchromatic genes , suggesting it may be involved in gene regulation . Our report provides the first description of histone variant exchange in P . falciparum , as we demonstrate a temporal modulation of PfH2A . Z occupancy in var genes . This mechanism is disturbed in parasites in which the histone deacetylase Sir2A is disrupted , suggesting a functional link between Sir2A and regulation of PfH2A . Z dynamics at the var promoter .
P . falciparum lines 3D7 , 3D7ΔSir2A [41] , 3D7ΔSir2B [41] and 3D7HP1GFP [33] were cultured in RPMI medium supplemented with 5% heat-inactivated human serum 0 . 25% albumax [108] . Synchronicity was maintained by weekly treatment with 5% sorbitol . Selection of var2CSA expressing parasites was performed by panning on plastic dishes coated with 50 µg/ml bovine trachea CSA ( Sigma ) , as described previously [108] . The full length PfH2A . Z ( PFC0920w ) sequence was amplified from 3D7 cDNA using oligonucleotides PfH2A . ZBamHIFor: 5′-GGGATCCGGATGGAAGTTCCAGGAAAAGT and PfH2A . ZEcoRIRev: 5′-GAATTCTTATTGAGCTGTTGGGGGAAGTG . The PCR product was cloned into pGEX-5X-3 ( GE Healthcare ) . PfH2A . ZGST fusion protein was expressed in BL21 cells in Luria Broth and induced with 1mM IPTG . Bacteria pellets were resuspended in Bugsbuster reagent and benzonase ( Novagen ) and the soluble proteins were purified on GST-Bind Resin ( Novagen ) . Antibodies were generated in rabbits by the Walter and Eliza Hall Institute Monoclonal Antibody Facility ( Bundoora , Vic . Australia ) . Primary antibodies employed in ChIP assays in this study were rabbit anti-PfH2A . Z , rabbit anti-H3 ( Abcam Ab1791 ) , rabbit anti-H2A ( Millipore 07-146 ) , rabbit anti-H2B ( Abcam Ab1790 ) , non-immune rabbit IgG ( Abcam Ab 46540 ) and pre-immune rabbit serum . Primary antibodies used for IFA co-localization and WB were rabbit anti-H3K9me3 ( Abcam Ab8898 ) , rabbit anti-H3K4me3 ( Millipore 04-745 ) and rabbit anti-H3K9ac ( Millipore 06-942 ) . Secondary antibodies for IFA were goat anti-rabbit AlexaFluor488 or chicken anti-rabbit AlexaFluor594 ( Molecular Probes ) . Secondary antibodies for WB were goat anti-rabbit HRP ( Invitrogen ) . Parasites were harvested in 8 hour intervals at 5–10% parasitemia . Lysates were generated by saponin lysis of cultures and extraction of the resulting parasite pellets with 2 x SDS PAGE loading buffer . Equivalents of 5×107 IE were separated by SDS-PAGE on 10% Bis-Tris gels ( Invitrogen ) and analysed by Western Blotting as described previously [109] . IFA was performed on paraformaldehyde/glutaraldehyde-fixed cultures as described previously [110] . For co-localization studies , histone modifications were first labelled using chicken anti-rabbit AlexaFluor594 ( 1∶1000 ) ( Molecular Probes ) as a secondary antibody . PfH2A . Z was subsequently detected with affinity purified rabbit anti-PfH2A . Z ( 10 ng/µl ) directly labelled using the Zenon Rabbit IgG labelling kit ( Invitrogen ) according to the manufacturers instructions . IE were mounted onto slides using ProLong antifade ( Invitrogen ) , left overnight to cure and analysed with an Olympus FV1000 Confocal Laser Scanning Microscope and the FluoView software . Parasites were fixed in 1% glutaraldehyde for 1 h at 4°C , dehydrated in increasing ethanol concentrations , then embedded in LR Gold resin ( Electron Microscopy Sciences , Fort Washington , PA ) . Ultrathin ( 90–100 nm ) sections were cut using a Leica Ultracut R microtome , labeled with rabbit anti-PfH2A . Z and goat-anti-rabbit IgG conjugated to 12 nm colloidal gold ( Jackson ImmunoResearch Laboratories ) . Sections were poststained with uranyl acetate and lead citrate and observed using a Philips CM120 BioTwin Transmission Electron Microscope . Chromatin was isolated at three time points during the intra-erythrocytic developmental cycle ( IDC ) from early ring ( 6–14 hpi ) , trophozoite ( 24–32 hpi ) and schizont ( 36–44 hpi ) stage parasites . Parasite cultures were cross-linked with 1% paraformaldehyde for 10 min at 37°C and the reaction subsequently quenched with 125 mM glycine . After one wash with PBS parasites were released from IE by saponin lysis , and nuclei were isolated by incubation for 30 min on ice in lysis buffer ( 10 mM Hepes pH 7 . 9 , 10 mM KCl , 0 . 1 mM EDTA , 0 . 1 mM EDTA , 1 mM DTT , 1x EDTA-free protease inhibitor cocktail ( Roche ) ) followed by dounce homogenization ( Pestle B ) . 0 . 25% NP40 was added to the parasite suspension prior to homogenization . Nuclei were pelleted by centrifugation at 21 , 000×g for 10 min at 4°C and resuspended in SDS lysis buffer ( 1% SDS , 10 mM EDTA , 50 mM Tris pH 8 . 1 , 1 x EDTA-free protease inhibitor cocktail ) . Chromatin was sheared into 200–1000 bp fragments by sonication for 2×8 min at 30 sec intervals using a Bioruptor UCD-200 ( Diagenode ) and diluted 1∶10 in ChIP dilution buffer ( 0 . 01% SDS , 1 . 1% Triton X-100 , 1 . 2 mM EDTA , 16 . 7 mM Tris pH 8 . 1 , 150 mM NaCl ) . Immunoprecipitation was performed with the EZ ChIP Kit ( Millipore ) . For each IP , 1×109 ring stage parasites , 2 . 5×108 trophozoites or 1×108 schizonts were used . Optimal antibody dilutions were determined empirically and ranged between 1∶50 and 1∶200 . Primers amplifying upstream regions near the TSS ( ups ) and in the open reading frame ( orf ) were designed using the Primer Express software ( Applied Biosystems ) ( Table S1 ) . For genes whose TSS had not previously been experimentally determined and reported , predictions were obtained from the MAPP algorithm implemented at PlasmoDB [111] . Immunoprecipitated DNA and input DNA was quantified by real time qPCR ( Applied Biosystems 7900HT ) with SYBR PCR master mix ( Applied Biosystems ) . Optimal PCR conditions were determined for each primer pair using serial dilutions of gDNA . PCR was performed in duplicates and melting curves were analysed after each run to confirm the specificity of the amplification . ChIP recoveries were normalized for input signals ( ΔCt ) and corrected for values obtained with non-immune control antibodies ( ΔΔCt ) . The site specific enrichment of histones was calculated as the 2−ΔΔCt value . To correct for differences in nucleosome density , enrichment of PfH2A . Z and H2A was expressed as a ratio over H3 . As in the course of the study different H3 antibody lots were used to precipitate biological replicates , the H3 recoveries and therefore ratios varied considerably between experiments . However , the pattern of histone variant enrichment across genes was reproducible between experiments ( Figure 3 , 4 & S4 ) . For statistical analysis of PfH2A . Z enrichment in 3D7ΔSir2A and 3D7 parasites , the ratio of PfH2A . Z enrichment near the TSS and in the open reading frame ( ups/orf ratio ) was calculated . Paired t-test was performed using the GraphPad Prism software ( Version 4 ) . Mononucleosomes were prepared from freshly isolated nuclei by MNase ( NEB ) digestion with 20 KU MNase per 1×109 IE and extraction with salt-free buffers [33] . Mononucleosomes from 2 . 5×108 IE were incubated with 10 µl of antiserum overnight at 4°C and precipitated with 30 µl protein G agarose ( Millipore ) . After extensive washing proteins were eluted with 2 x SDS PAGE loading buffer and analysed by SDS-PAGE and Western Blotting . Total RNA was harvested in parallel to the chromatin preparation by lysis of pelleted IEs in 20 pellet volumes of TRIzol ( Invitrogen ) . RNA was purified as described previously [112] and cDNA was generated using Superscript III Reverse Transcriptase ( Invitrogen ) . Quantitative RT-PCR was performed as described previously [47] using gene specific primers targeting the open reading frame , listed in Table S1 . The level of each sequence in cDNA was determined relative to its level in a constant quantity of 3D7 strain gDNA and the amounts of cDNA and gDNA were normalised using the housekeeping gene arginyl-tRNA synthetase or the sbp1 gene by 2−ΔΔCt analysis ( Applied Biosystems user bulletin 2 ) . | Plasmodium falciparum is a protist parasite that causes malaria and kills more than 800 , 000 people per year . The parasite escapes from the human immune response by antigenic variation through switching between expression of different var genes . These encode different variant antigens that are expressed on the surface of the infected erythrocyte and mediate pathogenic adhesion of the infected erythrocytes to host receptors . Understanding how this process is regulated may lead to the identification of factors that are essential for immune evasion and that could represent novel drug targets . Here , we have identified the parasite's histone variant PfH2A . Z as a novel contributor to the transcriptional regulation of antigenic variation . PfH2A . Z is enriched in the promoter of many genes , but enrichment correlates with gene expression only in var genes . Furthermore we show that PfH2A . Z enrichment in var promoters is antagonised by the var gene silencing factor PfSir2A . These findings further extend our knowledge of the complex mechanisms regulating gene expression in P . falciparum . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"infectious",
"diseases/protozoal",
"infections",
"genetics",
"and",
"genomics/epigenetics",
"cell",
"biology/gene",
"expression"
] | 2011 | Expression of P. falciparum var Genes Involves Exchange of the Histone Variant H2A.Z at the Promoter |
Middle East respiratory syndrome coronavirus ( MERS-CoV ) remains a threat to public health worldwide; however , effective vaccine or drug against CoVs remains unavailable . CoV helicase is one of the three evolutionary most conserved proteins in nidoviruses , thus making it an important target for drug development . We report here the first structure of full-length coronavirus helicase , MERS-CoV nsp13 . MERS-CoV helicase has multiple domains , including an N-terminal Cys/His rich domain ( CH ) with three zinc atoms , a beta-barrel domain and a C-terminal SF1 helicase core with two RecA-like subdomains . Our structural analyses show that while the domain organization of nsp13 is conserved throughout nidoviruses , the individual domains of nsp13 are closely related to the equivalent eukaryotic domains of Upf1 helicases . The most distinctive feature differentiating CoV helicases from eukaryotic Upf1 helicases is the interaction between CH domain and helicase core .
Severe acute respiratory syndrome coronavirus ( SARS-CoV ) and Middle East respiratory syndrome coronavirus ( MERS-CoV ) caused global pandemics in 2003[1] and 2012[2] with the fatality rates of 10–35% . Outbreak of MERS-CoV in the Republic of Korea[3] in 2015 highlighted that the newly emerged CoVs remain a major concern for the public health . Nevertheless , effective vaccine and drug against CoVs are still missing . MERS-CoV is a member of Coronaviridae family , one of the four distantly related virus families ( the other three are Arteriviridae , Mesoniviridae and Roniviridae ) in Nidovirales order[4–6] . This monophyletic group of viruses includes the largest known RNA genomes in families Roniviridae ( ~26 kb ) and Coronaviridae ( from 26 . 3 to 31 . 7 kb ) , as well as , small-sized Arteriviridae ( 12 . 7 to 15 . 7 kb ) and medium-sized Mesoniviridae ( 20 . 2 kb ) [7 , 8] . Coronaviridae family is divided into Torovirinae and Coronavirinae subfamilies , with the latter consisting of α-CoVs , β-CoVs , δ-CoVs and γ-CoV genera[5] . Six human CoVs have been identified to date , including α-CoVs 229E-CoV and NL63-CoV and β-CoVs OC43-CoV and HKU1-CoV from lineage A , SARS-CoV from lineage B and MERS-CoV from lineage C[9–11] . MERS-CoV has a positive single-stranded RNA ( +RNA ) genome of ~30kb , one of the largest among +RNA viruses[4] . To support the efficient replication of its exceptionally large genome , MERS-CoV encodes two replicase polyproteins pp1a and pp1ab , which are proteolytically processed into 16 nonstructural proteins ( nsps ) [12 , 13] . The nsps assemble into the membrane-associated replication-transcription complexes ( RTCs ) , which drive viral genome replication and translation . An RNA-dependent RNA polymerase ( nsp12 ) and a helicase ( nsp13 ) are central components of RTC[14 , 15] . It has been previously shown that +RNA viruses with genome larger than 7 kb encode helicases [16–18] . Helicases unwind DNA or RNA duplexes in an NTP hydrolysis dependent manner . They are classified into six superfamilies SF1-SF6 and participate in almost every aspect of nucleic acid metabolism[19] . Regardless of their functional diversity , helicases all contain core domains that hydrolyze NTPs . The enzymatic core is formed either by the tandem RecA-like domains within the same polypeptide chain ( SF1-SF2 superfamilies ) or between subunits of the functional oligomer of the helicase ( SF3-SF6 superfamilies ) [16] . The universal features of the RecA-like domain includes a Walker A motif forming the phosphate binding loop ( P-loop ) , a Walker B motif coordinating magnesium and an “arginine finger” engaging γ-phosphate of ATP[18 , 20 , 21] . In addition to the core domains , helicases also have accessory domains or inserts with various functions , such as assisting the catalytic activity or the interacting with other protein partner [16 , 17 , 22] . Sequence conservation analysis indicates that CoV nsp13 belongs to SF1 superfamily , including Rep , UvrD , PcrA , RecD , Pif1 , Dda , Upf1-like helicases and many +RNA virus helicases[18 , 23] . Nidovirus helicases share many structural features with the eukaryotic Upf1 helicase , a key factor in nonsense-mediated mRNA decay in cells[24 , 25] . Upf1 is a multi-domain protein comprising of an N-terminal Cys-His-rich domain ( CH domain ) coordinating three zinc atoms , a 1B domain with the β-barrel fold and a conserved SF1 helicase core with a 1C insert in the first RecA-like domain[24 , 26] . The crystal structure of nsp10 from equine arteritis virus ( EAV ) of Arterovirus genus is the first high-resolution structure of nidovirus helicase . EAV nsp10 has an N-terminal zinc-binding domain ( ZBD ) that is followed by the 1B and a SF1 helicase core , but it lacks the 1C insert . The ZBD of nsp10 coordinates two zinc ions by an N-terminal RING-like module and one zinc ion by a C-terminal treble-clef zinc finger . The 1B domain of nsp10 undergoes large conformational change upon substrate binding , and 1B together with the 1A and 2A domains of the helicase core form a channel that accommodates the single stranded nucleic acids . The CH domain of Upf1 mediates the binding with Upf2[26] . Similarly , the ZBD of nsp10 harbors a putative protein interaction surface , of which the binding partner remains to be identified[25] . The structural resemblance between Upf1 and EAV nsp10 suggests that nidovirus helicase may be involved in the posttranscriptional quality control of the viral RNAs . CoV helicase is one of the three evolutionary most conserved proteins in nidoviruses[27] , thus making it an important target for drug development[28] . Previous biochemical characterizations have shown that CoV nsp13 exhibits multiple enzymatic activities , which include hydrolysis of NTPs and dNTPs , unwinding of DNA and RNA duplexes with 5’-3’ directionality and the RNA 5’-triphosphatase activity[29 , 30] . Additionally , the RNA dependent RNA polymerase ( RdRP , nsp12 ) of CoV physically interacts with nsp13 and enhances its unwinding activity[31] . Although the molecular mechanism underlying these activities and the role of nsp13 in viral RNA synthesis are poorly understood , mutagenesis studies have identified a collection of residues important for the activity of nidovirus helicase . Disruption of the zinc binding function of 229E-CoV nsp13 or EAV nsp10 by replacing the conserved Cys/His residues at ZBD or deleting the entire zinc binding domain interfere with the ATPase activity of the helicases . Moreover , the activity of nsp10 is not complemented by providing wild-type ZBD in trans[32] . These results suggest that ZBD of nidovirus helicase modulates the ATPase/helicase activity in cis . CoVs nsp13 is essential for virus replication . ATPase/helicase deficient mutations of nsp13 ( either at the zinc-binding site or the Walker A motif ) can lead to the abolition of CoV replication . The mouse hepatitis virus ( MHV ) M protein and nsp13 are required for efficient replication An A335V mutation in the helicase core of nsp13 causes the attenuation of MHV replication both in vitro and in vivo [33] . SARS- and MERS-CoV outbreaks boosted nearly fifteen years of structural studies on the CoV proteins . However , despite extensive efforts , three-dimensional structural characterization of nsp13 , one of the most important CoV replication enzymes , remained absent .
To investigate the structure of nsp13 , we overexpressed the full-length MERS-CoV nsp13 ( 1-598aa ) in High-5 insect cells ( Fig 1A ) . To verify that the recombinant protein is enzymatically active we first performed ATPase assay . The purified nsp13 exhibited ATPase activity with a turnover number ( kcat ) of 2 . 03 ±0 . 1 s-1 and the catalytic efficiency ( kcat Km-1 ) of 0 . 32 μM-1 s-1 ( Fig 1B ) . The ATPase activity of MERS-CoV nsp13 is comparable with that reported for SARS nsp13[29] . Next , we assessed helicase activity of the recombinant nsp13 . Partial RNA duplex containing 5’ overhang was fully unwound by the purified nsp13 . By contrast , MERS-CoV nsp13 could not unwind RNA duplex containing 3’ overhang ( the duplex region remained the same as the RNA duplex with 5’ overhang ) . These results confirm that MERS-CoV nsp13 is a unidirectional helicase with the unwinding polarity of 5’-to-3’ ( Fig 1C , left ) . Mutant with E375Q within Walker B failed to unwind the RNA substrate with 5’ overhang clearly indicating that the helicase activity of MERS-CoV nsp13 is dependent on ATP hydrolysis ( Fig 1C , left ) . MERS-CoV nsp13 was able to hydrolyze different NTPs and dNTPs to support the unwinding of RNA substrate , with a clear preference towards ATP ( Fig 1C , right ) . Our results are consistent with the recent enzymatic characterization of the MERS-CoV nsp13 expressed in bacteria[30] . Our crystallization trials with the unliganded MERS-CoV nsp13 yielded crystal , which diffracted the X-rays poorly . Intriguingly , incubation with a 5’-triphosphate-15T DNA ( ppp-15T ) greatly improved diffracting power of nsp13 crystals . Benefiting from the presence of an N-terminal zinc-binding domain , we collected highly redundant multi-wavelength anomalous diffraction ( MAD ) data at the zinc absorption edge and subsequently solved the structure . Next , we collected a 3 . 0Å resolution native dataset , which we used for final structure refinement and further analysis . There are two nsp13 in the asymmetric unit ( ASU ) with three zinc ions bound to each N-terminal Cys/His rich domain . Unexpectedly , no additional electron density for ppp-15T DNA could be identified , indicating that a stable nsp13-DNA complex did not form . The final model of nsp13 has good stereochemistry quality . Data collection and refinement statistics are summarized in Table 1 . MERS-CoV nsp13 is composed of multiple functional domains ( Fig 2 , S1 and S2 Figs ) . The N-terminal CH domain has 15 conserved Cys/His residues , twelve of which participate in the coordination of three zinc ions ( Fig 3A & Fig 4A–4C ) . The C-terminal helicase belongs to the SF1 helicase family and consists of two “RecA-like” domains , referred to as RecA1 and RecA2 . The CH and helicase are connected via a region consisting two additional domains . A helical domain sandwiched between the CH and RecA1 domains is followed by a six-stranded anti-parallel β-barrel domain . Because of the structural resemblance to the domains of Upf1 helicases , these domains were named as Stalk and 1B , respectively . We compared the crystal structure of MERS-CoV nsp13 with all structures in Protein Data Bank using Dali server . The top hits were human Upf1 helicase[24] ( hUpf1 PDB code: 2GK7; Dali Z-score: 22 . 7 ) and arterivirus ( EAV ) helicase[25] ( PDB code: 4N0N; Dali Z-score: 22 . 2 ) . The structural superposition showed that while the helicase portion of these three proteins aligns well , the N-terminal CH ( or ZBD ) and 1B domains adopt various conformations relative to the helicase core . The CH domain of MERS-CoV nsp13 ( residues 1–112 ) is a compact domain with three zinc-binding motifs stabilizing the fold ( Fig 3A ) . The CH domain contains an N-terminal RING-like module ( 1-46aa , β1–5 and α1 ) with two zinc fingers and C-terminal RING-like module with single treble-clef zinc finger ( 47-87aa , β6–8 ) . A long loop ( 88-100aa , between β8-α2 ) spanning across the entire height of CH domain connects zinc coordination modules . The first zinc ( Zn1 ) is coordinated within a CCCC type treble-clef zinc finger with four cysteine ligands ( Cys5 , Cys8 Cys26 and Cys29 ) ( Fig 3B ) . Cys5 and Cys8 are located on a zinc knuckle between β1-β2 strands , whereas Cys26 and Cys29 are placed in the N-terminal region of α1 helix . The second zinc ( Zn2 ) is coordinated by a two-Cys , two-His ( C2H2 ) zinc finger motif . The cysteine ligands ( Cys16 and Cys19 ) are forming part of a zinc knuckle between β3-β4 strands , while histidine ligands His33 and His39 are located on α1 helix and the loop between α1 and β5 , respectively . The third zinc ( Zn3 ) is coordinated by the C-terminal treble-clef zinc finger motif ( Fig 3C ) . A zinc-knuckle on a hairpin loop between β5 and β6 strands provides Cys50 and Cys55 for Zn3 coordination , whereas Cys72 from the C-terminal region of β7 strand and His75 from the loop between β7 and β8 strands provide two more ligands for the zinc . All Cys/His involved in zinc coordination are invariant in CoVs ( Fig 4C ) , indicating the zinc binding is essential for the structure and function of nsp13 . We compared the isolated CH/ZBD of MERS-CoV nsp13 , EAV nsp10 and Upf1 ( Table 2 ) , and found that the CH of MERS-CoV nsp13 is structurally more related to the CH of Upf1 than to the ZBD of EAV nsp10 . Superimposition of yeast Upf1 CH ( scUpf1 PDB code: 2XZL ) and MERS-CoV nsp13 CH aligned 102 Cα atoms with the Dali Z-score of 10 . 2 and rmsd of 2 . 2Å . Superimposition of EAV nsp10 CH ( PDB code: 4N0N ) and MERS-CoV nsp13 CH aligned 59 Cα atoms with the Dali Z-score of 4 . 4 and rmsd of 3 . 4Å . Additionally , while the positions of three zincs of MERS-CoV nsp13 CH domain and Upf1 CH domain are nearly identical , Zn3 of EAV nsp10 ZBD is located much closer to Zn1 and Zn2 . Particularly , the distance from Zn3 to Zn2 is ~6 . 4Å shorter in EAV nsp10 than the corresponding distance in MERS nsp13 ( Fig 4B ) . However , the interaction between the CH domain and the helicase region is significantly different between nidovirus helicases and Upf1 , and similar between MERS-CoV nsp13 and EAV nsp10 . The MERS-CoV nsp13 CH domain is tightly attached to the Stalk domain ( Fig 5A ) . Helices α2-α4 form a three-helical bundle . The buried area between CH and Stalk is 1670Å2 . Similarly , the EAV nsp10 ZBD domain is connected to the stalk domain , which is significantly smaller than CH of MERS-CoV nsp13 ( Fig 5B ) . The buried area between ZBD and Stalk is 634 . 7 Å2 . By contrast , the CH of Upf1 helicase is linked to the helicase portion through a long and flexible loop ( often invisible in the electron density maps ) , allowing large movement of the CH domain ( Fig 5C and 5D ) [26] . Structural comparison of nsp13 CH and hUpf1 CH in complex with Upf2[34] revealed two hydrophobic pockets on the surface of nsp13 CH equivalent to Upf2 binding sites on Upf1 ( Fig 3D ) . While pocket 1 highly resembles Upf2 α-helix binding site , the pocket 2 has a much shorter β6-β7 loop than the equivalent loop in Upf2 β-hairpin binding site of Upf1 ( β5-β6 loop ) . Two hydrophobic pockets on CH domain may function as interaction interfaces for other CoV replicase or cellular protein . The nucleotide-binding pocket of MERS-CoV nsp13 is located between RecA1 and RecA2 domains . The RecA1 ( 241-443aa ) contains a seven-stranded parallel β-sheet sandwiched by two α-helices located near the Stalk domain on one side and three α-helices on the opposite side . RecA2 ( 444-596aa ) has a five-stranded parallel β-sheet surrounded by four helices on one side and three helices on the other side . Seven helicase motifs conserved in SF1/SF2 families are located in the cleft between RecA1-RecA2 . RecA1 contains motifs I , Ia , II and III , whereas RecA2 includes motifs IV , V and VI ( Fig 3E ) . Sulfate , crystallization condition precipitant , was found bound to the P-loop mimicking binding of the NTP’s phosphate moiety . Residues Gln404 , Arg443 and Arg567 from helicase motifs III , IV and VI form hydrogen bonds with the sulfate suggesting their involvement in NTP hydrolysis . The corresponding residues in human Upf1 helicase are Gln665 , Arg703 and Arg865[24] , while in EAV nsp10 Gln267 , Arg296 and Arg381[25] . Residues Arg865 and Gln665 of Upf1 helicase act as the “arginine finger” and “γ-phosphate sensor” during ATP hydrolysis[24] , suggesting that MERS-CoV nsp13 Gln404 and Arg567 have the same function . Moreover , Tyr442 of MERS-CoV nsp13 is structurally equivalent to Tyr702 of Upf1 , which stabilizes adenosine base ( Fig 3E ) . It has been previously shown that CoV nsp13 interacts with both RNA and DNA in a sequence-independent manner[29 , 35 , 36] . To analyze the nucleic acid binding pocket of MERS-CoV nsp13 , we generated a model of nsp13-ssRNA based on the superposition of the helicase domain of nsp13 with the helicase domains of hUpf1 , scUpf1[26] and EAV nsp10[25] . The model shows that while the 3’ end of single-stranded RNA is located in the channel formed by 1B , Stalk and RecA1 domains; the 5’ end of the RNA lies on top of RecA2 domain ( Fig 6A ) . Although the helical insertion equivalent to the Upf1 1C domain is missing in nsp13 , the protein has a topologically equivalent loop between β17 and β18 which fulfills similar function in RNA binding . The β17-β18-loop makes direct contacts with 1B domain , forming the 3’ end outlet of the putative RNA binding channel . The narrowest opening of the putative RNA binding channel has the width and height of 6 and 12Å , which is just enough to accept a single-stranded RNA or DNA ( Fig 6B ) . The dimension of the RNA binding channel of nsp13 is similar to that in EAV nsp10 , but smaller than the channel in Upf1 . Based on structural comparisons , we predicted the residues of nsp13 , which are likely involved in RNA recognition ( S2 Fig & S1 Table ) . Their structural equivalents are mostly conserved in EAV nsp10 and Upf1 helicases , suggesting that CoV nsp13 adopts the similar mechanism for nucleic acids binding .
Structural comparisons of the individual domains of Upf1-like helicases using Dali server showed that the CH of MERS-CoV nsp13 is structurally closer to the CH of Upf1 than to the ZBD of EAV nsp10 . In accordance with previously published results , the structure of the helicase cores is well conserved in all three Upf1-like helicases ( Table 2 ) . However , some subtle differences could also be identified . Firstly , while the helicase cores of MERS-CoV nsp13 and Upf1 have similar size , the helicase core of EAV nsp10 is more compact . MERS-CoV nsp13 contains seven and five parallel β-strands in RecA1 and RecA2 domains , respectively , Upf1 has seven and six β-strands in the RecA1 and RecA2 domain , respectively , whereas EAV nsp10 has only five and four β-strands in the RecA like domains . Secondly , while three Upf1 like helicases all contain the 1B domain with the β-barrel fold , only Upf1 has a helical 1C insertion in RecA1 domain . The equivalent 1C insertion is missing in both EAV nsp10[25] and MERS-CoV nsp13 ( Fig 4D and 4E ) . Thirdly , CoVs nsp13 does not contain a C-terminal domain homologous to the C-terminal portion of EAV nsp10 ( C-terminal 65 residues ) , which was shown to regulate ATPase and helicase activities[25] . The sequence conservation analysis showed that the C-terminal domain of arteriviruses and CoVs is indeed poorly conserved[25] . Our crystallographic study provides the complete structure of the extreme C-terminal region of MERS-CoV nsp13 , which shows that the C-terminus is an integral part of the RecA2 domain . Thus , we conclude that the C-terminal regulatory domain outside SF1 helicase core is completely missing in MERS-CoV nsp13 and other CoVs helicases . Previous mutagenesis studies of CoVs nsp13 have identified a collection of residues essential to the activity of nsp13 . Because nsp13 helicase is highly conserved among CoVs , the crystal structure of MERS-CoV nsp13 can provide three-dimensional information to understand previous phenotypes of nsp13 mutants . It has been shown that mutations of the conserved Cys/His at ZBD of 229E-CoV nsp13 interfered with its ATPase activity[32] . Ala or Arg replacement of C5003 , C5021 , C5024 and H5028 abolished or reduced the ATPase activity of 229E-CoV nsp13 . Our crystal structure of MERS-CoV nsp13 confirms that these residues , corresponding to C8 , C26 , C29 and H33 of MERS-CoV nsp13 ( Fig 4C ) are coordinating Zn1 and Zn2 . The loss of ATPase activity caused by Cys/His substitutes can be attributed to the disruption of zinc-binding and the integrity of the CH domain . Unexpectedly , only C5050A mutant retained significant ATPase activity ( ~60% ) . Based on the comparison with corresponding C55 of MERS-CoV nsp13 , C5050 of 229E-CoV coordinates Zn3 of the ZBD . Zn3 is the most distantly located zinc from the helicase core ( 6 . 4Å away from Zn2; Fig 1A ) , and the ZBD/CH is tightly attached to the helicase core in CoV nsp13 , it is therefore unsurprising that the impaired binding of Zn3 has the minimal effect on the enzymatic activity of nsp13[32] . This hypothesis is supported by the mutagenesis study of EAV nsp10[25 , 32] , which also showed that while mutant H2414A ( ligand for Zn3 ) retained residual ATPase/ helicase activity and wild-type-level nucleic acid binding activity , the activities of mutants C2395A ( ligand for Zn1 ) and H2399A ( ligand for Zn2 ) was completely abolished . Recently , Zhang et al found that mutation A335V of MHV nsp13 led to attenuation of virus growth in cells and resulted in ~30 fold reduced viral titer in the livers of infected mice[33] . Residue A335 of MHV nsp13 is highly conserved among CoVs nsp13 ( S2 Fig ) . Based on the crystal structure of MERS-CoV nsp13 , A335 ( corresponding to A336 of MERS-CoV nsp13 ) is located on the β17-β18-loop , a loop interacts with the 1B domain as well as participates in the formation of RNA binding channel . Because A335 is highly exposed to the surface of nsp13 , Valine replacement could increase hydrophobicity of this region . Therefore , A335V mutation may impair the function of nsp13 by destabilizing the local structure critical to RNA binding or promoting aggregation of the protein . In summary , our current study presents the first structural insight into the multiple functionality of the CoV nsp13 . Our analyses demonstrate that while the domain organization of MERS-CoV nsp13 and EAV nsp10 is conserved , the structures of the individual domains of nsp13 are closely related to their eukaryotic equivalents in Upf1 helicase . While the interaction between the CH ( or ZBD ) domain and the helicase core presents the most distinctive feature differentiating nidovirus helicases from the Upf1 helicases , the high resemblance between the CH domains of MERS-CoV nsp13 and Upf1 helicases is remarkable . This structural similarity not only supports a hypothesis that nidoviruses helicase may have a Upf1-like role in post-transcriptional quality control of viral RNAs synthesis[25] , but also implies a possibility that CoV nsp13 might use its Upf1-similar CH domain to interfere with nonsense-mediated mRNA decay ( NMD ) pathway . It has already been shown that NMD targets viral RNAs for degradation in the early phase of infection of +RNA viruses[37] . Based on the coevolution of virus and host mechanism , the viruses also develop strategies to suppress NMD . Nevertheless , whether CoV nsp13 is involved in NMD suppression requires experimental evidence . To investigate what proteins bind nsp13 through its CH domain and how do they modulate nsp13 function is important to expand our knowledge about the evolution and function of Upf1-like helicases . Finally , our results provide novel structural information essential for structure-based drug design against CoV .
The gene encoding full-length MERS-CoV nsp13 helicase ( 1-598aa , GeneBank accession: YP_009047202 ) was amplified by polymerase chain reaction ( PCR ) and inserted via BamHI/XhoI restriction sites into a modified pFastbac-1 baculovirus transfer vector with an N-terminal 6×His-SUMO tag[38] . Nsp13 protein was overexpressed in High-5 insect cells using Bac-to-Bac Baculovirus Expression System ( Invitrogen ) . One liter cell culture ( 2 . 0×106 cells ml-1 ) was infected with 30 ml baculovirus at 22°C . Forty-eight hours after infection , cells were harvested by centrifugation . The cell pellet was re-suspended in a lysis buffer containing 25mM Tris-HCl ( pH 7 . 5 ) , 1 . 5 M NaCl and 15mM imidazole and lysed by ultrasonification . High salt concentration in the lysis buffer was necessary to remove nucleic acids bound to nsp13 . Protein was purified using Ni-NTA resin ( QIAGEN ) . The eluted protein was digested with PreScission protease ( GE healthcare ) to remove the 6×His-SUMO tag . Finally , the untagged nsp13 was purified using size-exclusion chromatography ( Superdex-200 , GE healthcare ) . The purified nsp13 was concentrated to ~8 mg/ml in the buffer containing 10mM HEPES ( pH 7 . 0 ) and 100mM NaCl . Before crystallization trials , nsp13 was mixed with 5’-triphosphate 15 thymine single-stranded DNA ( ppp-15T ) with 1:1 . 5 molar ratio and incubated at 4°C overnight . Crystallization of nsp13 was achieved by mixing equal volume of sample and reservoir buffer ( 1 . 0μl ) containing 0 . 1 M Tris-HCl ( pH 8 . 5 ) , 1M ( NH4 ) 2SO4 and 15% glycerol . The crystals were grown in a hanging-drop vapor-diffusion system at 18°C . 5’-triphosphate DNA ( ppp-15T ) was synthesized and purified according to previously published procedures[39 , 40] . Briefly , ppp-15T oligonucleotide was analyzed by LC/ESI-MS on a XBridge C8 column ( 2 . 1 x 50 mm , 2 . 5μm ) . Buffer A was 95mM 1 , 1 , 1 , 3 , 3 , 3-hexafluoro-2-propanol , 16mM triethylamine in water and buffer B was 100% methanol . A gradient from 2% to 29% B over 26 . 8 min with flow rate of 0 . 25 mL/min was employed at a column temperature of 60°C . A DNAPac-200 column ( 4 x 250 mm ) was used for analytical IE-HPLC . Buffer A was 25mM Tris-HCl , 1mM EDTA in 10% acetronitrile ( pH = 8 ) and buffer B was buffer A plus 1 M sodium bromide . A gradient of 25 to 56% B over 21 . 5 min at a flow rate of 1 . 0 mL/min was used at a column temperature of 75°C . Final purity by IE-HPLC and LC/ESI-MS was above 94% ( S3 Fig ) . Prior to crystallization trials nsp13 protein was incubated with a 5’-triphosphate single-stranded DNA containing 15T ( ppp-15T ) . Highly redundant multi-wavelength anomalous diffraction data were collected using the X-ray with wavelengths close to the absorption edge of zinc ( Hrem: 1 . 2810Å , Peak: 1 . 2827Å and Infl: 1 . 2831Å ) . Crystal belonged to the space group P6122 , and contained two molecules per ASU . An interpretable electron density map was calculated using SHARP/autoSHARP[41] . An initial model of MERS-CoV nsp13 was manually built using Coot[42] . Finally , a native data with highest resolution ( 3 . 0Å ) was collected using the X-rays with the wavelength of 0 . 978Å . Higher resolution structure was solved by molecular replacement using the initial nsp13 structure as the searching model . The 3 . 0Å structure was refined using PHENIX[43] . In the final model , 145-230aa ( the entire 1B domain ) of molecule A are disordered , probably due to mobility of 1B and the lack of crystal contacts , whereas in molecule B , 591 out of 598 amino acids were located in the electron density maps ( S4 Fig ) . ATPase assay was carried out as previously described[44] . Briefly , reaction mixtures ( 50μl ) containing 100mM Tris-HCl ( pH 8 . 0 ) , 4mM MgCl2 , trace amount of [γ-32P]ATP ( ~1nM ) and the specified amount of ATP ( from 10μM to 320μM ) were incubated at 30°C . The reaction was initiated by the addition of MERS-CoV nsp13 ( 20nM ) . At each indicated time point , 2μl of quenching buffer ( 0 . 5M EDTA ) was added to the mixture to stop the reaction . Finally , 1μl the sample was spotted on the thin-layer chromatography cellulose TLC plates ( Sigma-Aldrich ) and resolved with running buffer containing 0 . 8M acetate and 0 . 8M LiCl . The plates were dried and analyzed using storage phosphor screen and Typhoon Trio Variable Mode Imager ( GE healthcare ) . ATP turnover was quantified using ImageQuant TL software ( GE Healthcare ) . The reaction mixture ( 10μl ) contained 50mM HEPES ( pH 7 . 5 ) , 5mM MgCl2 , 2mM dithiothreitol ( DTT ) , 1mM nucleotide ( ATP , GTP , CTP or TTP ) , 50nM of partial duplex RNA substrate and 300nM unlabeled trap RNA ( 5’-CGAAGCUGCUAACAUCAG-3’ ) [36] . The RNA substrate with 5’ overhang is prepared by mixing a top strand: 5’-UUUUUUUUUUCUGAUGUUAGCAGCUUCG-3’ with a bottom strand: 5’-HEX-CGAAGCUGCUAACAUCAG-3’ . The RNA substrate with 3’ overhang is prepared by mixing a top strand: 5’-CUGAUGUUAGCAGCUUCGUUUUUUUUUUUUUUUUUUUU-3’ with the same bottom strand as in RNA partial duplex with 5’ overhang . The reaction was initiated by the addition of MERS-CoV nsp13 ( 100nM ) and incubated at 30°C for 30 minutes . The reaction was terminated by addition of 2 . 5μl loading buffer ( 5X ) containing 100 mM Tris-HCl ( pH7 . 5 ) , 1% SDS , 50mM EDTA and 50% glycerol . Samples were resolved by 10% native-PAGE running on ice . The gel was scanned with Typhoon Trio Variable Mode Imager ( GE healthcare ) . Multiple sequence alignment was carried out using MUSCLE software ( http://www . ebi . ac . uk/Tools/msa/muscle/ ) . In some region , minor manual adjustments were performed in accordance to the structural superimposition of the Upf1 like helicases . The illustration of sequence alignment was either generated using ESPript 3 . 0 ( http://espript . ibcp . fr/ESPript/ESPript/ ) , or produced manually using Microsoft PowerPoint . The structural alignment was carried out using Dali server ( http://ekhidna . biocenter . helsinki . fi/dali_server ) . | Severe acute respiratory syndrome coronavirus ( SARS-CoV ) and Middle East respiratory syndrome coronavirus ( MERS-CoV ) caused global pandemics in 2003 and 2012 with the fatality rates of 10–35% . Outbreak of MERS-CoV in the Republic of Korea in 2015 highlighted that the newly emerged CoVs remain a concern for the public health . Nevertheless , effective vaccine and drug against CoVs are still missing . Among CoV-encoded nonstructural proteins ( nsps ) , nsp13 helicase is considered one of the most important drug targets . Nsp13 is a highly conserved protein in CoVs and nidovirales and one of the two central components of the membrane associated replication-transcription complex , which performs viral RNA synthesis . However , despite decades of structural characterization of CoV-encoded proteins , the structure of nsp13 remained unavailable . In this study , we determined the first crystal structure of the full-length MERS-CoV nsp13 . MERS-CoV helicase has an N-terminal Cys/His rich domain ( CH ) with three zincs , a beta-barrel domain and a C-terminal SF1 helicase core . While the domain organization of nsp13 is similar to arterivirus nsp10 , the individual domains of nsp13 are closely related to their equivalent domains of eukaryotic Upf1 helicases . Our findings provide novel structural information essential for structure-based drug design against CoV . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"coronaviruses",
"crystal",
"structure",
"pathology",
"and",
"laboratory",
"medicine",
"enzymes",
"pathogens",
"split-decomposition",
"method",
"condensed",
"matter",
"physics",
"microbiology",
"enzymology",
"phosphatases",
"viruses"... | 2017 | Crystal structure of Middle East respiratory syndrome coronavirus helicase |
Learning in biologically relevant neural-network models usually relies on Hebb learning rules . The typical implementations of these rules change the synaptic strength on the basis of the co-occurrence of the neural events taking place at a certain time in the pre- and post-synaptic neurons . Differential Hebbian learning ( DHL ) rules , instead , are able to update the synapse by taking into account the temporal relation , captured with derivatives , between the neural events happening in the recent past . The few DHL rules proposed so far can update the synaptic weights only in few ways: this is a limitation for the study of dynamical neurons and neural-network models . Moreover , empirical evidence on brain spike-timing-dependent plasticity ( STDP ) shows that different neurons express a surprisingly rich repertoire of different learning processes going far beyond existing DHL rules . This opens up a second problem of how capturing such processes with DHL rules . Here we propose a general DHL ( G-DHL ) rule generating the existing rules and many others . The rule has a high expressiveness as it combines in different ways the pre- and post-synaptic neuron signals and derivatives . The rule flexibility is shown by applying it to various signals of artificial neurons and by fitting several different STDP experimental data sets . To these purposes , we propose techniques to pre-process the neural signals and capture the temporal relations between the neural events of interest . We also propose a procedure to automatically identify the rule components and parameters that best fit different STDP data sets , and show how the identified components might be used to heuristically guide the search of the biophysical mechanisms underlying STDP . Overall , the results show that the G-DHL rule represents a useful means to study time-sensitive learning processes in both artificial neural networks and brain .
Most learning rules used in bio-inspired or bio-constrained neural-network models of brain derive from Hebb’s idea [1 , 2] for which “cells that fire together , wire together” [3] . The core of the mathematical implementations of this idea is multiplication . This captures the correlation between the pre- and post-synaptic neuron activation independently of the timing of their firing . Time is however very important for brain processing and its learning processes [4] . Differential Hebbian learning ( DHL ) rules [5 , 6] are learning rules that change the synapse in different ways depending on the specific timing of the events involving the pre- and post-synaptic neurons . For example , the synapse might tend to increase if the pre-synaptic neuron activates before the post-synaptic neuron , and decrease if it activates after it . As suggested by their name , DHL rules use derivatives to detect the temporal relations between neural events . Here we will use the term event to refer to a relatively short portion of a signal that first monotonically increases and then monotonically decreases . Events might for example involve the activation of a firing-rate unit in an artificial neural network , or the membrane potential of a real neuron , or a neurotransmitter concentration change . DHL rules use the positive part of the first derivative of signals to detect the initial part of events , and its negative part to detect their final part . By suitably multiplying the positive/negative parts of the derivative of events related to different signals , DHL rules can modify the synapse in different ways depending on how their initial/final parts overlap in time . To the best of our knowledge , current DHL rules are basically two: one proposed by Kosko [5] and one proposed by Porr , Wörgötter and colleagues [6 , 7] . These rules modify the synapse in specific ways based on the temporal relation between the pre- and post-synaptic events . Formulating other ways to modify synapses based on event timing is the first open problem that we face here . The development of dynamical neural-network models and learning mechanisms that , as DHL , are able to take time into consideration is very important . Indeed , the brain is an exquisitely dynamical machine processing the continuous flow of information from sensors and issuing a continuous flow of commands to actuators so its understanding needs such types of models [8–11] . In this respect , neuroscientific research on spike timing dependent plasticity ( STDP; [12] ) clearly shows how synaptic changes strongly depend on the temporal relation between the spikes of the pre- and post-synaptic neurons . Given the typical shape of spikes , an important class of STDP models , called phenomenological models [13] , abstracts over the features of the spike signals and directly links the synaptic strengthening , Δw , to the time interval separating the pre-synaptic and post-synaptic spikes , Δt , on the basis of a function of the type Δw = f ( Δt ) [12 , 14] . Such a function is usually designed by hand and reflects the synaptic changes observed in experimental data . [15] . The function f ( Δt ) generates a typical learning kernel that when plotted shows a curve where each Δt causes a certain Δw . Phenomenological models are simple but are applicable only to spike events . In comparison , DHL rules are more complex but have the advantage of computing the synaptic update as the step-by-step interaction ( based on multiplication ) between the pre-synaptic and post-synaptic events . Therefore they are applicable to any complex signal that might exhibit events with variable time courses . When applied to the study of STDP , the property of DHL rules just mentioned also opens up the interesting possibility of using them to investigate the actual biophysical neural events following and caused by the spikes that actually lead to the synaptic change , as first done in [16] . The chain of processes changing the synapse is also captured by biophysical models ( e . g . , see [14 , 17] ) . These models can capture those processes in much biological detail ( mimicking specific neurons , neuromodulators , receptors , etc . ) but at the cost of being tied to specific phenomena . Because the level of abstraction of DHL rules lies between that of phenomenological models and that of biophysical models , DHL represents an important additional research tool . Experimental study of STDP [18 , 19] shows that different types of neurons , for example excitatory/inhibitory neurons in different parts of the brain , implement a surprisingly rich repertoire of learning kernels . It is reasonable to assume that the brain employs such learning mechanisms to implement different computational functions . In this respect , an interesting fourth class of models appropriate for studying STDP , which might be called functional models , aims to derive , or to justify , specific STDP learning kernels based on normative computational principles [20–23] . Investigating the functions of different STDP kernels is not in the scope of this work . However , assuming that the variety of learning kernels discovered through STDP experiments supports different functions relevant to neural processing and that analogous functions might be needed in artificial neural networks , it is important to understand the computational mechanisms that might generate such a variety of learning kernels . In this respect , an important question is this: is there a DHL learning rule , or a set of them , that can generate the complete variety of learning kernels found in the brain ? Some existing research shows how different STDP learning kernels can arise from the same biophysical mechanisms [17] , or from the same DHL-based model [24] . However , these studies propose specific mechanisms to address a sub-set of STDP data sets rather than proposing a general way to systematically reproduce STDP learning kernels . Understanding the extent to which DHL can capture the known STDP phenomena , and how this can be done , is thus a second important open problem that we address here . The rest of the paper addresses the two open problems indicated above in the following ways . As a first contribution of the paper , the Section ‘G-DHL and the systematisation of DHL’ considers the first open problem—how different DHL rules can be generated in a systematic fashion—by proposing a general framework to produce DHL rules . In particular , the section first reviews the DHL rules proposed so far in the literature; then it presents the G-DHL rule and shows how it is able to generate the existing DHL rules and many others; and finally it shows how one can filter the neural signals to generate events that correspond to the features of interest and can use memory traces to apply the G-DHL rule to events separated by time gaps . As a second contribution of the paper , the Section ‘Using G-DHL to fit STDP data sets’ deals with the second open problem—understanding if and how G-DHL can be used to capture known STDP phenomena . To this end , the section first illustrates how the G-DHL synapse update caused by a pre- and post-synaptic spike pair can be computing analytically rather than numerically , and then it presents a collection of computational tools to automatically search the rule components and parameters to fit a given STDP data set . Addressing the same second open problem , and as a third contribution of the paper , the Section ‘Using G-DHL to fit STDP data sets’ uses those computational tools to show how the G-DHL rule is able to reproduce several learning kernels from the STDP literature . To this end , the section first uses G-DHL to fit the classic STDP data set of Bi and Poo [25]; then it illustrates how the G-DHL components found by the fitting procedure can be heuristically useful to search for the biophysical mechanisms underlying a given STDP data set; and finally it shows how to apply the G-DHL rule to systematically capture different aspects of all the STDP data sets reviewed by Caporale and Dan [18] ( such as their temporal span , long-term potentiation/depression , and variability around zero inter-spike intervals—e . g . sharp depression-potentiation passages , non-learning plateaus , Hebbian/anti-Hebbian learning ) . The Section ‘Discussion’ closes the paper by analysing the main features of G-DHL and its possible development . All software used for this research is available for download from internet ( https://github . com/GOAL-Robots/CNR_140618_GDHL ) .
As discussed in the introduction , different types of neurons exhibit surprisingly different STDP learning kernels . For this reason we tested the flexibility of G-DHL by using it to capture several different STDP learning kernels involving pairs of pre- and post-synaptic spikes . In the future G-DHL could be extended to capture STDP processes involving spike triplets or quadruplets ( [41]; see [42] for a model ) by considering three or more multiplication elements rather than only two as done here . To apply G-DHL to spike pairs , we first outline the procedure used to derive the formulas to compute G-DHL analytically , rather than numerically as done so far . The procedure is illustrated in detail in Section 2 . 1 in S1 Supporting Information in the case in which one assumes that spikes and traces are described with some commonly used formulas . Sections 2 . 7 and 2 . 8 in S1 Supporting Information show a method that leverages these formulas to use G-DHL to fit STDP data sets; examples of this fitting are shown in the Section ‘Results’ . Before presenting the formulas , we discuss two important points . The closed-form formulas for synaptic updates by the G-DHL rule have two main advantages . First , they allow the mathematical study of the G-DHL rule ( see Sections 2 . 2 and 2 . 6 in S1 Supporting Information ) . Second , the formulas allow a computationally fast application of G-DHL by computing the synaptic update through a single formula rather than as a sum of many step-by-step synaptic updates as done in its numerical application , an advantage exploited in the computationally intensive simulations of the Section ‘Results’ . A second observation concerns the relation between the G-DHL explicit formulas and phenomenological models discussed in the introduction . The G-DHL explicit formulas have the form Δw = f ( Δt ) typical of phenomenological models . This shortcut is possible because spikes have a fixed shape: this implies that Δt is the only information relevant for computing G-DHL . The resulting synaptic update is however the same as the one that would be obtained by numerically simulating the step-by-step interaction between the pre- and post-synaptic neural events mimicking more closely what happens in the real brain . Therefore , the possibility of computing Δw = f ( Δt ) formulas for DHL rules does not violate what we said in the introduction , namely that G-DHL captures the mechanisms causing the synaptic update at a deeper level with respect to phenomenological models .
The procedure for the automatic fit of STDP data sets was first employed to fit the classic STDP data set of Bi and Poo from rat hippocampal neurons [25] . Fig 8a summarises the results ( for ease of reference , henceforth we will refer to synapse strengthening/weakening as ‘LTP—long term potentiation’ and ‘LTD—long term depression’ ) . The model comparison technique selected two G-DHL components: an LTP component ( σpp = 0 . 73 ) and an LTD component ( ηps = −0 . 025 ) . The parameters σ and η differ in scale as they refer to differential and mixed G-DHL components involving signal-derivative or derivative-derivative multiplications . Fig 8b shows the target data and their fit obtained with the G-DHL components and parameters shown in Fig 8a . The G-DHL regression fits the data accurately ( FVU = 0 . 2725 ) . While the original paper performed the fit with the usual exponential function for both positive and negative Δt , the G-DHL regression captures the LTP with the σpp ‘sharp’ component ( Fig 6 ) , concentrated on small positive inter-spike intervals , and the LTD with the ηps = −0 . 025 ‘softer’ component ( Fig 7 ) , concentrated on negative intervals . We now illustrate with an example the idea of using the components found by the G-DHL regression to heuristically search for biophysical mechanisms possibly underlying a target STDP data set . This example involves the Bi and Poo’s data set [25] analysed in the previous section . The idea relies on the observation that each multiplication factor of the G-DHL components identified by the regression procedure has a temporal profile that might correspond to the temporal profile of the pre-/post-synaptic neuron electrochemical processes causing the synaptic change . The steps of the procedure used to search the biophysical mechanisms are as follows: ( a ) identify with an automatic procedure the G-DHL components and parameters fitting the target STDP data set; ( b ) define the temporal profile of the two pre-/post-synaptic factors of each found component , and the LTP/LTD effects caused by the component; ( c ) identify possible biophysical processes having a temporal profile similar to the one of the identified factors; ( d ) design experiments to verify if the hypothesised biophysical processes actually underlie the target STDP phenomenon in the brain . We now give an example of how to apply the steps ‘a’ and ‘b’ , and some initial indications on the step ‘c’ , in relation to the Bi and Poo’s data set [25] . The example aims to only furnish an illustration of the procedure , not to propose an in-depth analysis of this STDP data set . Regarding step ‘a’ , Fig 8 shows that the G-DHL regression identified two LTP and LTD components . Regarding step ‘b’ , Fig 9 shows the temporal profile of the factors of the two components . The first component is a ‘positive-derivative/positive-derivative’ component ( [ u ˙ 1 ] + [ u ˙ 2 ] +; Fig 9a , left graph ) with two factors ( Fig 9b , left graph ) : ( a ) a relatively long pre-synaptic factor ( [ u ˙ 1 ] + ) lasting about 30 ms; ( b ) a shorter post-synaptic factor ( [ u ˙ 2 ] + ) lasting about 7 ms . These two factors , amplified by a positive coefficient ( σpp = + 0 . 73 ) , produce LTP concentrated on small positive inter-spike intervals ( 0 ms < Δt < 30 ms; Fig 9a , left graph ) . The second component is a ‘positive-derivative/signal’ component ( [ u ˙ 1 ] + u 2; Fig 9a , right graph ) with other two factors ( Fig 9b , right graph ) : ( a ) a relatively long pre-synaptic factor ( [ u ˙ 1 ] + ) lasting about 30 ms; ( b ) a longer post-synaptic factor ( u2 ) lasting about 50 ms . The two factors , amplified by a negative coefficient ( ηps = −0 . 025 ) , produce LTD covering negative-positive inter-spike intervals ( −30ms < Δt < 20ms; see Fig 9a , right graph ) . When the two components are summed , LTP more than cancels out LTD for positive delays ( 0ms < Δt < 20ms ) . This causes the sharp passage from LTD to LTP around the critical Δt values close to zero , which characterise the target kernel ( Fig 8 ) . Regarding step ‘c’ of the procedure , directed to identify possible biological correspondents of the component factors identified in step ‘b’ , we now discuss some possible candidate mechanisms that might underlie the factors identified for the Bi and Poo’s data set . Note that these brief indications are only intended to show the possible application of the procedure , not to make any strong claim on the possible specific mechanisms underlying such STDP data set . Pioneering studies on hippocampus have shown that a repeated stimulation of the perforant path fibres enhances the population response of downstream dentate granulate cells ( long-term potentiation–LTP; [47–49] ) . LTP also takes place in other parts of brain such as the cortex [50] , amygdala [51] , and the midbrain reward circuit [52] . Other studies have shown the existence of long-term depression ( LTD ) , complementary to LTP , in various parts of brain , for example hippocampus [53 , 54] and motoneurons [55] . More recent research has shown that LTP and LTD , and their intensity , depend on the duration of the temporal gap separating the pre- and post-synaptic spikes ( spike time-dependent plasticity—STDP; e . g . [56] , see [18] for a review ) . The relation between the time-delay and the synaptic change depends on the types of neurons involved ( e . g . , glutamatergic vs . GABAergic neurons [57 , 58] ) , the position of the synapse ( e . g . , [59] ) , and the experimental protocols used ( e . g . , [60] ) . Early findings that blocking NMDA receptors ( NMDARs ) can prevent both LTP and LTD , while a partial blocking can turn an LTP effect into an LTD , has led to the proposal of several calcium-based models of synaptic plasticity ( e . g . , [61–64] ) . One view proposes that two independent mechanisms can account for the classic STDP learning kernel [19 , 65] . This is in line with the two components , and their factors , found by our G-DHL based regression of Bi and Poo data set . The first component was an LTP ‘positive-derivative/positive-derivative’ component ( [ u ˙ 1 ] + [ u ˙ 2 ] + ) formed by two factors . The first factor was a pre-synaptic factor ( [ u ˙ 1 ] + ) lasting about 30 ms , compatible with a short-lived effect involving the pre-synaptic glutamatergic neuron spike and affecting the post-synaptic NMDARs [66] . The second factor was a post-synaptic factor ( [ u ˙ 2 ] + ) lasting about 7 ms , compatible with a back-propagating action potential ( BAP; [67] ) . The second component was a ‘positive-derivative/signal’ LTD component ( [ u ˙ 1 ] + u 2 ) formed by two factors: a relatively slow pre-synaptic element , ( [ u ˙ 1 ] + ) , lasting about 30 ms , and a slow post-synaptic element , ( u2 ) , lasting about 50 ms . Different biological mechanisms might underlie these two factors . In this respect , there is evidence that post-synaptic NMDARs might not be necessary for spike-timing-dependent LTD [68] , while this might be caused by metabotropic glutamate receptors ( mGluR; [69] ) , voltage gated calcium channels ( VGCC; [25 , 69] ) , pre-synaptic NMDAR [70] , or cannabinoid receptors [68 , 69] . We tested the generality of G-DHL by fitting all STDP kernels reported in the review of Caporale and Dan [18] . The data sets addressed in this review encompass many different STDP experiments reported in the literature and proposes a taxonomy to group them into distinct , and possibly exhaustive , classes . The taxonomy is first based on the excitatory or inhibitory nature of the pre- and post-synaptic neurons , giving the classes: ( a ) excitatory-excitatory; ( b ) excitatory-inhibitory; ( c ) inhibitory-excitatory; ( d ) inhibitory-inhibitory . Some neurons in different parts of brain belong to the same class but exhibit different STDP learning kernels: in [18] , these have been grouped in ‘subtypes’ ( sub-classes ) called ‘Type I’ , ‘Type II’ , etc . For the G-DHL regressions we used the original data when the authors of the experiments could furnish them . When this was not possible , we used the data extracted from graphs in the publications . Figs 10 and 11 summarise the outcome of the G-DHL-based regressions for the different data sets . For each data set , the figures report this information: ( a ) left graph: original data and , when available , regression curve of the original paper; ( b ) right graph: regression curve based on G-DHL; ( c ) top-center small graph: function with which the review [18] proposed to represent the STDP class of the data set . In the following , we illustrate the salient features of these regressions . Section 3 in S1 Supporting Information presents more detailed data on all the regressions as those presented in Fig 8 for the data set of Bi and Poo .
Understanding the functioning and learning in dynamical neural networks is challenging but also very important for advancing our theories and models of the brain—an exquisitely dynamical machine . Differential Hebbian Learning ( DHL ) might become a fundamental means to do so . Existing DHL rules are few , basically two [5 , 7] , and are not able to model most spike-timing dependent plasticity ( STDP ) phenomena found so far in the brain . Building on previous pioneering research , this work addresses these limitations in multiple ways . First , it proposes a framework to understand , use , and further develop DHL rules . In particular , it proposes a general DHL ( G-DHL ) rule encompassing existing DHL rules and generating many others , and highlights key issues related to the pre-processing of neural signals before the application of DHL rules . Second , it proposes procedures and formulas for applying DHL to model STDP in the brain . Third , it shows how the proposed G-DHL rule can model many classes of STDP observed in the brain and reviewed in [18] . With respect to other approaches for modelling STDP , DHL represents a complementary tool in the toolbox of the modeller and neuroscientist . First , DHL differs from ‘phenomenological models’ . Although simple and elegant , these models update the synapse based on mathematical functions directly mimicking the synaptic changes observed in empirical experiments in correspondence to different inter-spike intervals [14 , 15] . Instead , DHL rules compute the synaptic update on the basis of the step-by-step interactions between levels of and changes in the neural variables of interest . DHL rules also differ from ‘biophysical models’ . These models can reproduce many biological details but have high complexity and rely on phenomenon-specific mechanisms ( e . g . , [14 , 17] ) . Instead , DHL rules reproduce fewer empirical details but at the same time , after the systematisation proposed here , they represent ‘universal mechanisms’ able to capture many STDP phenomena . G-DHL relies on two main ideas . The first idea , elaborated starting from previous proposals [5] ( see also [29] ) , is that the derivative of an ‘event’ , intended as a monotonic increase followed by a monototic decrease of a signal , gives information on when the event starts and terminates . This information is used by G-DHL to update the connection weight depending on the time interval separating the pre- and post-synaptic neural events . The second idea is that the actual synaptic update can rely on different combinations of the possible interactions between the pre-/post-synaptic events and their derivatives , thus leading to a whole family of DHL rules . Mathematically , this gives rise to a compound structure of the G-DHL rule which is formed by a linear combination of multiple components . In this respect , the capacity of G-DHL to capture different STDP phenomena is linked to the power of kernel methods used in machine learning [34 , 35] . The linear form of the rule facilitates its application through manual tuning of its parameters , as shown here and in some previous neural-network models of animal behaviour using some components of the rule [80–82] . The linear form of the rule also facilitates the automatic estimation of its coefficients when used to capture STDP data sets , as also shown here . G-DHL has a high expressiveness , as shown here by the fact that we could use it to accurately fit multiple STDP data sets . In particular , the G-DHL components form basis functions that are well suited to model key aspects of STDP , in particular its long-term potentiation/depression features , its time span , and its variability around the zero inter-spike interval ( e . g . , sharp depression-potentiation passages , non-learning plateau , Hebbian/anti-Hebbian learning ) . The regressions of the data sets targeted here employed seven out of eight components of the rule . The regressions are particularly reliable because the optimisation procedure used here is highly robust with respect to local minima , so they show the utility of most G-DHL components for modelling different STDP data sets . Future empirical experiments might search for STDP processes corresponding to the eighth non-used G-DHL component ( encompassing a multiplication between the pre-synaptic stimulus and the post-synaptic derivative negative part ) : this corresponds to a relatively long LTD peaking at a negative inter-spike interval but also involving low-value positive intervals . The results of our regression based on G-DHL of the classic STDP kernel , represented by the classic Bi and Poo data set [25] , suggests the possible existence of two distinct mechanisms underlying LTP and LTD involved in such STDP learning kernel , so it is interesting to compare this result with different views in the literature . A specific hypothesis on calcium control of plasticity was formulated in [83] and was followed by significant experimental evidence . According to this hypothesis , post-synaptic calcium transients above a lower threshold cause LTD whereas calcium transients above a second higher threshold produce LTP . In a detail model [84] , this phenomenon is captured with a single mechanism for which the synaptic change is caused by calcium concentrations at the post-synaptic neuron modulated by the temporal relation between the current at the pre-synaptic neuron ( causing NMDAR opening ) and the back-propagating action potential ( BAP ) at the post-synaptic neuron [67]: low levels of post-synaptic calcium cause the synapse depression whereas high levels cause its enhancement . Models of such type have been criticised on the basis of empirical evidence . According to [65] , calcium models require a long-fading BAP-induced transients to account for LTD when the BAP occurs before the pre-synaptic action potential [12] . Moreover , calcium models also predict a pre-post form of LTD even when the BAP occurs beyond a given time from the pre-synaptic action potential . While this pre-post form of LTD has been registered in hippocampal slices [74] , other data [25] indicate that it is not a general feature of STDP . In this respect , our findings agree with other proposals for which two independent mechanisms account for LTP and LTD in the classic STDP learning kernel [19 , 65] . Future work might extend these preliminary results . In particular , it could aim to understand in detail how some of the mechanisms mentioned above implement change detectors and these lead to STDP , as predicted by the G-DHL core functioning mechanisms based on derivatives . Moreover , G-DHL could be used to heuristically guide the identification of the biophysical mechanisms underlying different STDP data sets beyond the classic kernel . Future work might also investigate , both computationally and empirically , DHL rules different from G-DHL , namely: ( a ) DHL rules formed by three or more components ( useful to model STDP involving more than two spikes [41] ) ; ( b ) DHL rules using orders of derivatives higher than the first one used in G-DHL [32 , 33]; ( c ) DHL rules generated by other types of filters , rather than [ u ˙ ] + and [ u ˙ ] - used in G-DHL , to detect the increasing and decreasing parts of events . Another line of research might aim to investigate the possible computational and behavioural functions of the different G-DHL components . In this respect , the analysis presented here on the computational mechanisms underlying STDP might contribute to the current research on the possible functions of such plasticity [20–23] . Indeed , this research mainly focuses on the computational function of the classic STDP learning kernel [25] , whereas the research presented here , by stressing how the brain uses different DHL rules , calls for the investigation of their different possible functions . A different approach to understand the functions of different DHL rules and STDP kernels might use embodied neural models to understand their utility to support adaptive behaviour . The development of G-DHL was in fact inspired by the need to implement specific learning processes in neural-network models able to autonomously acquire adaptive behaviours [80–82] . Thus , it could for example be possible to establish a particular target computation or behaviour and then automatically search ( e . g . with genetic algorithms or other optimisation techniques ) the rule components and coefficients that are best suited for them . For example , previous work [85] used a learning rule based on Kosco’s DHL rule [5] to obtain interesting/surprising emergent behaviours in physical simulated agents . This approach might test other G-DHL components to produce different behaviours . | Which learning rules can be used to capture the temporal relations between activation events involving pairs of neurons in artificial neural networks ? Previous computational research proposed various differential Hebbian learning ( DHL ) rules that rely on the activation of neurons and time derivatives of their activations to capture specific temporal relations between neural events . However , empirical research of brain plasticity , in particular plasticity depending on sequences of pairs of spikes involving the pre- and the post-synaptic neurons , i . e . , spike-timing-dependent plasticity ( STDP ) , shows that the brain uses a surprisingly wide variety of different learning mechanisms that cannot be captured by the DHL rules proposed so far . Here we propose a general differential Hebbian learning ( G-DHL ) rule able to generate all existing DHL rules and many others . We show various examples of how the rule can be used to update the synapse in many different ways based on the temporal relation between neural events in pairs of artificial neurons . Moreover , we show the flexibility of the G-DHL rule by applying it to successfully fit several different STDP processes recorded in the brain . Overall , the G-DHL rule represents a new tool for conducting research on learning processes that depend on the timing of signal events . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"learning",
"medicine",
"and",
"health",
"sciences",
"action",
"potentials",
"engineering",
"and",
"technology",
"nervous",
"system",
"signal",
"processing",
"membrane",
"potential",
"social",
"sciences",
"electrophysiology",
"neuroscience",
"learning",
"and",
"memory",
... | 2018 | General differential Hebbian learning: Capturing temporal relations between events in neural networks and the brain |
Filamentous fungi produce a diverse array of secondary metabolites ( SMs ) critical for defense , virulence , and communication . The metabolic pathways that produce SMs are found in contiguous gene clusters in fungal genomes , an atypical arrangement for metabolic pathways in other eukaryotes . Comparative studies of filamentous fungal species have shown that SM gene clusters are often either highly divergent or uniquely present in one or a handful of species , hampering efforts to determine the genetic basis and evolutionary drivers of SM gene cluster divergence . Here , we examined SM variation in 66 cosmopolitan strains of a single species , the opportunistic human pathogen Aspergillus fumigatus . Investigation of genome-wide within-species variation revealed 5 general types of variation in SM gene clusters: nonfunctional gene polymorphisms; gene gain and loss polymorphisms; whole cluster gain and loss polymorphisms; allelic polymorphisms , in which different alleles corresponded to distinct , nonhomologous clusters; and location polymorphisms , in which a cluster was found to differ in its genomic location across strains . These polymorphisms affect the function of representative A . fumigatus SM gene clusters , such as those involved in the production of gliotoxin , fumigaclavine , and helvolic acid as well as the function of clusters with undefined products . In addition to enabling the identification of polymorphisms , the detection of which requires extensive genome-wide synteny conservation ( e . g . , mobile gene clusters and nonhomologous cluster alleles ) , our approach also implicated multiple underlying genetic drivers , including point mutations , recombination , and genomic deletion and insertion events as well as horizontal gene transfer from distant fungi . Finally , most of the variants that we uncover within A . fumigatus have been previously hypothesized to contribute to SM gene cluster diversity across entire fungal classes and phyla . We suggest that the drivers of genetic diversity operating within a fungal species shown here are sufficient to explain SM cluster macroevolutionary patterns .
Filamentous fungi produce a diverse array of small molecules that function as toxins , antibiotics , and pigments [1] . Although by definition , these so-called specialized or secondary metabolites ( SMs ) are not strictly necessary for growth and development , they are critical to the lifestyle of filamentous fungi [2] . For example , antibiotic SMs give their fungal producers a competitive edge in environments crowded with other microbes [3] . SMs can additionally mediate communication between and within species as well as contribute to virulence on animal and plant hosts in pathogenic fungi [4 , 5] . A genomic hallmark of SMs in filamentous fungi is that the biosynthetic pathways that produce them are typically organized into contiguous gene clusters in the genome [6] . These gene clusters contain the chemical backbone synthesis genes whose enzymatic products produce a core metabolite , such as nonribosomal peptide synthases ( NRPSs ) and polyketide synthases ( PKSs ) , tailoring enzymes that chemically modify the metabolite , transporters involved in product export , and , often , transcription factors that control the expression of the clustered genes [6] . These gene clusters also occasionally contain resistance genes that confer self-protection against reactive or toxic metabolites [6] . Filamentous fungal genomes , particularly those in the phylum Ascomycota [6] , typically contain dozens of SM gene clusters . However , most individual SM gene clusters appear to be either species specific or narrowly taxonomically distributed in only a handful of species [6 , 7] . SM gene clusters that are more broadly distributed show discontinuous taxonomic distributions and are often highly divergent between species . Consequently , the identity and total number of SM gene clusters can vary widely even between very closely related species whose genomes exhibit very high sequence and synteny conservation [8 , 9] . In the last decade , several comparative studies have described macroevolutionary patterns of SM gene cluster diversity . For example , studies centered on genomic comparisons of closely related species have identified several different types of interspecies divergence , from single nucleotide substitutions ( e . g . , differences in fumonisins produced by Fusarium species are caused by variants in 1 gene [10] ) to gene gain/loss events ( e . g . , the trichothecene gene clusters in Fusarium species and the aflatoxin family SM gene clusters in Aspergillus species ) [11–16] and genomic rearrangements ( e . g . , the trichothecene gene clusters in Fusarium ) [11] . Additionally , genetic and genomic comparisons across fungal orders and classes have identified several instances of gene gain or loss [17–19] and horizontal gene transfer ( HGT ) [13 , 20–23] acting on individual genes or on entire gene clusters , providing explanations for the diversity and discontinuity of the taxonomic distribution of certain SM gene clusters across fungal species . Although interspecies comparative studies have substantially contributed to our understanding of SM diversity , the high levels of evolutionary divergence of SM clusters make inference of the genetic drivers of SM gene cluster evolution challenging; put simply , it has been difficult to “catch” the mechanisms that generate SM gene cluster variation “in the act . ” Several previous studies have examined intraspecies or population-level differences in individual SM gene clusters , typically focusing on the presence and frequency of nonfunctional alleles of clusters involved in the production of mycotoxins . Examples of clusters exhibiting such polymorphisms include the gibberellin gene cluster in F . oxysporum [24] , the fumonisin gene cluster in F . fujikuroi [25] , the aflatoxin and cyclopiazonic acid gene clusters in A . flavus [26] , and the bikaverin gene cluster in Botrytis cinerea [27] . While these studies have greatly advanced our understanding of SM gene cluster genetic variation and highlighted the importance of within-species analyses , studies examining the entirety of SM gene cluster polymorphisms within fungal species are so far lacking . We currently do not know the types and frequency of SM gene cluster polymorphisms within fungal species , whether these polymorphisms affect all types of SM gene clusters , or the genetic drivers of SM gene cluster evolution . To address these questions , we investigated the genetic diversity of all 36 known and predicted SM gene clusters in whole genome sequence data from 66 strains , 8 of which were sequenced in this study , of the opportunistic human pathogen A . fumigatus , a species with cosmopolitan distribution and panmictic population structure [28] . We found that 13 SM gene clusters were generally conserved and harbored low amounts of variation . In contrast , the remaining 23 SM gene clusters were highly variable and contained 1 or more of 5 different types of genetic variation: single nucleotide polymorphisms ( SNPs ) , including nonsense and frameshift variants , individual gene gain and loss polymorphisms , entire cluster gain and loss polymorphisms , polymorphisms associated with changes in cluster genomic location , and clusters with nonhomologous alleles resembling the idiomorphs of fungal mating loci . Many clusters contained interesting combinations of these types of polymorphisms , such as pseudogenization in some strains and entire cluster loss in others . The types of variants we find are likely generated by a combination of DNA replication and repair errors , recombination , genomic insertions and deletions , and horizontal transfer . We additionally find an enrichment for transposable elements ( TEs ) around horizontally transferred clusters , clusters that change in genomic locations , and idiomorphic clusters . Taken together , our results provide a guide to both the types of polymorphisms and the genetic drivers of SM gene cluster diversification in filamentous fungi . As most of the genetic variants that we observe have been previously associated with SM gene cluster diversity across much larger evolutionary distances and timescales , we argue that processes influencing SM gene cluster diversity within species are sufficient to explain SM cluster macroevolutionary patterns .
It is well established that SNPs and short indel polymorphisms , which are caused by errors in DNA replication and repair , are a major source of genomic variation [36] . Nonsynonymous SNPs and indels with missense , frameshift , and nonsense effects were widespread across the 33 SM reference gene clusters ( Fig 1 , S2 Table ) . Every strain contained numerous missense mutations and at least 1 nonsense or frameshift mutation in its SM gene clusters . Although missense mutations are likely to influence SM production , the functional effects of nonsense and frameshift mutations are comparatively easier to infer from genomic sequence data because they often result in truncated proteins and loss of protein function . SNPs and short indel polymorphisms can affect secondary metabolite production , as in the case of the lack of trypacidin production in the A1163 strain because of a previously identified frameshift mutation in the PKS of the trypacidin gene cluster [37] . Interestingly , we identified a premature stop codon ( Gln273* ) in a transcription factor required for trypacidin production , tpcD ( Afu4g14550 ) , in a strain sequenced in this study ( MO79587EXP ) ( S2 Table ) . These data suggest that function of this SM gene cluster has been lost at least twice , independently , in A . fumigatus . Individual nonsense or frameshift variants varied in frequency . For example , the NRPS pes3 gene ( Afu5g12730 ) in SM gene cluster 21 harbors 16 nonsense or frameshift polymorphisms in 55 strains , 7 of which are common ( present in ≥10 strains ) and another 7 of which are rare ( ≤5 strains ) . Strains with lab-mutated null alleles of the pes3 gene are more virulent than strains with functional copies [38] , which may explain the widespread occurrence of null pes3 alleles within A . fumigatus . We additionally identified several SM gene clusters that gained or lost genes in some strains . These gene content polymorphisms were most likely generated through genomic deletion or insertion events and were sometimes found at high frequencies among strains ( Fig 1 , Table 1 ) . In 3 cases , these polymorphisms impacted backbone synthesis genes , rendering the SM gene cluster nonfunctional . One example involves SM gene cluster 14 , whose standard composition includes a pyoverdine synthase gene , an NRPS-like gene , an NRPS backbone gene , and several additional modification genes ( Fig 2A ) . Four of the 66 strains examined lack an 11-kb region on the 3′ end of the cluster , which normally contains an NRPS gene and 2 additional cluster genes , and the first non-SM genes on the 3′ end flanking the cluster . All A . fumigatus strains contain a copia family TE [39 , 40] at the 3′ end of the cluster , suggesting that TEs may have been involved in the generation of this polymorphism . While this polymorphism could have arisen through a deletion event , a homologous cluster lacking the 11-kb region is also present in the reference genomes of A . lentulus and A . fischeri , close relatives of A . fumigatus ( Fig 2A ) . The most parsimonious explanation is that the genome of the A . fumigatus ancestor contained an SM gene cluster that lacked the 11-kb region and that this genomic region was subsequently gained and increased in frequency within A . fumigatus . The remaining 2 gene content polymorphisms affecting SM backbone genes were restricted to 1 strain each and appear to have arisen through genomic deletion events . Specifically , strain IF1SWF4 lacks an 8-kb region near the helvolic acid SM gene cluster , resulting in the loss of the backbone oxidosqualene cyclase gene as well an upstream region containing 2 non-SM genes ( S1 Fig ) . Strain LMB35Aa lacks a 54-kb region on the end of chromosome 2 , which includes 5 genes from the telomere-proximal fumigaclavine C cluster ( S1 Fig ) . Three other cases of gene content polymorphisms involved gene loss or truncation events of non-backbone structural genes . The second half of the open reading frame ( ORF ) of the gliM O-methyltransferase gene in the gliotoxin gene cluster has been lost in 2 of 66 strains ( S1 Fig ) and the first half of the permease fmqE in the fumiquinazoline gene cluster has been lost in 4 of 66 strains ( S1 Fig ) . Finally , an ATP-binding cassette ( ABC ) transporter gene in SM cluster 21 has been almost entirely lost in 21 of 66 strains ( S1 Fig ) . This deletion event is found in strains that are related in the SNP-based strain phylogeny but does not perfectly mirror the phylogeny ( Fig 1 ) . Several SM gene clusters were gained or lost entirely across strains . We observed several instances in which a cluster present in the genome of either the reference Af293 or A1163 ( also known as CEA10 ) strain was absent or pseudogenized in other strains , which we present in this section . One of the novel SM gene clusters , cluster 34 , was present in all but 2 of the strains ( Af293 and F7763 ) . Cluster 34 contains a PKS backbone gene , 1 PKS-like gene with a single PKS-associated domain , 9 genes with putative biosynthetic functions involved in secondary metabolism , and 6 hypothetical proteins ( Fig 2B ) . The 2 strains that lack cluster 34 contain a likely nonfunctional cluster fragment that includes the PKS-like gene , 2 biosynthetic genes , and 3 hypothetical proteins . Interestingly , the 3′ region flanking cluster 34 is syntenic across all 66 strains but the 5′ region is not , suggesting that a recombination or deletion event may have resulted in its loss in the Af293 and F7763 strains . These 2 strains form a clade in the strain phylogeny ( Fig 1 ) , so it is likely that this deletion or recombination event occurred once . One notable example of an SM gene cluster present in the Af293 reference genome but absent or pseudogenized in others was SM cluster 4 . This cluster contains 5 genes on the tip of the Af293 chromosome 1 and contains orthologs to 5 of the 6 genes in the fusarielin-producing gene cluster in F . graminearum [41] . Cluster 4 is also present in several other Aspergillus species , including A . clavatus and A . niger [41] , as well as in whole or in part in other non-Aspergillus fungi in the class Eurotiomycetes and in fungi in the class Sordariomycetes ( S3 Fig ) [30 , 42–50] . Phylogenetic analysis of the genes in cluster 4 does not provide a clear view of the origin of this cluster , which is consistent either with extensive gene loss in both Sordariomycetes and Eurotiomycetes or , alternatively , with HGT between fungi belonging to the 2 classes ( S2 and S3 Figs ) . Cluster 4 is entirely absent in 4 of 66 strains , and its genes are undergoing pseudogenization in an additional 43 strains via multiple independent mutational events ( Fig 3 ) . The 4 strains lacking the cluster form a single clade on the strain phylogeny , suggesting that the cluster was lost in a single deletion event ( Fig 1 ) . Further , 19 strains shared a single frameshift variant in the PKS gene ( 4380_4381insAATGGGCT; frameshift at Glu1461 in Afu1g17740 ) and an additional 13 strains shared a single frameshift variant ( 242delG; frameshift at Gly81 ) in an aldose 1-epimerase gene ( Afu1g17723 ) ( Fig 3A , S2 Table ) . Eleven other strains each contained 1 to several frameshift or nonsense polymorphisms involving 9 unique mutational sites . Five of these strains contained multiple distinct frameshifts and premature stop codons in more than 1 gene in the cluster , indicating that the entire pathway is pseudogenized in these strains . A phylogeny of the entire cluster 4 locus across all 62 strains with short-read data shows that 2 pseudogenizing variants shared across multiple strains , one in the aldose 1-epimerase gene and one in the PKS , are found in loci that form well-supported clades ( Fig 3B ) , suggesting that these variants arose once . Similarly , a set of variants shared across 3 strains and 1 variant shared in 2 strains are found in loci that form well-supported clades in the locus phylogeny . Two strains sharing a pseudogenizing variant in the PKS do not group together in the locus phylogeny , a discordance likely stemming from within-locus recombination events . Finally , functional alleles of cluster 4 are distributed throughout the locus phylogeny , suggesting that the functional allele is ancestral and the pseudogenized variants are derived . Perhaps surprisingly , loss-of-function polymorphisms ( from nonsense and frameshift mutations to wholesale cluster loss ) are common and sometimes frequent within A . fumigatus . The majority of these polymorphisms are presumably neutral and reflect the fact that any mutation is more likely to result in loss of a function than in gain . Consistent with this hypothesis is our observation that these loss events were often found at low frequencies . However , the possibility also exists that some of the high-frequency , recurrent loss-of-function polymorphisms may be adaptive . Given that many secondary metabolites are primarily secreted in the extracellular environment and can benefit nearby conspecifics that are not themselves producing the metabolite [51] , individual strains may be circumventing the energetically costly process of producing the metabolite themselves in a situation analogous to the Black Queen Hypothesis [52] . By searching for novel SM gene clusters in the genomes of the other 65 A . fumigatus strains , we found 3 SM gene clusters that were absent from the genome of the Af293 reference strain . As SM gene clusters are often present in repeat-rich and subtelomeric regions that are challenging to assemble [53 , 54] , the strains analyzed here might harbor additional novel SM gene clusters that were not captured here . One of these SM gene clusters , cluster 34 , was mentioned earlier as an example of whole gene cluster loss polymorphism ( Fig 2B ) and is present in most strains but has been lost in 2 strains . The other 2 SM gene clusters absent from the Af293 genome are present at lower frequencies and likely reflect gene cluster gain events; cluster 35 is present in 2 of 66 strains and cluster 36 in 4 of 66 strains . Cluster 35 is located in a region syntenic with an Af293 chromosome 4 region and is flanked on both sides by TEs ( S4 Fig ) . Eight of the 14 genes in this SM gene cluster are homologous to genes in an SM gene cluster in the genome of the insect pathogenic fungus Metarhizium anisopliae ( S4 Fig ) [55] . Phylogenetic analysis of these 8 genes is consistent with a horizontal transfer event ( S5 Fig ) . The 2 strains that contain this novel cluster are not sister to each other on the strain phylogeny ( Fig 1 ) . Cluster 36 is an NRPS-containing cluster located on shorter genomic scaffolds that lack homology to either the Af293 or A1163 genomes , making it impossible to determine on which chromosome this cluster is located ( S4 Fig ) . Two of the strains containing this novel cluster are sister to each other on the strain phylogeny , while the third is distantly related to these 2 ( Fig 1 ) . The evolutionary histories of the genes in the cluster are consistent with vertical inheritance , and these genes are present in multiple Aspergillus species . One of the most peculiar types of polymorphisms that we identified is a locus containing different unrelated alleles of SM gene clusters , reminiscent of the idiomorph alleles at the fungal mating loci [56] . This locus , which resides on chromosome 3 and corresponds to cluster 10 in the Af293 genome ( Fig 4 ) , was previously described as being strain specific in a comparison between Af293 and A1163 strains [30] and is thought to reside in a recombination hot spot [28] . Our analysis showed that there are at least 6 different alleles of this cluster in A . fumigatus containing 4 different types of key enzymes involved in natural product biosynthesis: a PKS-NRPS hybrid , a highly reducing ( HR ) PKS , a nonreducing ( NR ) PKS , and an NRPS-like enzyme ( Fig 4 ) . Two additional alleles were present in only 1 strain each ( S6 Fig ) . In the Af293 reference genome , the cluster present at the idiomorph locus contains 1 NR-PKS along with an NRPS-like gene ( allele B ) . In the A1163 reference genome and 17 other strains , there is a PKS-NRPS and an HR-NRPS at this locus ( allele E ) . These alleles show an almost complete lack of sequence similarity except for a conserved hypothetical protein and a fragment of the HR-PKS in the Af293 allele; in contrast , the upstream and downstream flanking regions of the 2 alleles , which do not contain any backbone genes , are syntenic . Remarkably , another allele , present in 12 strains , contains all of the genes from both the Af293 and A1163 clusters ( allele D ) . The remaining 3 alleles contain various combinations of these genes . One allele found in 22 strains contains some A1163-specific genes , including the HR-PKS , and no Af293-specific genes ( allele F ) , while another allele found in 3 strains contains some Af293-specific genes , including the NRPS-like gene , but no A1163 genes ( allele A ) . The final allele , present in 8 strains , contains the entire Af293 allele as well as part of the A1163 allele containing the HR-PKS ( allele C ) . Every allele is littered with multiple long terminal repeat sequence fragments from gypsy and copia TE families as well as with sequence fragments from DNA transposons from the mariner family [39] . In some cases , these TEs correspond with break points in synteny between alleles , suggesting that the diverse alleles of this SM gene cluster may have arisen via TE-driven recombination . Furthermore , both of the alleles that are restricted to a single strain have an insertion event of several genes near a TE , while the rest of the locus is highly similar to one of the more common alleles ( S6 Fig ) . Untargeted XCMS analysis [57] of an allele D strain ( 08-19-02-30 ) and 2 allele F strains ( 08-12-12-13 and 08-19-02-10 ) and comparison of their metabolite profiles revealed the presence of 2 unique masses in 08-19-02-30 ( S4 Table; S7 Fig ) , raising the possibility that variation at the idiomorph locus is functional . Further analysis is underway to investigate whether any of these mass to charge ratios can be directly linked to the allele D sequence . To gain insight into the evolutionary history of this locus , we constructed a phylogeny based on its conserved downstream flanking region ( Fig 4B ) . The resulting phylogeny shows some grouping of strains that share alleles , but there are no clades that contain all instances of a particular allele . This is likely to be the consequence of within-locus recombination between strains of A . fumigatus , which has been previously described at this locus [28] and which is potentially driven by the high number of repetitive sequences at this locus . While it is tempting to speculate that allele D , the longest allele containing all observed genes , represents the ancestral state , this does not explain the presence of a shared hypothetical protein and PKS gene fragment between allele C and allele B . Furthermore , 2 close relatives of A . fumigatus , A . lentulus and A . fischeri , contain a similar region with conserved upstream and downstream flanking genes that is highly dissimilar to any of the alleles observed in A . fumigatus ( S8 Fig ) . In both species , this locus contains numerous TEs as well as genes homologous to portions of allele E in A . fumigatus ( S8 Fig ) . A . fischeri additionally contains 2 hypothetical proteins from the PKS-NRPS region of A . fumigatus and an additional hybrid PKS-NRPS-containing gene cluster not found in either A . lentulus or any A . fumigatus strain ( S8 Fig ) . Other genes at this locus in both A . lentulus and A . fischeri have functions likely not related to SM . Interestingly , A . lentulus contains a gene with a heterokaryon incompatibility protein domain , which may be involved in determining vegetative incompatibility [58] . Only 1 representative genome from each species has been sequenced , but based on the high concentration of TEs and lack of sequence similarity with any A . fumigatus alleles , it is likely that this locus is highly variable within both A . lentulus and A . fischeri . It is possible that polymorphism at this locus originated via SM gene cluster fusion or fission events driven by TEs , which are present in large numbers . Interestingly , 2 other previously described instances of SM gene cluster variation bear some resemblance to the A . fumigatus idiomorphic SM gene cluster 10 locus . The first is the presence of 2 nonhomologous A . flavus alleles , for which some strains contain a 9-gene sesquiterpene-like SM gene cluster and others contain a nonhomologous 6-gene SM gene cluster at the same genomic location [35] . The second is the presence of 2 nonhomologous SM gene clusters at the same well-conserved locus in a comparison of 6 species of dermatophyte fungi [34] . Based on these results , we hypothesize that idiomorphic clusters may be common in fungal populations and contribute to the broad diversity of SM gene clusters across filamentous fungi . The final type of polymorphism that we observed is associated with SM gene clusters that are found in different genomic locations in different strains , suggesting that these SM gene clusters are behaving like mobile genetic elements . This type of polymorphism was observed in SM gene clusters 1 and 33 , both of which produce as-yet-identified products and are present at low frequencies in A . fumigatus strains . SM gene cluster 1 , which is present in 6 strains at 3 different genomic locations ( Fig 5A ) , consists of a PKS and 4 other structural genes that are always flanked by a 15-kb region ( upstream ) and a 43-kb region ( downstream ) containing TEs . In the reference Af293 strain and in strain F7763 , cluster 1 and its flanking regions are located on chromosome 1 , while in strains 08-31-08-91 , F13619 , and Z5 they are located between Afu4g07320 and Afu4g07340 on chromosome 4 . In contrast , in strain JCM10253 , the cluster and flanking regions are located on chromosome 8 immediately adjacent to the 3′ end of the intertwined fumagillin and pseurotin SM gene supercluster [59] . The strains containing the allele on chromosome 1 are sister to each other on the strain phylogeny , while the other strains are scattered across the tree and do not reflect the phylogeny ( Fig 1 ) . In 5 of 6 strains , cluster 1 appears to be functional and does not contain nonsense SNPs or indels . However , the cluster found on chromosome 1 in strain F7763 contains 2 stop codons in the oxidoreductase gene ( Gln121* and Gln220* ) and 2 premature stop codons in the PKS ( Gln1156* and Gln1542* ) , suggesting this strain contains a null allele . This “jumping” gene cluster is not present in any other sequenced genome in the genus Aspergillus , and phylogenetic analysis of its constituent genes is consistent with HGT between fungi ( S9 Fig ) . Specifically , this gene cluster is also present in Phaeosphaeria nodorum [60] , a plant pathogen from the class Dothideomycetes , Pseudogymnoascus pannorum [61] , a fungus isolated from permafrost from the Leotiomycetes , and Escovopsis weberi [62] , a fungal parasite of fungus-growing ants from the Sordariomycetes ( Fig 5B ) . One additional species , the endophyte Hypoxylon sp . CI4A from the class Sordariomycetes [63] , contains 4 of the 5 cluster genes but is missing Afu1g00970 , an MFS drug transporter . However , this species contains a gene unrelated to Afu1g00970 that is annotated as an MFS drug transporter immediately adjacent to this cluster ( Fig 5B ) . None of these fungi contain the upstream or downstream TE-rich flanking regions present in A . fumigatus , and each fungus contains additional unique genes with putative biosynthetic functions adjacent to the transferred cluster . The most likely explanation for this change in flanking regions is that this SM gene cluster was transferred into A . fumigatus once and has subsequently “jumped” in different genomic locations in different strains . The second SM gene cluster that shows variation in its genomic location across strains , cluster 33 , contains a terpene synthase . This cluster is present in only 5 strains at 3 distinct locations ( Fig 5C ) . Similar to cluster 1 , cluster 33 is also flanked by TEs , and in 1 strain the cluster is located in a new region 58 Kb from SM gene cluster 34 . Two strains that contain the cluster in the same genomic location are sister to each other on the strain phylogeny , while the placement of the other 3 strains containing the cluster does not reflect the phylogeny ( Fig 1 ) . In contrast to cluster 1 , cluster 33 does not appear to have been horizontally transferred between fungi and its genes are present in other sequenced Aspergillus species [64] , suggesting that the mobility of clusters 1 and 33 may be driven by different mechanisms . Interestingly , both cases of mobile gene clusters are located near or immediately adjacent to other SM gene clusters in some strains . Cluster 33 is located 58 kb away from cluster 34 in one strain , and cluster 1 is located immediately adjacent to the intertwined fumagillin and pseurotin supercluster [59] in another . This supercluster is regulated by the transcriptional factor fapR ( Afu8g00420 ) and is located in a chromosomal region controlled by the master SM regulators laeA ( Afu1g14660 ) and veA ( Afu1g12490 ) [59 , 65] , raising the hypothesis that mobile gene clusters might be co-opting the regulatory machinery acting on adjacent SM gene clusters . Previous work has hypothesized that the fumagillin and pseurotin supercluster formed through genomic rearrangement events , placing the once-independent gene clusters in close proximity to each other [59] . Our observation that the mobile cluster 1 is located in this same region not only supports this hypothesis but also implicates TEs as one of the mechanisms by which superclusters are formed . These superclusters may also represent an intermediate stage in the formation of new SM gene clusters . Supercluster formation , potentially mediated by mobile gene clusters and followed by gene loss , could explain macroevolutionary patterns of SM gene clusters in which clustered genes in one species are found to be dispersed over multiple gene clusters in other species [9 , 11] .
Our examination of the genomes of 66 strains of A . fumigatus revealed 5 general types of polymorphisms that describe variation in SM gene clusters . These polymorphisms include variation in SNPs and short indels , gene and gene cluster gains and losses , nonhomologous ( idiomorph ) gene clusters at the same genomic position , and mobile clusters that differ in their genomic location across strains ( Fig 6 ) . Previous work has demonstrated that SM gene clusters , like the metabolites that they produce , are highly divergent between fungal species [8 , 9 , 19 , 64] . Our examination of genome-wide variation shows that these SM gene clusters are also diverse across strains of a single fungal species . These results also demonstrate that the diversity of SM gene clusters within A . fumigatus cannot be captured by sequencing a single representative strain , which is the current standard practice for determining the SM gene cluster content of a fungal species . The quantification of diversity in SM gene clusters within a species is dependent on both numbers and types of strains analyzed . The types of polymorphisms detected as well as their observed frequency , especially for rare polymorphisms , will increase with the number of genomes examined . In addition , both the frequencies of the different types of polymorphisms and the polymorphisms themselves may also change with sampling design or in a manner corresponding to the population structure or ecology of the species under study . A . fumigatus is a cosmopolitan species with panmictic population structure [28] , characteristics that do not always apply to other filamentous fungi . Fungi exhibiting strong population structure or fungi adapted to different ecological niches might contain different patterns of genetic diversity . Nevertheless , the variants and genetic drivers we observe at the within-species level are also implicated as driving SM gene cluster variation at the between-species level , suggesting that the observed microevolutionary processes are sufficient to explain macroevolutionary patterns of SM gene cluster evolution . For example , the narrow and discontinuous distribution of SM gene clusters across the fungal phylogeny has been attributed to HGT as well as to gene cluster loss [13 , 15 , 20 , 22 , 30 , 66–68] . Here , we find evidence that both processes also influence the distribution of SM gene clusters within a species ( Figs 2 and 5 , S2–S5 Figs ) . Interestingly , the fraction of SM gene clusters within A . fumigatus that harbor loss of function polymorphisms is substantial , consistent with the macroevolutionary view that SM gene cluster loss is rampant [18 , 19 , 68] . However , our within-species observations are also consistent with the macroevolutionary importance of HGT to SM gene cluster evolution . Once thought to be nonexistent in eukaryotes , HGT is now considered to be responsible for the presence of several different SM gene clusters in diverse filamentous fungi [13 , 68 , 69] . The instances of HGT of SM gene clusters within A . fumigatus suggests that acquisition of foreign genetic material containing SM gene clusters is likely a common and ongoing occurrence in fungal populations . One recurring theme across different types of SM gene cluster polymorphisms in A . fumigatus was the perpetual presence of TEs adjacent to or within clusters . One particularly striking case is the “idiomorphic” cluster 10 , in which TEs seem to correspond with break points in synteny both within A . fumigatus and also between A . fumigatus and its close relatives ( Fig 4 , S8 Fig ) . TEs were also present flanking mobile and horizontally transferred SM gene clusters and were located adjacent to gene gain sites . There are several potential explanations for the observed TE enrichment . First , TE presence may promote repeat-driven recombination and gene rearrangement , or the TEs themselves may be the agents of horizontally transferred clusters ( either on their own or through a viral vector ) . Alternatively , it may simply be the case that SM gene clusters preferentially reside in TE-rich genomic regions . In summary , examination of SM gene cluster variation within a single fungal species revealed 5 distinct types of polymorphism that are widespread across different types of SM gene clusters and are caused by many underlying genetic drivers , including errors in DNA transcription and repair , nonhomologous recombination , gene duplication and loss , and HGT . The net effect of the observed variation raises the hypothesis that the chemical products of filamentous fungal species are in a state of evolutionary flux , each population constantly altering its SM gene cluster repertoire and consequently modifying its chemodiversity .
Eight strains of A . fumigatus were isolated from 4 patients with recurrent cases of aspergillosis in the Portuguese Oncology Institute in Porto , Portugal . Each strain was determined to be A . fumigatus using macroscopic features of the culture and microscopic morphology observed in the slide preparation from the colonies with lactophenol solution [70] . Based on the morphological characterization , all clinical strains were classified as A . fumigatus complex-Fumigati . After whole genome sequencing , retrieval and examination of the beta tubulin and calmodulin sequences of each strain confirmed that all strains belonged to A . fumigatus ( see Phylogenetic analysis and S9 Fig ) . The genomes of all 8 strains were sequenced using 150-bp Illumina paired-end sequence reads at the Genomic Services Lab of Hudson Alpha ( Huntsville , Alabama , USA ) . Genomic libraries were constructed with the Illumina TruSeq library kit and sequenced on an Illumina HiSeq 2500 sequencer . Samples of all 8 strains were sequenced at greater than 180X coverage or depth ( S1 Table ) . Short-read sequences for these 8 strains are available in the NCBI Sequence Read Archive ( SRA ) under accession SRP109032 ( https://trace . ncbi . nlm . nih . gov/Traces/sra/ ? study=SRP109032 ) . In addition to the 8 strains sequenced in this study , we retrieved 58 A . fumigatus strains with publicly available whole genome sequencing data , resulting in a dataset of 66 strains ( S1 Table ) . The strains used included both environmental and clinical strains and were isolated from multiple continents . Genome assemblies for 10 of these strains , including the Af293 and A1163 reference strains , were available for download from GenBank [28–32 , 71] . For 6 of these strains , short-read sequences were also available from the NCBI SRA , which were used for variant discovery only ( see Single nucleotide variant [SNV] and indel discovery ) and not for genome assembly . Short-read sequences were not available for the remaining 4 strains . Short-read sequences were downloaded for an additional 48 strains from the NCBI SRA if they were sequenced with paired-end reads and at greater than 30X coverage . All strains with available short-read data ( 62 of 66 strains ) were aligned to both the Af293 and A1163 reference genomes using BWA mem version 0 . 7 . 12-r1044 [72] . Coverage of genes present in the reference genome was calculated using bedtools v2 . 25 . 0 [73] . SNV and indel discovery and genotyping were performed relative to the Af293 reference genome and were conducted across all samples simultaneously using the Genome Analysis Toolkit version 3 . 5-0-g36282e4 with recommended hard filtering parameters [74–76] and annotated using snpEff version 4 . 2 [77] . All 56 strains without publicly available genome assemblies were de novo assembled using the iWGS pipeline [78] . Specifically , all strains were assembled using SPAdes v3 . 6 . 2 and MaSuRCA v3 . 1 . 3 and resulting assemblies were evaluated using QUAST v3 . 2 [79–81] . The average N50 of assemblies constructed with this strategy was 463 kb ( S1 Table ) . Genes were annotated in these assemblies as well as in 5 GenBank assemblies with no predicted genes using augustus v3 . 2 . 2 trained on A . fumigatus gene models [82] . Repetitive elements were annotated in all assemblies using RepeatMasker version open-4 . 0 . 6 [83] . Secondary metabolic gene clusters in the Af293 reference genome were taken from 2 recent reviews , both of which considered computational and experimental data to delineate cluster boundaries [84 , 85] ( S3 Table ) . The genomes of the other 65 strains were scanned for novel SM gene clusters using antiSMASH v3 . 0 . 5 . 1 [86] . To prevent potential assembly errors from confounding the analysis , any inference about changes in genomic locations of genes or gene clusters was additionally verified by manually inspecting alignments and ensuring that paired end reads supported an alternative genomic location ( see Single nucleotide variant [SNV] and indel discovery ) . Cases in which paired end reads did not support the change in genomic location ( i . e . , all 3′ read mapping to chromosome 1 and all 5′ pairs mapping to chromosome 8 ) or mapping was ambiguous or low quality were discarded . To confirm all strains in this analysis belonged to the species A . fumigatus , the genomic sequences of the beta tubulin and calmodulin genes were extracted from the assembled genomes of all strains . Gene phylogenies were constructed using A . fischerianus as an out-group using RAxML v8 . 0 . 25 with the GTRGAMMA substitution model [87] . The tree was midpoint rooted and all branches with bootstrap support less than 80% were collapsed ( S10 Fig ) . To construct an SNP-based strain phylogeny , biallelic SNPs with no missing data were pruned using SNPRelate v1 . 8 . 0 with a linkage disequilibrium threshold of 0 . 8 [88] . A total of 15 , 274 SNVs were used to create a phylogeny using RAxML v8 . 0 . 25 with the ASC_BINGAMMA substitution model [87] . The tree was midpoint rooted and all branches with bootstrap support less than 80% were collapsed . The phylogeny was visualized using ITOL version 3 . 0 [89] . To understand the evolutionary histories of specific SM gene clusters showing unusual taxonomic distributions , we reconstructed the phylogenetic trees of their SM genes . Specifically , SM cluster protein sequences were queried against a local copy of the NCBI nonredundant protein database ( downloaded May 30 , 2017 ) using phmmer , a member of the HMMER3 software suite [90] , using acceleration parameters—F1 1e-5—F2 1e-7—F3 1e-10 . A custom perl script sorted the phmmer results based on the normalized bitscore ( nbs ) , in which nbs was calculated as the bitscore of the single best-scoring domain in the hit sequence divided by the best bitscore possible for the query sequence ( i . e . , the bitscore of the query aligned to itself ) . No more than 5 hits were retained for each unique NCBI Taxonomy ID . Full-length proteins corresponding to the top 100 hits ( E-value < 1 × 10 − 10 ) to each query sequence were extracted from the local database using esl-sfetch [90] . Sequences were aligned with MAFFT v7 . 310 using the E-INS-i strategy and the BLOSUM30 amino acid scoring matrix [91] and trimmed with trimAL v1 . 4 . rev15 using its gappyout strategy [92] . The topologies were inferred using maximum likelihood , as implemented in RAxML v8 . 2 . 9 [87] , using empirically determined substitution models and rapid bootstrapping ( 1 , 000 replications ) . The phylogenies were midpoint rooted and branches with less than 80% bootstrap support were collapsed using the ape and phangorn R packages [93 , 94] . Phylogenies were visualized using ITOL version 3 . 0 [89] . To understand the evolutionary histories of SM gene clusters 4 and 10 , full-length nucleotide sequences of all 62 strains with short-read sequence data were extracted for the entire cluster region ( SM gene cluster 4 ) or the downstream flanking region ( SM gene cluster 10 ) using the previously described SNV analysis procedure followed by Genome Analysis Toolkit’s “ExtractAlternativeReferenceFasta” tool [75] . The resulting nucleotide sequences were aligned using MAFFT v7 . 310 [91] . Phylogenies were constructed using maximum likelihood as implemented in RAxML v 8 . 0 . 25 , using the GTRGAMMA substitution model and rapid bootstrapping ( 1 , 000 replications ) [87] . Phylogenies were midpoint rooted and branches with less than 80% bootstrap support were collapsed . Phylogenies were visualized using ITOL version 3 . 0 [89] . All sequence alignments and phylogenies generated in this study are available on the Figshare repository ( https://figshare . com/projects/Data_for_Drivers_of_genetic_diversity_in_secondary_metabolic_gene_clusters_within_a_fungal_species_/26089 ) . For natural product analysis , 5 × 106 spores/mL for the indicated strains were grown in 50 mL liquid GMM [95] for 5 days at 25°C and 250 rpm in duplicates . Supernatants were extracted with equal volumes of ethyl acetate , dried down and resuspended in 20% acetonitrile ( ACN ) . Each sample was analyzed by ultra high-performance liquid chromatography ( UHPLC ) coupled with mass spectrometry ( MS ) . The samples were separated on a ZORBAX Eclipse XDB-C18 column ( Agilent , 2 . 1 × 150 mm with a 1 . 8 μM particle size ) using a binary gradient of 0 . 5% ( v/v ) formic acid ( FA ) as solvent A and 0 . 5% ( v/v ) FA in ACN as solvent B that was delivered by a VanquishTM UHPLC system ( Thermo Scientific ) with a flow rate of 0 . 2 mL/min . The binary gradient started with 20% B that was increased with a linear gradient to 100% B in 15 min followed by an isocratic step at 100% B for 5 min . Before every run , the system was equilibrated for 5 min at 20% B . The UHPLC system was coupled to a Q Exactive hybrid quadrupole OritrapTM MS ( Thermo Scientific ) . For electrospray ionization , the ion voltage was set at ±3 . 5 kV in positive and negative mode . Nitrogen was used as sheath gas at a flow rate of 45 and as sweep gas at a flow rate of 2 . Data analysis was performed using XCMS [57] and Maven [96] software . | All organisms produce metabolites , which are small molecules important for growth , reproduction , and other essential functions . Some organisms , including fungi , plants , and bacteria , make specialized forms of metabolites known as “secondary” metabolites that are ecologically important and improve their producers’ chances of survival and reproduction . In fungi , the genes in pathways that synthesize secondary metabolites are typically located next to each other in the genome and organized in contiguous gene clusters . These gene clusters , along with the metabolites they produce , are highly distinct , even between otherwise similar fungi , and it is often difficult to reconstruct how these differences evolved . To understand how secondary metabolic pathways evolve in fungi , we compared secondary metabolic gene clusters in 66 strains of one species of filamentous fungus , the human pathogen Aspergillus fumigatus . We show that these gene clusters vary extensively within this species , and describe the genetic processes that cause these differences . We identify 5 types of variants: single nucleotide changes , gene and gene cluster gain and loss , different gene clusters at the same genomic position , and mobile gene clusters that “jump” around the genome . These results provide a road map to the types and frequencies of genomic changes underlying the extensive diversity of fungal secondary metabolites . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"taxonomy",
"aspergillus",
"fumigatus",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"aspergillus",
"fungal",
"genetics",
"pathogens",
"microbiology",
"fungi",
"phylogenetics",
"data",
"management",
"metabolites",
"fungal",
"patho... | 2017 | Drivers of genetic diversity in secondary metabolic gene clusters within a fungal species |
As a result of evolution , the biology of triatomines must have been significantly adapted to accommodate trypanosome infection in a complex network of vector-vertebrate-parasite interactions . Arthropod-borne parasites have probably developed mechanisms , largely still unknown , to exploit the vector-vertebrate host interactions to ensure their transmission to suitable hosts . Triatomines exhibit a strong negative phototaxis and nocturnal activity , believed to be important for insect survival against its predators . In this study we quantified phototaxis and locomotion in starved fifth instar nymphs of Rhodnius prolixus infected with Trypanosoma cruzi or Trypanosoma rangeli . T . cruzi infection did not alter insect phototaxis , but induced an overall 20% decrease in the number of bug locomotory events . Furthermore , the significant differences induced by this parasite were concentrated at the beginning of the scotophase . Conversely , T . rangeli modified both behaviors , as it significantly decreased bug negative phototaxis , while it induced a 23% increase in the number of locomotory events in infected bugs . In this case , the significant effects were observed during the photophase . We also investigated the expression of Rpfor , the triatomine ortholog of the foraging gene known to modulate locomotion in other insects , and found a 4 . 8 fold increase for T . rangeli infected insects . We demonstrated for the first time that trypanosome infection modulates the locomotory activity of the invertebrate host . T . rangeli infection seems to be more broadly effective , as besides affecting the intensity of locomotion this parasite also diminished negative phototaxis and the expression of a behavior-associated gene in the triatomine vector .
The triatomine bug Rhodnius prolixus ( Hemiptera: Reduviidae ) is an important vector of Chagas disease in Northern South America . Parasites transmitted by blood-sucking insects are responsible for nearly 16% of the global burden of transmissible diseases [1] and their dispersal is highly dependent on the behavior of their arthropod vectors . Some vector-borne parasites have been shown to modify physiological and behavioral traits of both vectors and vertebrate hosts , in a way that increases the probability of transmission [2] . Identifying the underlying bases of these parasite-induced alterations is of great importance to understand disease biology which may lead to the design of new control measures [3] . T . cruzi , the causative agent of Chagas disease , is generally considered as non-pathogenic to triatomines because several previous works showed no pathological effects of this parasite to triatomines [4–6] . Nevertheless some other reports have shown costs induced by T . cruzi infections on its insect vectors [7–10] . A recent study published in our laboratory has clearly shown that T . cruzi infection decreases the fecundity and fertility of R . prolixus adults [10] . Trypanosoma rangeli is also transmitted by Rhodnius species and does not cause disease to humans , but its pathogenicity for triatomines has been confirmed by many authors [11–13] . Studies recently published by our laboratory have shown that infection with T . rangeli extends intermolt periods [14] and affects reproductive parameters in R . prolixus [14] . Triatomines present a bimodal pattern of daily activity , leaving their refuges when light intensity declines and displaying most of their activity during the first hours of the scotophase [15–17] . Before sunrise , they return to their refuges mainly guided by chemical signals [18] . During daylight hours , bugs remain aggregated inside these shelters and present low locomotory activity [17] . Different factors are considered to modulate the locomotion of insect vectors , e . g . Culex annulirostris varies its activity pattern as an adaptation to local climate [19] . The locomotory activity of Cimex lectularius is known to be controlled by nutritional status , as starved insects are less active than recently fed ones [20] . Alterations in locomotory activity induced by physiological changes may be mediated by modulation of gene expression and/or posttranscriptional mechanisms [21] . Interestingly , the sand fly Lutzomyia longipalpis shows a reduction in locomotory activity correlated with a downregulation in the expression of period and timeless genes after blood meals [22] . cGMP-dependent protein kinases ( PKGs ) are serine/threonine kinases [23] found in diverse organisms from paramecia to humans [24] . A PKG encoded by the gene named foraging was first reported to control the locomotory activity of Drosophila melanogaster [25] . The role of these proteins in modulating foraging behavior is highly conserved across species [26] . Recently our laboratory group has been investigating the effects of trypanosomes on their invertebrate hosts [10 , 14 , 27] . Understanding how pathogens modify activities such as locomotion will contribute to our understanding of pathogen transmission dynamics . In the present study two behavioral parameters were evaluated in R . prolixus nymphs infected with either T . cruzi or T . rangeli . We first evaluated the potential effects of parasites on the negative phototaxis of these insects , i . e . , tested whether they have a weaker avoidance for illuminated places . Secondly , we developed an actometer to evaluate patterns of locomotory activity and characterized alterations induced by trypanosome infection . Finally , we characterized the expression levels of the R . prolixus PKG orthologue ( Rpfor ) in the brain of healthy and infected insects .
All experiments using live animals were performed in accordance with FIOCRUZ guidelines on animal experimentation and were approved by the Ethics Committee in Animal Experimentation ( CEUA/FIOCRUZ ) under the approved protocol number L-058/08 . This protocol adheres to the guidelines of CONCEA/MCT ( http://www . cobea . org . br/ ) , which is the maximum ethics committee of the Brazilian government . Rhodnius prolixus: insects were obtained from a laboratory colony derived from insects collected in Honduras around 1990 . The colony was maintained at 26±1°C , 50±5% RH and exposed to a natural illumination cycle . Insects were consistently fed on diverse sources of blood that included mice , chicken and a membrane feeder offering citrated rabbit blood at 37°C . Living hosts were anesthetized with an intraperitoneal injection of a ketamine ( 150 mg/kg; Cristália , Brazil ) /xylazine ( 10 mg/kg; Bayer , Brazil ) mixture . Fifth instar nymphs starved for 30 days after ecdysis were used in all the experiments . Trypanosoma cruzi CL and Trypanosoma rangeli CHOACHI strains originally isolated from naturally infected T . infestans [28] and R . prolixus [29] , respectively , were used in this study . The epimastigote forms used to infect insects were cultured by twice weekly passages at 27°C in liver-infusion tryptose ( LIT ) medium supplemented with 15% fetal bovine serum , 100 mg/ml of streptomycin and 100 units/ml of penicillin . Negative phototaxis , i . e . the guidance reaction in which animals steer their way against light , allows the location of shelters and also helps the insect to avoid exposure to predators [31] . To study the phototactic responses of these R . prolixus , experiments were performed as described before [32] . A rectangular glass box ( 14×5×5cm ) was divided lengthwise into two experimental arenas ( 2 . 5cm width each ) and covered with a rectangular acrylic lid . The parallel design allowed the simultaneous evaluation of individual uninfected and infected insects . Half of each arena was maintained in the dark by a black piece of cardboard fixed on the cover of the box and the other half remained illuminated ( 190 lux ) . An initial batch of 70 nymphs ( control = 35; infected = 35 ) was used in the assays that evaluated T . cruzi infection . For T . rangeli infections , 84 nymphs were used ( control = 42; infected = 42 ) . Insects that did not move during the trials were excluded from the analysis . Therefore , the analyses were performed comparing the responses of 34 control vs 28 T . cruzi infected bugs , and 42 control vs 39 T . rangeli infected bugs . In each trial , both nymphs were individually placed at one end of each arena inside small dark bowls ( light and dark sides were alternated in subsequent assays ) . After 40s the bowls were removed and the trial started . Each trial lasted for 10min during which two behavioral parameters were measured , negative phototaxis ( proportion of time spent in the dark side of the arena ) and activity ( number of times that the insect crossed the middle line of the arena ) . All assays were performed during the three first hours of scotophase , period in which triatomines normally exhibit an activity peak related to food search [16] . In order to record the locomotory activity of triatomine bugs , we developed an automatic actometer system ( Fig 1 ) which was set up inside a controlled environment chamber ( 25±°C , 50±5%RH , photoperiod of 12:12 L/D ) . This device consisted of 40 individual arenas ( 10x5x2cm ) arranged on an aluminum plate , each one presenting three light barriers , each constituted by a light emitting diode ( LED ) and a phototransistor ( Fig 1 ) . The activity of each insect was restricted to the interior of an acrylic container that acted as arena . During a trial , every time a moving insect interrupted a light beam , a signal was generated and recorded by an ad hoc software . Therefore , each of these signals was considered a locomotory activity event . During the photophase , fluorescent tubes located overhead illuminated the chamber at a light intensity of ca . 60 LUX . For the measurement of locomotory activity , 36 nymphs ( control , n = 18; infected , n = 18 ) were placed individually in the arenas . Each container had filter paper as substrate and was covered with a rectangular acrylic lid . Insects were maintained in these conditions for 6 days , during which their activity was continuously recorded . This procedure was replicated three times for T . cruzi ( uninfected , n = 54; infected , n = 54 ) and six times for T . rangeli ( uninfected , n = 108; infected , n = 108 ) infection experiments . Brains and pieces of surrounding muscles were dissected on a freeze cold dissecting dish ( BioQuip , Gardena , CA , US ) and remained frozen during the entire dissection procedure . Samples consisted of pools of these tissues dissected from five 5th instar nymphs ( n = 3 biological replicates for each treatment , except for insects used as controls for T . rangeli infection , n = 2 ) . RNA was extracted on the same day in which dissection happened using TRI Reagent ( Sigma-Aldrich , St . Louis , MO , US ) according to the manufacturer’s instructions . RNA concentrations were determined using a BIOPhotometer ( Eppendorf , Hamburg , Germany ) . Total RNA ( 500ng ) was used for reverse transcription using M-MLV Reverse Transcriptase ( Promega , Fitchburg , WI , US ) and a modified oligo dT primer ( MgdT 5’-CGGGCAGTGAGCAACG ( T12 ) -3’ ) as described [39] . Quantitative PCR ( qPCR ) was used to assess whether Rpfor expression levels were affected by trypanosome infection . All reactions contained 1μl of cDNA , 5ng/μl of each primer and 6μl of PerfeCTa SYBR Green Super Mix ( Quanta Biosciences , Gaithersburg , MD , US ) in a final volume of 10μl . The reactions were conducted in a RotorGene 3000 thermal cycler ( Corbett Research , Sydney , Australia ) . The qPCR conditions used were: 95°C: 2min , 35 cycles of 95°C: 10s , 60°C ( Rpfor ) , 55°C ( β-actin and GADPH ) : 10s and 72°C: 30s , followed by a melting curve analysis to confirm the specificity of the reaction . In all qPCR experiments , no-template controls ( NTC ) were included for each primer set to verify the absence of exogenous DNA and primer-dimers . Reactions on each sample were run in duplicate . Relative differences in abundance of Rpfor transcripts were calculated using the 2–ΔΔCt method [40] with β-actin as reference gene as described [41] . The PCR efficiencies ( E ) and repeatability ( R2 ) for each primer were determined using the slope of a linear regression model ( Table 1 ) [42] . All data were normalized relative to values recorded for control insects . Statistical analyses were performed using R 3 . 0 . 2 [43] . Negative phototaxis ( i . e . proportion of time spent in the dark side of the arena ) in uninfected vs . infected insects was analyzed using linear regression with beta distribution and logit link function ( function betareg in betareg package ) [44] , since values could only range from 0 to 1 . To compare the activity ( i . e . number of times insects crossed the middle line ) in the same experiment , generalized linear regression with Poisson error distribution and log link function was used . The locomotory activity of R . prolixus individuals in both experiments ( i . e . effects of T . cruzi and T . rangeli ) was analyzed with linear mixed-effects models ( function lmer in lme4 package ) [45] fitted by restricted maximum likelihood ( REML ) . Square root transformed locomotory activity data were used as a response variable whereas treatment ( uninfected vs . infected ) and time of the day ( 24 hour period ) were set as fixed variables . To take into account that the data were measured repeatedly on the same individuals with time intervals of one hour , individuals were included as random factor . Likewise , time of the day was also measured repeatedly over the experiment and it was hence included as random factor nested within the variable day . Pairwise contrasts ( function testInteractions in phia package ) [46] were used to evaluate the locomotory activity of uninfected and infected individuals at every hour of the day . P-values of the contrasts were adjusted by Holm-Bonferroni method to correct for the problem of multiple comparisons . D . melanogaster ( Q03043 ) ; Apfor ( H6V8U7 ) ; Agfor ( V5GSC1 ) ; Amfor ( T1SGQ4 ) ; Btfor ( C6GBY7 ) ; Tcfor ( D6WXB3 ) ; Phfor ( E0VGN7 ) .
Triatomine insects exhibit a strong aversion to light as an adaptive aspect of their behavior . As trypanosome infection has been shown to affect several parameters of the biology of their invertebrate vectors , we decided to investigate whether these parasites could influence triatomine negative phototaxis . By analyzing the amount of time that uninfected controls , as well as T . cruzi or T . rangeli-infected insects , spent on either the dark or the light sides of a chamber we evaluated this behavior . During the assays , uninfected and T . cruzi-infected insects crossed the middle line between the two sectors of the arena on average 14 and 13 times , respectively ( Z = 55 . 77 , p = 0 . 39 ) . Uninfected R . prolixus spent more than 70% of the time on the dark side of the arena and T . cruzi-infected insects showed a similar behavior ( Fig 3; Z = 4 . 74 , p = 0 . 92 ) . Likewise , our experiments with T . rangeli-infected insects showed that the number of crossings between the two sections of the arena was not significantly different between uninfected or T . rangeli-infected triatomines ( 12 and 11 times respectively on average , Z = 55 . 36 , p = 0 . 12 ) . However , the percentage of time spent in the dark sector was significantly reduced for T . rangeli-infected insects compared to uninfected controls ( Fig 3; Z = 6 . 83 , p = 0 . 01 ) . In addition to evaluating the effect of trypanosome infection on triatomine phototaxis we tested whether locomotion could also be affected . R . prolixus exhibit a bimodal pattern of daily locomotion activity and we used an actometer to evaluate if infection by T . cruzi or T . rangeli could interfere with this aspect of triatomine behavior . The movements of uninfected and infected insects were monitored in order to detect variations in their locomotory activity . Uninfected and T . cruzi-infected insects showed a similar pattern of locomotion consisting of two main peaks: one at the second hour of the scotophase and the other , during the first hour of the photophase . During the remaining time , especially during the photophase , insects exhibited almost no locomotion . However , we found a statistically significant interaction between infection status and hour of the daily cycle in insects infected by T . cruzi ( Fig 4; F = 4 . 02 , p<0 . 0001 ) . Pairwise contrasts revealed that T . cruzi infected insects showed significantly decreased locomotory activity during the second hour of the scotophase ( 20:00–21:00h; S1 Table ) . In overall , the total number of movements recorded for T . cruzi-infected insects was about 20% less than that observed in uninfected control insects . When T . rangeli-infected insects were analyzed , a similar general pattern of daily activity with two peaks was observed . As seen for the T . cruzi experiment , infection by T . rangeli also promoted alterations in R . prolixus motility at a particular time of the day ( Fig 5; F = 8 . 70 , p<0 . 0001 ) . However , differently from the former results , T . rangeli-infected nymphs showed an increase in their general activity that was significant during the 11th and 12th hours of the photophase ( 17:00–18:00 , 18:00–19:00h; S2 Table ) when compared to uninfected controls . The total number of movements displayed by T . rangeli-infected insects was on average 23% higher than that shown by uninfected control insects . Many aspects of vector behavior are controlled by complex and yet unknown genetic interactions and signaling pathways . Since parasites can interfere with gene expression from their invertebrate vectors , we speculated whether the levels of the foraging gene ( shown to modulate locomotion in other species ) would be altered in trypanosome-infected R . prolixus . We identified an ortholog of the foraging gene ( Rpfor ) in the supercontig GL553754 , between the positions 4 , 659 and 26 , 665 on the forward strand , in the R . prolixus genome database . The predicted protein for Rpfor gene ( code: RPRC000321-PA ) contains 551 amino acids ( 1 , 656 bp ) is coded by fourteen exons and is similar in length to those from T . castaneum and P . humanus with 535 and 542 amino acids , respectively . Furthermore , the presence of functional domains characteristic of this gene was confirmed in the Rpfor protein sequence ( Fig 2 ) . Using qPCR analyses we quantified the relative expression of Rpfor in uninfected and trypanosome-infected R . prolixus . We found no difference in Rpfor expression levels between T . cruzi infected and uninfected insects ( Fig 6A ) . In contrast , however , when the insects were infected with T . rangeli , there was an important decrease in Rpfor expression when compared with uninfected controls ( Fig 6B ) . We observed a decrease of 4 . 8 fold in the relative expression levels of Rpfor in T . rangeli infected insects .
Triatomines are nocturnal insects and remain inside protected shelters during daylight hours [17] . Such behaviors are highly adaptive as they allow triatomines to decrease their exposure to predators , which eventually become hosts from whom these insects obtain their blood meals . In a natural context , a higher exposure during daylight hours could potentially increase insect mortality by predation . Uninfected and T . cruzi-infected nymphs of R . prolixus showed a marked negative phototaxis , spending more than 70% of the time in the dark sector of the arena . Interestingly , T . rangeli-infected triatomines seem to be less averse to light , and spent a significantly shorter proportion of time in the dark . Therefore , the decreased negative phototaxis of T . rangeli-infected insects shown in the present work could be an indication of a behavioral alteration potentially costly to the insects . Independent of infection status , all R . prolixus nymphs presented a bimodal pattern of locomotory activity with pronounced peaks at the start of the scotophase and photophase , in agreement with patterns described for triatomines [15 , 16] . Interestingly , T . cruzi-infected nymphs showed decreased spontaneous locomotory activity during the first half of the scotophase . This interval represents the period in which triatomines search for food and sexual partners [15 , 16 , 47] . Altered host activity is a common effect of parasitism [48] . In some cases , a declined activity may be caused by tissue destruction as seen in A . aegypti infected by Brugia pahangi [49] and in Gryllus integer infected by Ormia ochracea [50] . However , T . cruzi does not invade the celomic cavity of its vectors and is restricted to the intestinal tract . Triatomine tissue injuries have not been reported during T . cruzi infection [51] , indicating that a decrease of activity as a consequence of direct damage is unlikely . Different effects on triatomine fitness as a result of T . cruzi infection have previously been described , including resistance to starvation [8] , delayed molt and increased mortality [52] . Recently , our laboratory team showed that T . cruzi can be pathogenic to its vectors depending on the environmental temperature and insect nutritional status [10 , 30] . Taken together , these data suggest that T . cruzi competes with its vector for nutrients , since starved infected insects have reduced survival and increased susceptibility to other stress factors [53] . Furthermore , T . cruzi-infected T . infestans need more blood for molting than uninfected controls , probably as a compensation for the nutrients lost to trypanosomes [7] . In addition , a reduction in gonad weight as a consequence of nutrition curtailment was observed in the triatomine Mepraia spinolai infected with T . cruzi [54] . As well , infected M . spinolai find hosts almost twice as fast as uninfected bugs [9] . In our experimental design , insect activity was evaluated in the absence of host cues . Therefore , it is possible that the decreased locomotory activity observed in T . cruzi-infected insects could represent an energy saving mechanism designed to avoid the loss of already lowered nutritional resources in the absence of host cues . This lack of activity when host stimuli are absent and no indication that a potential blood meal is available might be intended to preserve insect fitness . A similar protective behavior has been suggested for C . lectularius in the absence of host cues [20] . Contrarily to what was observed with T . cruzi infected bugs , infection of R . prolixus with T . rangeli promoted instead , an increased locomotory activity during most of the daily cycle . Exceptionally , T . rangeli infected insects consistently showed a decreased locomotory activity at the onset of light . It is worth noting that the lights-on period normally induces a certain increase in locomotion in animals as a consequence of such abrupt change in light intensity , independently of having a circadian component superimposed . In the case of bugs infected with this parasite , their decreased negative phototaxis could at least partially explain this altered pattern in comparison to control insects . Nevertheless , the results for the remaining of the photophase together with the decrease in negative phototaxis shown in our experiment suggest that T . rangeli promotes an increased exposure of insects to predators . Whether this alteration increases parasite transmission still needs to be tested . An increase in the locomotory activity of infected vectors has also been reported in A . aegypti/Dengue 2 [55] and A . aegypti/Wolbachia pipientis [56] associations . In both cases , increases in pathogen transmission rates have been suggested . The origin and evolutionary distance of T . cruzi and T . rangeli is controversial , as a new perspective , the bat seeding hypothesis , proposes that T . cruzi evolved from bat trypanosomes [57] . According to this , T . cruzi would have evolved more recently than proposed in the prevailing Southern super-continent hypothesis [57] . In spite of this , T . cruzi and T . rangeli are classified unequivocally in the same clade [58 , 59] . In case of an ancient T . cruzi origin , the divergence time between T . cruzi and T . rangeli , would have occurred several millions of years ago [60] . This scenario would have possibly promoted distinct evolutionary associations with their insect vectors which concur with the obvious lifecycle and morphological differences between them . For T . rangeli , it has been accepted that the parasite presents a close evolutionary association with Rhodnius spp . [61] which has promoted the appearance of parasite strains closely associated to specific Rhodnius species [62] . Whether co-evolution enabled these parasites to manipulate vector behavior to increase their own transmission is an intriguing question for future studies . The molecular mechanisms underlying the modification of vector locomotion by parasites are still largely unknown , but likely require alterations in gene expression . Insect locomotory activity can be modulated regulating the expression of a gene coding for a PKG named foraging ( see rev . [63 , 64] ) . Therefore , we investigated whether Rpfor expression levels were altered in trypanosome-infected triatomines . Rpfor expression was differently affected in R . prolixus infected with T . cruzi and T . rangeli . While infection by T . cruzi promoted a trend for increased Rpfor mRNA abundance , T . rangeli infection , in contrast , significantly decreased Rpfor expression . The relation between locomotory activity , food search and the levels of foraging gene expression has been studied in different organisms ( see rev . in [63 , 64] ) . Foraging behavior is largely influenced by cGMP-activated protein kinase pathways across taxa , although the mechanisms involved still remain elusive . Two opposite models connecting insect locomotion/pattern of behavior and foraging gene expression have been described to date . In the first case , increased for expression has been related to increased locomotory activity in fruit flies [65] , bees [66] , bumblebees [67] and locusts [68] . In contrast , higher levels of for expression have been related to a decreased locomotory activity in ants [69] , wasps [70] and nematodes [71] . Interestingly , our data from T . rangeli-infected insects suggest a relation fitting the second model . These data strongly support the need for future studies evaluating Rpfor activity in triatomines under different physiological conditions , as well as gene expression manipulation or modulation of PKG activity . To our knowledge , this is the first demonstration that both , the expression of the for gene and the behavior of an insect host vectoring human disease have been shown to be altered by a parasite . It is relevant to consider that potential limitations may affect the conclusions of this study according to the methods used to test our hypotheses . One first case would be represented by the fact that we suggest that infection by T . rangeli induces a decrease in bug negative phototaxis . While we have followed the methodology used by Reisenman and colleagues [72] and Reisenman and Lazzari [32] because we consider those reports to be very relevant in relation to the existing literature about triatomine behavior , one may argue that these laboratory conditions can have restricted predictor power for natural environments . We highlight here that the size of an arena and the duration of an experiment only serve the purpose of evincing an effect hypothesized previously . If such effects can be put in evidence using those conditions , then they seem adequate . In fact , our results seem to indicate that our methods were adequate to prove that T . rangeli infection affects this parameter . The relative weight of such an effect in a natural scenario is not evident from them , and should be the focus of future studies . A similar case applies to the study of the locomotory activity of R . prolixus . Our arenas seemed effective for measuring bug locomotory activity and evince alterations . Furthermore , the scale of arena used was similar to those traditionally used for actometer studies with diverse insects [16 , 73 , 74] . Another issue that could potentially be raised would be that of the infection procedures used . In the case of T . rangeli , we have injected a small inoculum , based on our previous study evaluating the effects of different amounts of T . rangeli on R . prolixus nymphs [14] . For T . cruzi infections , we used a relatively large inoculum based on the fact that there is a relevant reduction in T . cruzi populations after the initial five days of infection [75] . It is worth mentioning that natural infections in sylvatic cycles probably include quite different conditions represented by diverse mammals experiencing acute or chronic infections . This includes opossums , which are known to show very high parasitemia [76] . It is difficult for us to determine an adequate standard for this , but we tend to assume that our conditions are quite reasonable . Finally , since our intent was to test whether parasites alter locomotory patterns in R . prolixus , the use of nymphs was planned to exclude the potential interference of sexual activities that could have interfered if adult bugs were used . Confirming an effect of trypanosome infection on 5th instar nymphs supports our claim . It is worth mentioning that a recently published report has shown that adult reproduction is compromised by infection with either of these parasites [10] , suggesting that adults may suffer similar consequences of trypanosome infection . This report proposes a new approach in the study of trypanosome-triatomine interactions , showing that these parasites alter bug locomotory activity and , in the case of T . rangeli , the phototatic behavior and the expression of a gene that has been shown to modulate insect behavior . Taken together , these alterations would possibly affect parasite transmission rates . Interestingly , it has recently been shown that T . cruzi induces alterations in the dispersion of Triatoma dimidiata females [77] , as well as in the wing size of adults of this species [78] . In addition , T . rangeli promotes longer flights in Rhodnius pallescens , possibly affecting its dispersion ability [79] . Altogether , these facts evidence that triatomine locomotion and trypanosome infection seem to be connected but the mechanisms through which these effects take place remain obscure . We suggest that functional genomics studies should enable a better understanding of the molecular mechanisms underlying the trypanosome induced alterations of triatomine behavior . | The control of Chagas disease , an infection that affects ca . 8 million people in Latin America , is mostly based on vector control activities . Understanding vector biology and how these insects interact with their environment , hosts and pathogens is crucial to improve vector control strategies . The behavior of triatomines has been largely studied , yet few reports have focused on the behavioral effects of the interaction that these insects endure with their natural parasites . Trypanosoma cruzi and Trypanosoma rangeli are two protozoan parasites found naturally infecting Rhodnius species . In this study , we showed for the first time that the locomotory activity of Rhodnius prolixus , a relevant vector of Chagas disease , is affected by trypanosome infection . T . cruzi was found to decrease bug locomotory activity during night hours , while T . rangeli promoted a generally increased insect locomotion . In addition , we searched for the R . prolixus orthologue ( Rpfor ) of a gene associated with the modulation of insect activity ( foraging gene ) and found that Rpfor expression was also affected by trypanosome infection . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Trypanosomes Modify the Behavior of Their Insect Hosts: Effects on Locomotion and on the Expression of a Related Gene |
Retroviruses and Long Terminal Repeat ( LTR ) -retrotransposons have distinct patterns of integration sites . The oncogenic potential of retrovirus-based vectors used in gene therapy is dependent on the selection of integration sites associated with promoters . The LTR-retrotransposon Tf1 of Schizosaccharomyces pombe is studied as a model for oncogenic retroviruses because it integrates into the promoters of stress response genes . Although integrases ( INs ) encoded by retroviruses and LTR-retrotransposons are responsible for catalyzing the insertion of cDNA into the host genome , it is thought that distinct host factors are required for the efficiency and specificity of integration . We tested this hypothesis with a genome-wide screen of host factors that promote Tf1 integration . By combining an assay for transposition with a genetic assay that measures cDNA recombination we could identify factors that contribute differentially to integration . We utilized this assay to test a collection of 3 , 004 S . pombe strains with single gene deletions . Using these screens and immunoblot measures of Tf1 proteins , we identified a total of 61 genes that promote integration . The candidate integration factors participate in a range of processes including nuclear transport , transcription , mRNA processing , vesicle transport , chromatin structure and DNA repair . Two candidates , Rhp18 and the NineTeen complex were tested in two-hybrid assays and were found to interact with Tf1 IN . Surprisingly , a number of pathways we identified were found previously to promote integration of the LTR-retrotransposons Ty1 and Ty3 in Saccharomyces cerevisiae , indicating the contribution of host factors to integration are common in distantly related organisms . The DNA repair factors are of particular interest because they may identify the pathways that repair the single stranded gaps flanking the sites of strand transfer following integration of LTR retroelements .
Retroviruses integrate their DNA sequence into the chromosomes of infected cells to achieve permanent and reliable replication . A substantial amount of biochemical and genetic information is known about the catalysis of integration and the host factors responsible for the virus specific positions of integration [1–3] . The bulk of information about the factors required for integration is derived from high throughput sequencing of insertion profiles . Specific patterns of integration such as the promoter sequences selected by gamma retroviruses or the actively transcribed genes selected by lenti-retroviruses , result from direct interactions between the viral integrase and chromosome bound host proteins [3] . These diverse patterns of integration suggest the host pathways that promote integration are virus specific . This understanding remains to be tested since genome-wide siRNA screens for host factors have only been performed for HIV-1 infection and the complexity of these results provided little consensus [4–7] . Long terminal repeat ( LTR ) retrotransposons are mobile elements that are the progenitors of retroviruses [8 , 9] and are studied extensively as important models for retrovirus replication [10–13] . LTR-retrotransposons model the same processes of particle formation , reverse transcription and integration that are central to retrovirus propagation . One advantage of LTR-retrotransposons is they are highly active in well-characterized model organisms , Saccharomyces cerevisiae and Schizosaccharomyces pombe . Extensive study of these model systems has resulted in significant understanding of particle formation , reverse transcription , and integration [10–13] . Genetic assays that measure retrotransposon mobility rely on single copy elements tagged with a drug resistance gene or on plasmids that express retrotransposon mRNA [14–17] . These assays were used with collections of deletion strains or insertion mutants to identify host factors important for transposition of Ty1 , and Ty3 in S . cerevisiae [18–22] and extensive screens were performed to identify host factors that restrict transposition in S . cerevisiae [20 , 23 , 24] . Host factors important for transposition are involved in chromatin modification , transcription , translation , vesicle trafficking , nuclear transport , and DNA repair . These genetic screens provide a broad view of what cellular systems support transposition in S . cerevisiae . However , it is not known how general these processes are in supporting transposition in other eukaryotes . More importantly , none of these screens were designed to identify host factors that promote integration . S . pombe is distantly related to S . cerevisiae having diverged approximately 350 million years ago [25–27] . The identification of host factors in S . pombe important for retrotransposition would provide a valuable means for determining whether the cellular processes that support retrotransposition are conserved between distantly related eukaryotes . A significant body of research on the LTR-retrotransposon Tf1 of S . pombe describing protein expression , particle assembly , reverse transcription , and transposition activity has established Tf1 as a valuable model system [12] . The transposition of Tf1 in S . pombe is measured by expressing a drug resistant copy of Tf1 from a multi-copy plasmid [14 , 28 , 29] . This genetic assay combined with high throughput sequencing shows that Tf1 has a pronounced pattern of integration that favors the promoters of stress response genes [30 , 31] . Recent studies revealed that the DNA binding protein Sap1 plays an important role in directing integration to stress response promoters [32 , 33] . Although two-hybrid assays detected interaction between Sap1 and IN , biochemical and immunoprecipitation experiments fail to detect this interaction [32 , 33] . We therefore believe other factors necessary for integration bridge the Sap1-IN interaction . To identify potential bridging proteins we applied a genome-wide screen for factors involved in integration . For this , we applied a unique combination of assays that together detect defects in integration . We identified a set of 61 host factors that promote integration relative to recombination and participate in key cellular processes such as transcription , chromatin structure , mRNA processing , translation , vesicle trafficking , and DNA repair . With these results we discovered there is a surprising diversity in processes involved in integration . Although it’s not clear with this type of genetic screen which factors impact integration directly , we found strong similarity in the host factors that promote integration in distantly related eukaryotes .
The haploid deletion strains of the Bioneer 2 . 0 collection were individually transformed to introduce the Tf1-natAI plasmid ( Fig 2A ) . Fifty strains did not grow on plates lacking uracil , which was used to select for uptake of the plasmid ( Fig 2B , S1 Table ) . While some of these strains contained deletions in uracil catabolism genes other deletion strains had very slow growth rates , might be unable to tolerate the lithium treatment of transformation , or might be incapable of transferring the plasmid DNA into the nucleus . Despite the strains that were poor growers or transformation defective , the expression plasmid was successfully introduced into 2 , 954 deletion strains . For each of these strains , four independent isolates containing the plasmid were assayed for transposition and recombination activities as diagramed in Fig 2A and listed in S1 Table . All four independent isolates of each deletion strain were scored for transposition activity on a scale of 0 to 5 by comparing growth to wild-type strains of S . pombe which received a score of 5 ( S2 Fig and S1 Table ) . Ten deletion strains were unable to grow on FOA-containing medium and therefore could not be scored ( Fig 2B and S1 Table ) . Additional strains that could not be scored include 30 deletions that had poor viability , and eight strains with unidentified genetic defects ( S1 Table ) . A total of 150 deletion strains had a significant defect in transposition frequency and were scored 2 . 5 or lower ( Fig 3 and S1 Table ) . For the recombination assays the patches were also scored on a scale of 0 to 5 where wild-type was assigned the score of 5 ( S3 Fig ) . A total of 183 deletion strains exhibited a notable defect in homologous recombination and were scored lower than 4 ( S1 Table ) . Of the 150 deletion strains with low levels of transposition 41 also exhibited recombination activity lower than 4 indicating these genes were important for intermediate stages of transposition such as particle assembly , reverse transcription or nuclear import ( S2 Table ) . For example , deletion of nup124 resulted in low recombination and transposition , a result previously described in studies that found Nup124 interacts with Gag and promotes nuclear import of Tf1 protein and cDNA [41–46] . Importantly , we identified 109 deletion strains that had strong homologous recombination scores ( 4 or higher ) but had significantly reduced transposition activities scoring 2 . 5 or less ( Fig 3 , S1 Table ) . These strains represented our initial list of candidates that could be important for integration ( S3 Table ) . One concern with our list of integration deficient candidates was that the homologous recombination assay relied on the growth of cells in patches and reductions of two to four-fold in the growth of a patch is not reliably detected . To test whether integration deficient candidates had reductions in recombination not observed with patches we screened the integration candidates with a quantitative assay that measures the fraction of cells in liquid cultures that have recombination events ( S4 Fig ) [29] . Each deletion strain was assayed in triplicate and each replica was an independent plasmid containing isolate ( S3 Table ) . The results of this assay were highly reproducible . Although the homologous recombination activity of Tf1 looks to be independent of integration as Tf1 lacking IN ( Tf1 INfs ) has approximately the same amount of activity as Tf1 with IN ( Fig 1 ) , quantitative measures show that approximately 50% of the recombination response is IN dependent [29 , 36] . The quantitative recombination assays reported here confirmed this finding that with Tf1-natAI the INfs reduced recombination activity to 45% ( SD 4 . 0% ) of wild-type Tf1-natAI ( Fig 4A and S3 Table ) . Therefore , deletion strains with reduced integration but intact homologous recombination would be expected to exhibit the same recombination levels as the INfs , 45% . Surprisingly , 91 of the integration deficient candidates possessed recombination activity greater than exhibited by the INfs ( Fig 4B and S3 Table ) . This high number of deletions that had more recombination than the INfs suggests that in the absence of integration the IN protein might promote homologous recombination of the cDNA . This was tested by measuring recombination frequencies of deletions in rhp18 and pht1 with the INfs . In addition to showing high levels of cDNA recombination , these two genes were selected because of interesting potential roles in Tf1 integration as described below . While Tf1-natAI expressed in these strains produced recombination activities higher than the INfs in wild-type cells , expression of INfs in the deletion strains resulted in reduction in recombination activity to levels similar in wild-type S . pombe containing INfs ( Fig 4A ) . These results indicate that the presence of IN stimulates cDNA recombination independent of IN catalysis . Wild-type strains with single amino acid substitutions in catalytic residues of IN had recombination activities averaging 67% of the strain with intact IN ( Fig 4C ) . Since the catalytic mutations disrupted integration without reducing levels of IN protein ( S5 and S6 Figs ) , the 67% recombination vs . the 45% of the INfs indicates that IN lacking catalytic activity does stimulate homologous recombination . It is possible this occurs because IN protects cDNA from degradation . In considering these IN contributions to recombination we chose the average activity of the catalytically inactive IN mutants , 67% ( Fig 4C ) as the level for recombination activity expected in the absence of integration . Deletion strains with recombination levels below 60% were deemed to have a defect in Tf1 recombination . By this criterion 8 of the integration deficient candidates had reduced homologous recombination ( S3 Table ) and were excluded from the final list of candidates ( Table 1 ) . Another question when validating candidate strains was whether the deletion mutations reduced the levels of Tf1 proteins . We addressed this possibility for the primary set of 101 integration deficient candidates by performing quantitative immunoblotting of whole cell extracts ( Materials and methods ) ( Fig 3 and S3 Table ) [47] . Candidate strains with reduced Gag or IN levels by two-fold or greater were considered to have poor expression of Tf1 protein and as a result these factors were removed from the list of candidates that mediate integration . Results of these quantitative immunoblots identified 64 integration candidates that expressed normal IN and Gag levels ( S3 Table ) . The majority of the candidates identified as having reduced homologous recombination also had low Tf1 protein expression . However , three deletion strains had normal levels of Gag and IN but had reduced homologous recombination , as measured with the quantitative assay . As a result our final list of candidates that impact integration had 61 factors ( Table 1 ) . To understand how the candidate factors in our final list may contribute to integration , we grouped them by biological function encompassing two broad categories; non-chromatin associated and chromatin associated processes ( Table 1 ) [48 , 49] . Factors that lack association with chromatin are less likely to participate directly in integration . Nevertheless , 40 of the 61 candidates with defects in integration have no established association with chromatin ( Table 1 ) . These non-chromatin associated proteins function in nuclear transport , protein synthesis , mRNA processing , vesicle transport , ubiquitination , signal transduction , metabolic processes , chromosome segregation , and cytoskeleton structure . Among the list of non-chromatin associated factors that promote Tf1 integration are three nuclear transport proteins , Nup61 , Rsm1 , and Syo2 . Previous studies found several nuclear pore factors contribute to LTR-retrotransposon activity by mediating transport of transposon factors and cDNA into the nucleus ( Table 2 ) [18–20 , 22 , 46] . Nuclear pore factors also mediate the replication of retroviruses by mediating nuclear entry of the IN complexes [50–54] . However , these functions would not be expected to mediate integration as nuclear pores are imbedded in the nuclear envelope . The contribution of Nup61 to Tf1 integration could be indirect by transporting other factors that mediate integration . However , there is evidence that some nuclear pore complexes can interact directly with chromatin [55–57] and in the case of HIV-1 , integration appears to favor the nuclear periphery [53] . It is notable that Nup61 is a homolog of Nup2 , which in S . cerevisiae binds to a set of promoters and activates gene expression [58–60] . Other studies found that tRNA genes localize at nuclear pore complexes of S . cerevisiae via an interaction between DNA sequence and Nup2 [61] . These results suggest the possibility that Nup61 could bind promoters in S . pombe and directly stimulate Tf1 integration . This possibility also exists for nuclear pore factors that promote Ty1 and Ty3 transposition in S . cerevisiae . Deletion of five different nuclear pore factors inhibits Ty1 transposition and deletion of NUP59 results in reduced Ty3 transposition . In all these cases cDNA production is not reduced suggesting the Nups may contribute to integration in S . cerevisiae ( Table 2 ) . A collection of 10 factors that promote Tf1 integration are associated with protein synthesis and mRNA processing . While there is no unified understanding of how translation might contribute directly to integration , Risler et al . found 33 and Griffith et al . identified nine components of ribosomes or translation factors that promote Ty1 transposition [18 , 22] . At least four translation factors and two translation inhibitors are involved in Ty3 transposition [11 , 20 , 21] . Many of the ribosome constituents and ribosome biogenesis factors that promote Ty1 and Ty3 transposition are important for translation of transposon mRNA [18 , 20 , 22 , 62] . However , among the factors important for transposition , 20 ribosomal proteins and translation factors required for Ty1 transposition and four factors important for Ty3 transposition mediate a stage of the transposition process after reverse transcription either related to nuclear import or integration ( Table 2 ) . Factors that supported Tf1 integration include Pdc2 , and Ski3 , proteins involved with deadenylation , decapping , or 3’ end mRNA degradation . Although the function of these proteins suggests they would influence Tf1 mRNA translation , the results of the homologous recombination assay indicate these mRNA stability factors impact integration . While there are no obvious means for mRNA stability to have a direct contribution to integration , factors mediating deadenylation , decapping , and 3’ mRNA decay also contribute to Ty1 ( Ccr4 , Lsm1 , Lsm6 , Ski8 , Rpb4 , Trf5 , and Mpp6 ) [18] at stages after cDNA synthesis . The similarity in these factors that contribute to late stages of Tf1 , Ty1 and Ty3 transposition suggests these distantly related LTR-retrotransposons may share aspects of integration that are regulated by mRNA processing and translation . A significant cluster of host factors involved in vesicle transport was found to contribute to Tf1 integration ( Table 2 ) . This group included factors responsible for ER maintenance , ER to Golgi transport , transport within the Golgi , and two components of the ESCRT III complex associated with sorting of cargo proteins . As explained for nuclear transport , protein synthesis , and mRNA decay , vesicle transport is a process not known to be directly involved in retrotransposon integration . However , vesicle traffic and membranes are critical for the replication of many viruses . Examples include gamma and Type-D retroviruses , both of which require the endosomal system to traffic Gag and Gag-Pol to the plasma membrane [63] . This contribution to the replication of retroviruses occurs much earlier in the lifecycle than integration , the stage of Tf1 activity that requires vesicle transport . Nine ( 15% ) of the candidate integration factors identified in our screen are associated with vesicle transport . Several vesicle formation , cargo loading , and vesicle transport factors are involved in Ty1 retrotransposition including functions that occur after cDNA is synthesized ( Table 2 ) . Interestingly , a particularly large set of vesicle trafficking factors contribute to Ty3 retrotransposition post reverse transcription [20] . These include several components of ESCRT complexes I , II , and III ( Snf7 , Vps4 , Vph1 , Vps20 , Bro1 , Vps28 , Snf8 , Vps36 , Clc1 , Fab1 , and Vma7 ) . The high numbers of vesicle trafficking factors involved in Ty1 and Ty3 transposition indicate late stages of LTR-retrotransposition require highly conserved features of vesicle trafficking . Four Tf1 candidate integration factors function in ubiquitination , deubiquitination , or assembly of the proteasome ( Table 1 ) . These factors contribute to the degradation of a wide range of proteins , any number of which could be important for integration . It is therefore difficult to propose specific functions of these candidates that promote integration . With such broad impact on cellular systems , it’s not surprising that ubiquitin modifications and the proteasome factors promote activities of Ty1 and Ty3 [18–21] . A set of eight metabolic enzymes was identified in the list of candidate integration factors ( Table 1 ) . They are mostly unrelated making it difficult to identify a specific pathway that might mediate integration . The exception is that the metabolic factors included both enzymes responsible for synthesis of trehalose , a disaccharide that mitigates the impact of heat and oxidative stress [64–66] . What is more intriguing is that one of these enzymes trehalose-phosphate synthase ( Tps2 ) is important for Ty1 transposition [22] . Although with Ty1 , Tps2 is required for an early step in the transposition cycle that is necessary for cDNA production . Candidate integration factors that are chromatin associated included the histone variant H2A . Z ( Table 1 ) . H2A . Z is concentrated in the +1 and -1 nucleosomes that flank the nucleosome depleted region of promoters [67 , 68] . Genome-wide profiles of 1 . 6 million insertions show that Tf1 targets the nucleosome-depleted region of promoters in a window of 150 bp immediately adjacent to the -1 and +1 nucleosomes [32] . With this pattern of integration , it is feasible that H2A . Z participates in integration via a direct interaction with IN . Alternatively , it is possible that H2A . Z recruits the binding of a targeting factor or contributes to a form of chromatin structure that facilitates efficient integration . Nucleosomes are determinants of integration for retroviruses due to structural perturbations of the DNA [69–71] . Interestingly , H2A . Z and the remodeling factor that assembles H2A . Z in nucleosomes , Swr1 , are important for Ty1 transposition [18 , 19] . Importantly , H2A . Z and Swr1 may function directly in Ty1 integration since deletion of swr1 does not reduce Ty1 cDNA and H2A . Z associates with RNA pol III promoters [18 , 72 , 73] . A role of H2A . Z in integration is consistent with the strong association observed between H2A . Z and sites of Ty1 integration [74] . For a factor to promote the integration of two highly divergent LTR-retrotransposons such as Tf1 and Ty1 suggests that H2A . Z contributes to a feature of chromatin structure that is important for the integration of a broad range of LTR-retrotransposons . Candidate integration factors associated with chromatin included components of histone modifying complexes ( Set1 , Nts1 , and Ubp8 ) and Snf5 , a subunit of the SWI/SNF chromatin-remodeling complex ( Table 1 ) . A number of factors with similar functions contribute to Ty1 and Ty3 transposition , possibly in integration ( Table 2 ) . Set1 is the histone H3 lysine 4 methylase component of the COMPASS complex . A different component of this complex , Swd1 , contributes to Ty3 transposition post-reverse transcription [21] . Nts1 is a component of the histone H3 deacetylase complex Clr6 and Ubp8 is a subunit of the SAGA histone acetylation complex . A number of factors controlling histone acetylation promote transposition in S . cerevisiae ( Table 2 ) . The SWI/SNF complex has global impact on gene regulation including Ty1 transcription [75] . As a result , Snf5 and other components of SWI/SNF contribute to Ty1 transposition . Although chromatin modifications and remodeling have broad effects on expression of the genome , the similarities in the chromatin complexes that promote transposition of Tf1 , Ty1 , and Ty3 suggest certain features of chromatin structure may play a common role in integration of LTR-retrotransposons . Another class of candidate integration factors we identified is associated with transcription ( Table 1 ) . Ckb1 is a regulatory subunit of casein kinase 2 and Lkh1 is a kinase . Both factors mediate the phosphorylation of a broad range of substrates including transcription factors , and subunits of RNA polymerases [76–79] . Srb11 is a cyclin-like component of RNA polymerase II involved in phosphorylation of the RNA polymerase II C-terminal domain [80 , 81] . Any of these kinase functions have the potential to modulate a protein important for integration . Paf1 , and Tfs1 associate directly with RNA pol II and have the potential to target integration directly . Interestingly , deletion of paf1 abolishes the methylation of histone H3K4 . Paf1 controls H3K4 methylation by promoting ubiquitylation of histone H2B , which is required to recruit Set1 , [82] a factor our screen identified . This connection suggests that the role of Paf1 in integration is to promote H3K4 methylation . Interestingly , the Paf complex and rad6 inhibit integration of Ty1 and prevent disruption of ORFs [23 , 74 , 83–85] . Genome-wide integration of Ty1 upstream of pol III genes does not change in a rad6 deletion [74] . Proposed models for these observations suggest Paf and Rad6 strengthen target specificity and restrict integration . These effects are mechanistically distinct from the contribution Set1 and Paf1 make to Tf1 integration . Two core splicing factors , Cwf12 , and Smd3 , and the splicing coactivator Pwi1 were identified in our screen as candidate integration factors . Although it’s possible that splicing factors were identified because the transposition assay relies on splicing of the artificial intron , this is unlikely because the intron must also be spliced for cDNA recombination to be detected . Cwf12 is a member of the NineTeen Complex that plays a central role in splicing by tethering the U6 snRNA to the activated spliceosome [86–89] . Smd3 is one of seven Sm proteins that are common components of the U1 , U2 , U4 , and U5 snRNPs [86] . It is not clear whether these splicing components directly contribute to integration as it is possible their absence changed expression of proteins that mediate integration . However , several core splicing complexes including the NineTeen Complex contribute to stages of Ty1 transposition after reverse transcription and the splicing regulator Sqs1 promotes stages of Ty1 and Ty3 transposition post reverse transcription ( Table 2 ) . While there is no information about how splicing could contribute to integration in yeast , recent studies of HIV-1 found that the host factor LEDGF/p75 interacts with splicing factors and targets integration to highly spliced genes [90 , 91] . Our genetic screen found deletion of genes encoding four DNA repair factors , Rhp18 , Rhp23 , Rad50 , and Rad51 resulted in significant reductions in transposition without lowering homologous recombination or expression of Gag and IN ( Table 1 ) . If these factors mediate integration it is possible they function with the targeting factor Sap1 which can be a replication fork barrier [92] . Rad50 and Rad51 mediate homologous recombination and this activity can contribute to DNA replication by assisting recovery of arrested replication forks [93] . As a result , it’s possible that Rad50 and Rad51 interact with Sap1 at arrested forks in a configuration that stimulates integration . This is consistent with the model that Sap1 induces Tf1 integration at stalled forks [33] . However , the functions of the DNA repair factors in Table 1 are broad suggesting the intriguing possibility that these factors are responsible for repairing the unattached 5’ ends of the integrated cDNAs . The integrases of LTR-retrotransposons and retroviruses catalyze DNA-strand transfer reactions where the 3’ ends of the cDNAs attack staggered phosphodiester bonds on opposite strands of the target DNA [94 , 95] . The inserts are flanked by single stranded gaps with 5’ ends of the cDNA unattached to the target site . These gaps must be repaired and this process is of great interest as it is unknown which factors are responsible for integration repair of any LTR-retrotransposon or retrovirus . Deletion strains unable to repair the single stranded DNA gaps would have reduced transposition activity but potentially maintain normal frequencies of homologous recombination . Rhp18 is the S . pombe homolog of Rad18 , an E3 ubiquitin ligase that binds single stranded DNA and functions both in postreplication repair and in translesion synthesis [96–100] . Additional evidence indicates that Rad18 in mammalian cells mediates homologous recombination repair of double-strand breaks [101] . Rad18 localizes to double-strand breaks and facilitates homologous recombination by interacting directly with Rad51 , a RecA family recombinase . Rad51 was also identified as a candidate integration factor suggesting that Rad18 and Rad51 could function together in homologous recombination to repair integration sites ( Table 1 ) . If replication occurs before the single strand gaps are repaired , then the resulting double strand breaks could be repaired by homologous recombination . Rad50 , another DNA repair factor that promoted Tf1 integration ( Table 1 ) , is a subunit of Mre11-Rad50-Xrs2 MRX complex responsible for resection of double-strand breaks [102] . This function not only contributes to homologous recombination but is also thought to be important for processing unusual DNA structures . One possibility is that MRX is important for repairing integration sites because it displaces IN . Studies of Mu phage show that the transpososome adheres tightly to integration sites and is removed by the ClpX protease [103 , 104] . Rhp23 , the S . pombe homolog of Rad23 is another DNA repair factor found to promote integration ( Table 1 ) . Rhp23 is a subunit of Nuclear Excision Repair Factor 2 with Rad4p that binds damaged DNA and excises fragments of 24 to 27 nucleotides [105] . One other candidate integration factor with the potential to repair DNA is Mhf2 , discussed above as a component of kinetochores ( Table 1 ) . Mhf2 is a component of the MHF histone-fold complex that in human cells interacts with both DNA and the Fanconia anemia associated factor FANCM to repair damaged DNA and stabilize replication forks stalled by DNA interstrand crosslinks [106] . This function may participate with Rhp18 in conducting translesion synthesis . In all , five candidate integration factors identified with our screen , Rhp18 , Rhp23 , Rad50 , Rad51 , and Mhf2 have DNA repair activity and therefore have the potential to repair integration sites . They participate in translesion synthesis ( Rhp18 and Mhf2 ) , double strand break repair ( Rhp18 , Rad51 , and Rad50 ) , and nuclear excision repair ( Rhp23 ) . It is possible these factors function in concurrent repair processes that serve redundant functions . It is also possible that there are other factors important for repairing integration sites that were not identified by our screen because they contribute to homologous recombination of cDNA . These would be factors such as Rad52 that are important for both homologous recombination and transposition ( S2 Table ) . It is significant that similar DNA repair factors are involved in retrotransposition in S . cerevisiae ( Table 2 ) . In particular a subunit of the MRX complex ( Xrs2 ) and a Rad51 mediator ( Rad52 ) contribute to Ty1 transposition [22] . Consistent with a role in integration site repair , the contribution of these DNA repair factors occurs after cDNA synthesis . Several other studies independently found members of the MRX complex and the Rad51-Rad52 recombination pathway are involved in Ty1 transposition [23 , 24 , 107–110] . However in these studies the DNA repair factors inhibit transposition as measured with a single copy Ty1 carrying the his3AI reporter . Amounts of Ty1 cDNA produced by single copy Ty1 increase in the absence of Rad51-52 factors . The dramatic increase in cDNA in these assays is triggered by DNA damage and requires S-phase checkpoint factors [111] . It is not clear why single copy Ty1 with the his3AI reporter produces such differences from Ty1 and Tf1 expressed from a plasmid . Nevertheless , the overlap of DNA repair factors that can promote Ty1 and Tf1 transposition argues these factors may mediate a conserved feature of integration . In a previously published study designed to identify factors that repair DNA at HIV-1 integration sites , 232 genes associated with DNA repair were tested with RNAi methods [112] . A cluster of six genes involved in short patch base excision repair were identified that when deleted in mouse embryo fibroblasts resulted in decreased HIV-1 replication . The proteins identified included damage recognition glycosylases ( OGG1 and MYH ) and the late repair factor POLβ . Consistent with a role in integration site repair these proteins promote late steps in replication that occur after reverse transcription and nuclear entry [113] . While these proteins as well as the candidate integration factors we identified may be involved in the repair of integration sites , further studies are needed that can directly test this model . The overlap of factors and pathways that promote late stages of retrotransposition in S . cerevisiae and S . pombe suggest these represent cellular processes that are fundamental to delivery of cDNA or integration . The list of genes identified in screens of S . cerevisiae and S . pombe in Table 2 shows a number of overlapping pathways but it is not a formal test that can be evaluated statistically . To address this , we assembled lists of genes important for Ty1 and Ty3 transposition that when mutated do not result in significant reduction of cDNA . The genes of S . cerevisiae along with those we identified from S . pombe were grouped by gene ontology using Fission Yeast gene ontology slim terms ( S4 Table ) . We calculated the enrichment of these genes in each slim term relative to the total number of non-essential genes in the slim term that are included in the deletion sets ( S7 Fig ) . Although the overall number of genes identified by these genetic screens are relatively low to calculate enrichment values for non-essential genes , several had enrichments with p values <0 . 05 ( S7 Fig ) . The slim terms that showed statistically significant enrichment for genes important for late stages of transposition in S . pombe and S . cerevisiae were RNA metabolic processes and protein catabolic processes . While not reaching p values <0 . 05 , other terms showed enrichment near two-fold for late stage transposition genes of S . pombe and S . cerevisiae such as cell adhesion , chromatin organization , nucleocytoplasmic transport , regulation of transcription , DNA repair , and protein targeting . While the slim terms are broader than what we described in Table 2 , they do reflect the overlap between functions implicated in late stages of transposition in both S . pombe and S . cerevisiae . The 61 factors that promote integration participate in a wide range of cellular processes . We sought additional evidence about whether these processes are directly involved in integration by testing a representative set of candidate integration factors for contributions to cDNA levels and insertion site distribution . We evaluated strains lacking DNA repair factors ( Rad50 and Rad51 ) , chromatin factors ( Pht1 and Set1 ) , the chromatin remodeler Snf5 , the nuclear pore protein Nup61 , and the splicing factor Cwf12 . Although our recombination assays indicated cells lacking these candidates had wild-type levels of cDNA in the nucleus ( S3 Table ) , it was possible that incomplete cDNAs or intermediates were responsible for the recombination . We used a DNA blot to detect altered structure and accumulation of the Tf1-natAI cDNA . The cDNA produced from the plasmid expressed Tf1-natAI was digested with BsrGI and quantified on DNA blots ( S8 Fig ) . The 2 . 9 kb band detected with a probe for natAI is produced by BsrGI cleavage of the 3’ section of the cDNA . This terminal double stranded portion of cDNA is synthesized only after minus and plus strand transfers and as a result is a measure of mature Tf1-natAI cDNA . The intensities of the cDNA bands were quantified and normalized relative to the amount of expression plasmid in each strain . No reduction in cDNA was observed in cells lacking Rad50 , Rad51 , Pht1 , Set1 , Snf5 , Nup61 , or Cwf12 . Interestingly , cDNA was elevated in cells lacking Rad51 and was modestly increased in the absence of Pht1 , Snf5 , and Nup61 . We determined whether these seven representative candidates contributed to integration site distribution by high throughput sequencing inserts produced by plasmid-derived expression of Tf1-natAI ( Materials and methods ) . We used the Illumina platform and sequenced ligation-mediated PCR libraries of integration sites ( S5 Table ) [30–32] . We quantified integration in ORFs divided into 15 equal segments . For insertions upstream and downstream of ORFs we summed them in 100 bp windows ( Fig 5 ) . As observed in previous studies , integration clustered upstream of ORFs ( Fig 5A ) [30–32 , 114 , 115] . Although all the deletion strains tested exhibited this clustering upstream of ORFs , deletion of nup61 resulted in a modest increase of integration within ORFs ( Fig 5B ) . When integration sites are selected at random in the Matched Random Control ( MRC ) 58 . 55% occurred within ORFs ( Fig 5I ) . Reproducible measures show integration levels in intergenic sequences vary over a wide range with the bulk of insertions occurring in 1 , 000 of the 5 , 000 intergenic regions in the genome [30–32] . We asked whether this subgroup of candidate integration factors contribute to the distribution of integration among intergenic regions . When comparing amounts of integration in intergenic regions , strains lacking the candidate integration factors had strong correlations with the wild-type strain ( Fig 6 ) . These correlations were comparable to what we observed between two independent experiments with integration sites produced by wild-type cells ( Fig 6A ) . This indicates that these factors did not significantly contribute to the targeting of integration in intergenic regions . Our study of Tf1 integration showed that insertions cluster adjacent to positions of Sap1 binding at bases -9 and +19 relative to the motif recognized by Sap1 [32] . We asked whether the residual integration in the representative set of deletion mutations occurred at the -9 and +19 positions relative to the 5 , 000 best matches to the Sap1 motif . The integration pattern relative to the Sap1 motif was largely unchanged in the deletions ( Fig 7 ) . These patterns suggest the residual integration in the deletion mutations retains its dependence on Sap1 . Candidate integration factors that contribute directly to integration may interact physically with IN . In a report to be published separately , we applied the two-hybrid system of S . cerevisiae to identify host factors that interact with Tf1 IN . We found that the DNA repair factor Rhp18 reproducibly interacts with IN ( Fig 8 ) . Rhp18 was one of the candidate integration factors ( Table 1 ) , indicating that our screen was able to identify factors directly involved in integration . The interaction of a DNA repair factor with IN is intriguing and suggests the possibility that Tf1 recruits repair factors to integration sites to facilitate repair . This IN mediated recruitment may be a conserved function of integration since the human homolog of Rhp18 , hRad18 interacts and co-localizes with HIV-1 IN in HEK293T cells [116] . Our two-hybrid survey also identified an interaction between the Cwf3 component of the NineTeen splicing complex and Tf1 IN ( Fig 8 ) . One of the candidate integration factors , Cwf12 , is also a member of the NineTeen complex indicating that the NineTeen complex is directly involved in integration [88 , 89] . A role of splicing has been observed for HIV-1 where integration is directed to genes that are highly spliced [90] . Perhaps the NineTeen complex plays a similar role in S . pombe by recruiting Tf1 IN to sites of integration . The two-hybrid interactions described above resulted from a screen of a cDNA library . Since such screens are not exhaustive it is possible and even likely that IN interacts directly with other candidate integration factors listed in Table 1 . Our screen of 3 , 004 non-essential genes represents the first comprehensive study of host factors in S . pombe that promote retrotransposition . With our combination of genetic assays we were able to identify factors that may contribute directly to integration . However , other experiments are needed to evaluate the candidates for a direct role in integration . In addition to promoting integration some of our candidates could mediate a different step late in transposition such as the localization of cDNA in a nuclear compartment . Our data makes it possible to compare the candidate integration factors we identified in S . pombe to the factors of S . cerevisiae that through a number of studies are likely to promote integration . Factors we identified function in nuclear transport , protein synthesis , mRNA processing , vesicle transport , chromatin structure , transcription , spicing , and DNA repair . Although this wide range of host factors suggests many could make indirect contributions to integration , we found a surprising overlap with pathways and factors important for integration of Ty1 and Ty3 in S . cerevisiae ( Table 2 ) . These overlaps support the model that many of these processes contribute directly to integration . The extent of overlap is significant because of the great evolutionary distance between these yeasts and because Ty1 belongs to the copia family , a distinct superfamily of LTR-retrotransposons from gypsy , the family that includes Tf1 and Ty3 . The consensus this study provides serves as an opportunity to design experiments that test these pathways for mechanisms that drive integration of retroviruses in humans . Our data also provide an important first view of factors that may repair integrated DNA . We expect there are other factors that repair integrated DNA that we did not identify because they also contribute to cDNA recombination . To ask whether repair of integration is broadly conserved , assays will be needed that detect integrated cDNA with unrepaired 5’ ends . These experiments will be able to measure the contribution of each factor to the repair of integrated cDNA .
Edinburgh Minimal Medium ( EMM ) was prepared as described [117] . PM was identical to EMM except the nitrogen source was 3 . 74 gm/l monosodium glutamate . Minimal media were supplemented with 2 gm/l of a dropout mixture that contained equal weights of all amino acids and adenine was added to 2 . 5 times the weight of the other components [29] . When indicated vitamin B1 was added to a final concentration of μM and 5-Fluoroorotic acid ( FOA ) ( U . S . Biologicals , Swampscott , MA . ) was added to a final concentration of 1 mg/ml . When FOA is used in EMM the final concentration of uracil is lowered to 50 μg/ml . The rich medium , yeast extract plus supplements ( YES ) contained 5 g/l Difco yeast extract , 30 g/l glucose , and 2 g/l dropout powder . When indicated the drug nourseothricin ( Nat ) , ( ClonNAT , Jena Bioscience , Germany ) was added to a final concentration of 100μg/ml . The plasmids for this study are listed in S6 Table . The plasmid pHL2882 , used to measure transposition in the deletion strains , includes the nmt1 promoter to express Tf1 with nat disrupted with an artificial intron ( natAI ) ( S1 Fig and Fig 1A ) . pHL2883 and pHL2884 were equivalent to pHL2882 except they have frame shift mutations in PR and IN , respectively . These plasmids were derived from pHL2803 , which expressed Tf1 with a nat marker that lacks the AI . pHL2803 was constructed starting with pHL2673 by replacing the BsrGI-BamHI fragment containing IN sequence and neo with a BsrGI-BamHI fragment that was generated by fusion PCR to introduce restriction sites for AsiSI , SacII , and NotI just upstream of the polypurine tract . The primers for this fusion PCR and all other oligonucleotides are listed in S6 Table . To complete pHL2803 , nat was PCR amplified with primers containing AsiSI and NotI restriction sites and the product was inserted with nat in reverse orientation to Tf1 into pHL2673 with the AsiSI and NotI sites . To produce pHL2804 ( PRfs ) and pHL2805 ( INfs ) , the AvrII-BsrGI fragments of Tf1 from pHL415-2 ( PRfs ) and pHL431-25 ( INfs ) were inserted into the AvrII-BsrGI backbone of pHL2803 . pHL2882 was generated by inserting natAI synthesized commercially by DNA 2 . 0 into the AsiSI and NotI sites of pHL2803 ( S1 Fig ) . The synthetic fragment contained the AI located after the 60th amino acid of Nat ( S1 Fig ) . The codon usage of the nat ORF was optimized for S . pombe without changing the amino acid sequence . pHL2883 and pHL2884 were created by inserting the BsrGI-BamHI fragment with natAI from pHL2882 into the backbones of pHL2804 and pHL2805 , respectively . pHL2898 , pHL2900 , and pHL2902 express Tf1-neoAI from the nmt1 promoter and encode the IN mutations D987N , D1047N , and E1083Q , respectively . These plasmids were made by replacing the BsrGI-NarI fragment of pHL449-1 with PCR fusion products of the BsrGI-NarI fragment containing the mutations . The primers for these PCRs are listed in S7 Table . The deletion library contained 3 , 004 haploid deletion strains from the V2 library of Bioneer ( Alameda , CA , Cat . # M2030 ) [34] . The deletions were derived from two haploid parents ED666 ( h+ ade6-M210 ura4-D18 leu1-32 ) and ED668 ( h+ ade6-M216 ura4-D18 leu1-32 ) . These strains and others are listed in S8 Table . To transform pHL2882 in all 3 , 004 deletion strains , we modified previously published protocols [118] . Using a sterile 96 pin multi-replicator ( Model-VP408FS2AS-1 , V&P Scientific , Inc , San Diego , California , USA ) , each 96 well plate of the library was pined onto single well YES agar plates , and incubated at 32°C for 72hrs ( Fig 2A ) . Each strain was inoculated with an initial OD600nm of 0 . 05 units in 5ml YES liquid media in 15ml tubes . All 96 deletion strains from each plate were independently transformed with pHL2882 ( 10μg ) and 5μg of sonicated herring sperm DNA . Half of each culture was transformed with herring sperm DNA and no pHL2882 as a control for contamination . The transformed cells were processed as indicated in Fig 2A . We isolated four independent transformants for each deletion strain . Strains containing Tf1-natAI ( pHL2882 ) were grown as patches on agar plates with PM-U+L+B1 . These patches were then replica printed on to agar plates with PM-U+L-B1 to induce the nmt1 promoter . After 4 days of incubation , the patches were replica printed onto agar plates with EMM+U+L+B1+FOA twice in succession , the first print was incubated 3 days and the second for 2 days . The patches were then replica printed on to YES+Nat+FOA agar and incubated for 44hrs at 32°C ( Fig 3 ) . Each transposition plate contained patches of PRfs and INfs as controls and the growth of the deletion strains was scored relative to a set of standards ( S2 Fig ) . Four independent transformants of each deletion strain were assayed . An average transposition score was determined if all four transformants had scores within a window of three units . Outliers were excluded from the average if a single transformant had a difference in score three units or greater from the other three . If two transformants had scores that differed by three or more units from the other transformants , the score for the deletion was considered to be inconsistent and were excluded from the screen . To measure amounts of Tf1-cDNA in the nucleus , we used a homologous recombination patch assay as described [29 , 36] . Deletion strains containing pHL2882 were grown on PM-U+L+B1 agar plates for 3 days at 32°C . The patches were replica printed onto PM-U+L-B1 for induction . After 4 days of incubation , the patches were replica printed onto YES+Nat agar and incubated for 24hr at 32°C . The patches were compared with the PRfs and INfs controls from the same plate and scored for homologous recombination using standards ( S3 Fig ) . Four independent transformants of each deletion strain were assayed . The adjusted average scores were determined as previously described in the transposition assay . Strains tested with the quantitative homologous recombination assay were grown on PM-U+L+B1 plates for 3 days at 32°C ( S4 Fig ) . Cells were then suspended into 5ml of PM-U+L-B1 liquid media , and washed six times with 5 ml of PM-U+L-B1 liquid media to remove residual B1 . Cells were then inoculated in 5ml of PM-U+L-B1 media at a starting OD600nm of 0 . 05 units . Following 4 days of incubation the cultures were diluted to OD600nm 1 . 0 ( 2x107 cells/ml ) in PM-U+L+B1 medium and serially diluted from 2x107 cells/ml to 2 x104 cells/ml using PM-U+L+B1 , then spread on YES and YES+Nat ( 100 μg/ml ) agar plates and grown for 3 days at 32°C . Colonies were counted per plate , and the homologous recombination frequencies were determined with the following equation: QuantitativeHomologousRecombinationFrequency= ( numberofcoloniesonYES+NAT ) *100 ( numberofcoloniesonYES*dilutionfactor ) Recombination frequencies for wild-type Tf1-natAI in wild-type strains without deletions ranged from 3% to 1 . 5% in individual experiments . Values for each deletion strain were normalized to wild-type strains assayed during the same experiment . 10 ml cultures were inoculated with a starting OD600nm of 0 . 05 units . After 18 hours , cells were washed with sterile deionized water . The cell pellets were suspended in 0 . 4ml of extraction buffer consisting of 15 mM KCl , 10 mM HEPES-KOH ( pH 7 . 8 ) , 5 mM EDTA , 5 mM dithiothreitol , protease inhibitor cocktail tablet ( Complete , Roche Lifesciences ) , 2 mM phenylmethylsulfonylfluoride ( PMSF ) , Pepstatin ( 0 . 7mg/ml , 1000x stock ) , leupeptin ( 0 . 5mg/ml 1000x stock ) , and Aprotinin ( 1 . 0mg/ml 1000x stock ) . An equal volume of acid-washed glass beads was added and vortexed in a bead beater for a total of 3 min in 30 sec intervals separated by 30 sec rest . 0 . 1 ml of extraction buffer was mixed into the extract , and the liquid was removed . Extracts were combined with 2X sample buffer and boiled . The samples were loaded onto an SDS–10% polyacrylamide gel . The gels were electrotransferred to Immobilon-FL membranes ( Millipore ) . The production bleeds of 660 ( anti-Gag ) and 657 ( anti-IN ) were used to probe Tf1-IN and Gag protein levels , and monoclonal anti-α-Tubulin antibody ( Sigma-Aldrich , USA ) was used as a loading control on all immunoblot experiments . The anti-alpha tubulin , 660 ( anti-Gag ) and 657 ( anti-IN ) were used with 1:5000 , 1:10 , 000 , and 1:5000 , respectively . The fluorescently-tagged secondary mouse IR-Dye 700 and rabbit anti-body IR-Dye 800 ( Rockland Immunochemicals Inc . Limerick , PA ) were used with 1:20 , 000 dilutions . The Immobilon-FL membranes were scanned with an Odyssey infrared imaging system ( LI-COR Biosciences ) . Fluorescence levels from antibodies specific for Gag and IN were normalized to amounts of tubulin and measured with a Li-COR digital instrument ( Materials and methods ) ( S3 Table ) . For the deletion strains tested two independent transformants were assayed for Tf1-IN and Gag protein levels . Geometric means of Gag and IN levels of these replica pairs of deletion strains were compared to the geometric means of the wild-type strains lacking the deletion . The Tf1-IN and Gag protein expression levels were measured and normalized to alpha tubulin . The fold change in Tf1-IN and Gag protein expression were calculated using below equation: ChangeinINandGaglevels=NormalizedgeometricmeanofINandGagproteininmutantNormalizedgeometircmeanofINandGaginwildtypestrain Tf1 was expressed by incubating the cells for 2 days in 50 ml of EMM –B1 starting at OD600 = 0 . 05 to induce the nmt1 promoter after washing them 4 times in EMM –B1 . Genomic DNAs were isolated from 200 OD units of the resulting cultures . Southern blots were performed as described previously [36 , 47] with the following modifications . The nat probe was produced by digesting 5μg of pHL2597 with 160 units of EcoRI , isolating the 1 . 2 kb fragment from a 0 . 7% agarose gel and random-priming labelling with 32P-CTP . One microgram of gDNAs were digested with 40 units of BsrGI , separated on a 1 . 0% agarose gel and transferred to a nylon membrane . The blot was hybridized with the nat probe . BsrGI digestion resulted in Tf1 cDNAs being detected at 2 . 8 kb and the Tf1 expression plasmid at 14kb . Tf1 cDNA was quantified with phosphoimaging and normalized to the amount of expression plasmid . Briefly , the 32P-signal was detected by phosphoimaging on a Typhoon FLA-9500 . The relative level of cDNA was determined by normalizing the signal intensity of the 2 . 9kb cDNA band to the signal intensity of the 14kb plasmid band . Tf1 transposition was induced in strains containing Tf1-natAI ( pHL2882 ) and deleted for pht1 , rad51 , set1 , cwf12 , snf5 , nup61 , rad50 or wild-type ( S8 Table ) as described previously but with some modifications [31] . Briefly , cells were washed 4 times in EMM –B1 before being inoculated at OD600 = 0 . 05 in EMM –B1 to induce the nmt1 promoter , then grown 4 days for each of two passages . Cells with transposition events were selected by incubating them in 50 ml of EMM+B1+FOA for 4 days followed by 4 days in YES+FOA+Nat . Genomic DNAs were isolated from 200 OD units of the resulting cultures . Libraries were prepared for Illumina sequencing according to Chatterjee et al . [31] and sequenced on a MiSeq System ( Illumina ) with custom primers . The sequence of linker oligonucleotides and primers used are given in S7 Table . To determine the genome-wide integration profiles raw sequence reads were processed through a custom suite of Perl scripts [30] modified to accommodate sequences of Tf1-natAI and Illumina technology . Maps of integration relative to ORFs and Sap1 motifs were performed according to previous work [32] . Density plots were obtained using the R function densCol from the package grDevices [119] . R: A language and environment for statistical computing . R Foundation for Statistical Computing , Vienna , Austria . URL https://www . R-project . org/ ) . The sequence data can be obtained from the SRA database with the accession SRA Study: SRP100942 . The Biological Process slim terms optimized for S . pombe were applied to genes important for late stages of transposition of Tf1 , Ty1 and Ty3 . Term enrichments were calculated against the list of non-essential genes available from the Bioneer S . pombe deletion library and the S . cerevisiae ORF deletion collection in strain BY4741 from Invitrogen MapPairs . The p-values were calculated using hypergeometric distance and corrected for multiple comparison with false discovery rate . Full-length Tf1 IN was fused to the C-terminus of a truncated DNA binding domain of LexA by ligating the IN sequence into the EcoRI and SalI sites of pSH2-1 [120] . Full-length Tf1 IN was also fused to the C-terminus of the Gal4 activation domain by ligating IN sequence into the XhoI site of pACT [121] . The host strain used in the two-hybrid screen was S . cerevisiae strain CTY10-5d ( MATα ade2 trp1-901 leu2-3 , 112 his3-200 gal4 gal80 URA3::lexAop-lacZ ura3-52 ) [122] . The two-hybrid assays detected production of lacZ by lifting colonies to 3MM nitrocellulose filter ( Whatman ) that was then stored at -80°C overnight . The filters were thawed and at room temperature tested for galactosidase activity using X-gal [122] . The sequences of Cwf3 ( amino acids 3–284 ) and Rhp18 ( amino acids 16–308 ) were inserted into pACT . | Retroviruses and retrotransposons are genetic elements that propagate by integrating into chromosomes of eukaryotic cells . Genetic disorders are being treated with retrovirus-based vectors that integrate corrective genes into the chromosomes of patients . Unfortunately , the vectors can alter expression of adjacent genes and depending on the position of integration , cancer genes can be induced . It is therefore essential that we understand how integration sites are selected . Interestingly , different retroviruses and retrotransposons have different profiles of integration sites . While specific proteins have been identified that select target sites , it’s not known what other cellular factors promote integration . In this paper , we report a comprehensive screen of host factors that promote LTR-retrotransposon integration in the widely-studied yeast , Schizosaccharomyces pombe . Unexpectedly , we found a wide range of pathways and host factors participate in integration . And importantly , we found the cellular processes that promote integration relative to recombination in S . pombe are the same that drive integration of LTR-retrotransposons in the distantly related yeast Saccharomyces cerevisiae . This suggests a specific set of cellular pathways are responsible for integration in a wide range of eukaryotic hosts . | [
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Materials",
"and",
"methods"
] | [
"recombination-based",
"assay",
"dna-binding",
"proteins",
"fungi",
"model",
"organisms",
"experimental",
"organism",
"systems",
"dna",
"epigenetics",
"molecular",
"biology",
"techniques",
"chromatin",
"schizosaccharomyces",
"homologous",
"recombination",
"research",
"and",
... | 2017 | Host factors that promote retrotransposon integration are similar in distantly related eukaryotes |
CD8+ T cells have been shown to play a crucial role in Trypanosoma cruzi infection . Memory CD8+ T cells can be categorised based on their distinct differentiation stages and functional activities as follows: stem cell memory ( TSCM ) , central memory ( TCM ) , transitional memory ( TTM ) , effector memory ( TEM ) and terminal effector ( TTE ) cells . Currently , the immune mechanisms that control T . cruzi in the chronic phase of the infection are unknown . To characterise the CD8+ T cell subsets that could be participating in the control of T . cruzi infection , in this study , we compared total and T . cruzi-specific circulating CD8+ T cells with distinctive phenotypic and functional features in chronic chagasic patients ( CCPs ) with different degrees of cardiac dysfunction . We observed a decreased frequency of total TSCM along with an increased frequency of TTE in CCPs with severe disease . Antigen-specific TSCM cells were not detectable in CCPs with severe forms of the disease . A functional profile of CD8+ T cell subsets among CCPs revealed a high frequency of monofunctional CD8+ T cells in the most severe patients with IFN-γ+- or TNF-α+-producing cells . These findings suggest that CD8+ TSCM cells may be associated with the immune response to T . cruzi and outcome of Chagas disease , given that these cells may be involved in repopulating the T cell pool that controls infection .
The memory CD8+ T cell compartment comprises cells that represent distinct differentiation stages and different functional activities . This cellular compartment has been divided into stem cell memory ( TSCM ) , central memory ( TCM ) , transitional memory ( TTM ) , effector memory ( TEM ) and terminal effector ( TTE ) cells [1] . TSCM are considered an early differentiated and long-lived human memory T cell population with an enhanced capacity for self-renewal and a multipotent ability to generate other subsets of memory cells [2] . As shown in a viral infection model in non-human primates , TSCM cells also demonstrate better survival capacity compared with conventional memory T cells , even in the presence of little or no antigen stimulus [3] . Therefore , as a potential reservoir to maintain and to replenish the memory T cell pool , TSCM cells represent a potential tool for cellular immune therapies in chronic infectious diseases . Chagas disease ( CD ) , caused by the intracellular parasite Trypanosoma cruzi , is a public health problem that affects nearly 8 million people in Latin America , and almost 25 million people are at risk for contracting this disease [4] . In addition , cases are reported on different continents due to the migration of people from CD-endemic countries [5] . Classically , the course of CD consists of consecutive acute and chronic phases . The acute phase , lasting several weeks , is associated with a high parasitaemia that can be controlled but not eliminated by the immune system . Indeed , parasite persistence at low levels is the hallmark of the indeterminate or asymptomatic phase , which can last a lifetime . However , between 30–40% of infected individuals in the chronic phase develop a symptomatic phase with heart or gastrointestinal involvement [6] . Currently , the immune mechanisms that control T . cruzi infection and do not permit chronic phase progression are unknown . However , in mouse models of T . cruzi infection , it was shown that CD8+ T cells contribute to the control of intracellular pathogen infection by secreting cytokines and perforin . For example , CD8+ T cell knockout ( KO ) , IFN-γ KO and perforin KO mice infected with T . cruzi were unable to control parasitemia and succumbed faster to infection than wild-type infected mice [7] , [8] . In humans with severe cardiac forms of CD , it has been demonstrated that CD8+ T cells decline both in number and function , and there is a low frequency of early differentiated cells along with a high frequency of late differentiated cells compared with patients with less severe forms of the disease [9] . Additionally , patients with severe disease forms have a lower frequency of IFN-γ-producing T cells than patients with mild forms [9] , [10] . Indeed , a low frequency of IFN-γ-producing CD4+CD8+ T cells , reduced proliferative capacity and CD28 expression in T cells have been observed in patients with severe forms of the disease in previous group studies [11] , [12] . As CD8+ T cells are a heterogeneous population with distinct proliferative , survival and functional abilities , it is important to characterise CD8+ T cell subsets in chronic chagasic patients ( CCPs ) to define the types of cellular immune responses participating in the control of T . cruzi . The aim of the present study was to compare circulating CD8+ T cell subsets in CCPs with different degrees of disease severity , with particular focus on TSCM cells , which have the capability to generate all memory subsets .
The Research and Ethics Committees from the Pontificia Universidad Javeriana , Instituto Nacional de Salud , Fundación Abood Clínica Shaio and Hospital Universitario San Ignacio approved this study following the national regulations and the Declaration of Helsinki . Signed informed consent was obtained from all individuals prior to their inclusion in the study . A total of 32 cardiac CCPs from endemic areas were recruited at the Instituto Nacional de Salud , Fundación Abood Clínica Shaio and Hospital Universitario San Ignacio in Bogotá , Colombia . Additionally , nine healthy donors ( HD ) from non-endemic areas were included . All subjects were tested for Trypanosoma cruzi antibodies using an indirect immunofluorescence assay ( IFI ) and an enzyme-linked immunosorbent assay ( ELISA ) . CCPs were classified into groups A , B , C or D according to their disease severity score as previously described [13] . Group A included individuals with a normal electrocardiogram ( ECG ) , heart size and left ventricular ejection fraction ( LVEF ) and a New York Heart Association ( NYHA ) class I designation . Group B individuals had an abnormal ECG but normal heart size and LVEF and a NYHA class I designation . Group C individuals had an abnormal ECG , increased heart size , reduced LVEF and a NYHA class II or III designation . Finally , group D individuals had an abnormal ECG , increased heart size , reduced LVEF and were NYHA class IV . Patients from groups A and B correspond to patients with mild forms of disease severity , and those from groups C and D are patients with severe forms . Clinical characteristics and the classification of study participants are reported in Table 1 . Blood samples were obtained from all study participants in EDTA and heparinised tubes ( BD Vacutainer; Franklin Lakes; NJ , USA ) . The absolute number of lymphocytes was determined from the EDTA tube by a standard differential blood count . Peripheral blood mononuclear cells ( PBMCs ) were isolated with a Ficoll-Hypaque density gradient ( GE Healthcare; Uppsala , Sweden ) from the heparinised tubes . Non-frozen cells were used in phenotypic and functional activity analyses . The following conjugated antibodies were used for cell-surface staining: CD3-Pacific Blue ( BD Pharmingen; Clone UCHT1; Cat . No . 558117; San Diego , CA , USA ) , CD8-APC H7 ( BD Pharmingen; Clone SK1; Cat . No . 641400 ) , CD45RA-PE ( BD Pharmingen; Clone HI100; Cat . No . 555489 ) , CCR7-PE-Cy7 ( BD Pharmingen; Clone 3D12; Cat . No . 557648 ) , CD28-PerCP-Cy5 . 5 ( BD Biosciences; Clone L293; Cat . No . 337181; San Jose , CA , USA ) , CD27-Alexa Fluor 700 ( BD Pharmingen; Clone M-T271; Cat . No . 560611 ) , CD95-APC ( BD Pharmingen; Clone DX2; Cat . No . 558814 ) and CD127-FITC ( BD Pharmingen; Clone HIL-7R-M21; Cat . No . 560549 ) . Conjugated antibodies for intracellular staining included the following: IFN-γ-FITC ( BD Pharmingen; Clone 4S . B3; Cat . No . 554551 ) , IL-2-PerCP-Cy5 . 5 ( BD Pharmingen; Clone MQ1-17H12; Cat . No . 560708 ) and TNF-α-AlexaFluor 700 ( BD Pharmingen; Clone MAb11; Cat . No . 557996 ) . To exclude dead cells , the Fixable Aqua Dead Cell Stain viability marker was used ( Invitrogen; Cat . No . L34957; Eugene , OR , USA ) . All conjugated antibodies were titrated , and each multicolour panel of conjugates was evaluated as previously described [14] . To evaluate the frequency of CD8+ T cell subsets , one million PBMCs were stained with the viability marker for 20 min in the dark at room temperature and then washed with PBS 0 . 001 M pH 7 . 4 ( 1X PBS ) ( Eurobio; Les Ulis , France ) . Cells were stained with antibodies against CD3 , CD8 , CD45RA , CCR7 , CD28 , CD27 , CD127 and CD95 molecules for 30 min in the dark at 4°C and washed with 1X PBS . To evaluate the cytokine production of CD8+ T cell subsets , one million PBMCs were cultured with anti-CD28 and anti-CD49d antibodies and incubated for 6 hours in the presence of brefeldin A ( BD Biosciences , San Jose , CA , USA ) with Staphylococcal enterotoxin B ( SEB ) ( Sigma-Aldrich; Saint Louis , MO , USA ) , Trypanosoma cruzi trypomastigote lysate or medium . Parasite lysate was obtained as previously described [14] . First , cells were stained with the viability marker and then with surface antibodies against CD3 , CD8 , CD45RA , CCR7 and CD95 molecules for 30 min in the dark at 4°C . Cells were washed with 1X PBS , fixed and permeabilised with Cytofix/Cytoperm ( BD Biosciences ) for staining with antibodies against IFN-γ , TNF-αand IL-2 for 30 min in the dark at 4°C , followed by washing with 1X Perm/Wash ( BD Biosciences ) . At least 50 , 000 events gated on CD3+CD8+ cells were acquired on a FACS Aria II flow cytometer . Analysis was performed using FlowJo 9 . 3 . 2 ( Tree Star; Ashland , OR , USA ) , Pestle 1 . 7 ( National Institutes of Health ( NIH ) , Bethesda , MD , USA ) and SPICE 5 . 3 ( NIH ) software [15] . Dead and doublet cells were excluded from the analysis , as previously described [14] . A positive cytokine response was defined by subtracting the cytokine background ( cells cultured with medium ) from a frequency of >0 . 05% for each CD8+ T cell subset ( average frequency of the response of CD8+ T cells from HDs cultured with parasite lysate plus 3 SD ) . Conventional PCR ( cPCR ) and quantitative PCR ( qPCR ) were used to assess parasite DNA in blood samples in guanidine hydrochloride-EDTA stored at 4°C from 30 CCPs . DNA from blood was extracted using a High Pure PCR template preparation kit ( Roche , Mannheim , Germany ) . Afterwards , cPCR was run using initiators of β-globin FR , as described previously [16] , to check DNA integrity and to rule out the presence of inhibitors in the sample . cPCR was performed with the S35 ( AAATAATGTACGGG ( T/G ) GAGATGCATGA ) and the S36 ( GGGTTCGATTGGGGTTGGTGT ) primers , which amplify the kDNA variable mini-circle region from T . cruzi , using PCR reaction conditions described previously [17] , [18] . qPCR was performed with Cruzi 1 ( ASTCGGCTGATCGTTTTCGA ) and Cruzi 2 ( AATTCCTCCAAGCAGCGGATA ) primers and the Cruzi 3 ( 6FAM-CACACACTGGACACCAA-BBQ ) probe , which amplify a 166-bp segment of the satellite DNA from T . cruzi [19] . Each sample was run in duplicate , and the parasite load was estimated based on a standard curve . The curve was constructed with different concentrations of genomic DNA mixed with blood from one uninfected donor ranging from 105–100 parasite equivalents per mL . Samples were run using a LightCycler 1 . 5 Instrument ( Roche ) . For both PCR methods , different controls were included: reaction ( water added in the room containing the reaction mixture ) , grey ( water added in the room where the sample was added to the reaction ) , negative ( genomic DNA from an HD ) and positive ( genomic DNA of T . cruzi ) . A statistical analysis was performed using the Mann-Whitney test or a one-way ANOVA non-parametric Kruskal–Wallis test with Dunn's test for multiple comparisons . Correlations between the frequencies and the absolute numbers of CD8+ T cell subsets were analysed using Spearman's rank correlation coefficient . A Wilcoxon signed-rank test was performed to compare stimulated cells with 2 functions or 1 function . All tests were two-tailed , and statistical significance was achieved with p<0 . 05 . GraphPad Prism 6 . 0 for Mac OS X ( San Diego , CA , USA ) software was used for statistical analyses .
T cells are a highly heterogeneous cell compartment comprising different phenotypes , functional activities , gene expression and survival capacities . Recently , CD45RA ( or CD45R0 ) , CCR7 , CD28 and CD95 were proposed as canonical markers to identify T cell subsets using multiparametric flow cytometry [1] . In addition , CD127 and CD27 were included to accurately define T stem cell memory ( TSCM ) cells as described previously [2] . To compare the frequencies of total CD8+ T cell subsets among CCPs with different degrees of disease severity and HDs , PBMCs isolated from 32 CCPs and 9 HDs were labelled with a panel of conjugated antibodies as described in the Materials and Methods . Representative contour plots depicting the ex vivo selection of CD8+ T cell subsets based on differential expression of CD45RA , CCR7 , CD28 , CD27 , CD95 and CD127 are shown in Fig . 1A . On the basis of previous data , CD8+ T cell subsets were defined as TSCM cells ( CD45RA+CCR7+CD28+CD27+CD95+CD127+ ) , central memory ( TCM ) cells ( CD45RA−CCR7+CD28+CD27+CD95+CD127+ ) , transitional memory ( TTM ) cells ( CD45RA−CCR7−CD28+CD27+CD95+CD127+ ) , effector memory ( TEM ) cells ( CD45RA−CCR7−CD28−CD27+CD95+CD127− ) and terminal effector ( TTE ) cells ( CD45RA+CCR7−CD28−CD27−CD95+CD127− ) [1] . The frequencies of CD8+ T cell subsets were similar for group A and B patients and for HDs . A significant difference was observed when the frequencies of TSCM and TTE cells were compared between CCPs with severe and mild forms of the disease ( Fig . 1B ) . No significant differences were observed for TCM , TTM and TEM frequencies between CCPs and HDs . Given that we found significant differences in the frequencies of TSCM and TTE cells from CCPs , we evaluated whether the reduced frequency of TSCM cells was associated with changes in the frequency of TTE cells . Indeed , in CCPs , the frequency of TSCM cells correlated negatively with the frequency of TTE cells ( Spearman r = −0 . 7204 , p<0 . 0001 ) ( Fig . 1C ) . In addition , we found a negative trend in the frequency of TSCM cells and a positive trend in the frequency of TEM and TTE cells in CCPs with various degrees of disease severity ( S1 Fig . ) , as has been shown in previous reports [9] . As variations in the numbers of CD8+ T cells could consequently affect the absolute values of the studied subsets , absolute cell numbers were compared . The absolute numbers for CD8+ T cell subsets demonstrated a trend similar to that for the frequencies of CD8+ T cell subsets observed in CCPs and HDs . A correlation analysis of the absolute numbers of TSCM and TTE cells from all CCPs showed that TSCM cells also correlated negatively with TTE cells from CCPs ( Spearman r = −0 . 4240 , p = 0 . 0156 , S2 Fig . ) . We next compared the frequencies of T . cruzi-specific CD8+ T cells bearing different T cell phenotypes in CCPs with various degrees of disease severity . To evaluate the frequency of antigen-specific CD8+ T cell subsets , PBMCs from CCPs were stimulated with parasite lysate and labelled with a panel of conjugated antibodies to assess the cytokine production of CD8+ T cell subsets . Assessment of antigen-specific CD8+ T cells was carried out for cytokine production in response to parasite lysate as described in the Materials and Methods . Representative density plots for the selection of T . cruzi-specific CD8+ T cells producing IFN-γ , TNF-αand IL-2 are shown in Fig . 2A . An increased frequency of antigen-specific TEM cells was observed in all CCPs ( Fig . 2B ) . CCPs with severe disease demonstrated a low frequency of antigen-specific TCM cells and a high frequency of antigen-specific TTE cells compared with patients with mild disease . Interestingly , we did not detect antigen-specific CD8+ TSCM cells in patients from group D , who had the most severe form of the disease ( Fig . 2C ) . Using the panel described above , we assessed the functional profiles of T . cruzi-specific CD8+ T cell subsets in CCPs and HDs in cultures with medium , SEB and parasite lysate . When comparing the frequencies of TSCM , TCM , TEM and TTE cells from CCPs with one or two functions in lysate-stimulated cells , we observed a high frequency of CD8+ T cell subsets with one function compared with cells with two functions ( p = 0 . 0210 , p = 0 . 0008 , p = 0 . 0353 and p = 0 . 0006 , respectively ) . However , among SEB-stimulated cells , there was a higher frequency of CD8+ T cell subsets with two functions than of cells with one function in all CCPs ( S3 Fig . ) . As expected , in SEB-stimulated cells , we observed TSCM , TCM , TEM and TTE cells with three functions from CCPs in all groups , similar to those observed in HDs . In contrast , we observed an absence of cells with three functions among lysate-stimulated TSCM , TEM and TTE cells; however , patients from groups A and B had TCM cells with three functions , which was not observed in patients from groups C and D . In lysate-stimulated cells in group D patients , there were no TSCM cells with one , two or three functions among SEB-stimulated cells from both CCPs and HDs ( Fig . 3 ) . These findings were also observed in the proportion of cells because the patients from group D are composed of cells producing a single cytokine ( Fig . 4 ) . The most prevalent population in lysate-stimulated cells with two functions consisted of IFN-γ+TNF-α+-producing cells in CCPs with mild forms of the disease , and monofunctional cells were predominantly IFN-γ+ or TNF-α+ in patients from group D . Of note , IL-2-producing cells were not detected in any CCPs ( Fig . 5 ) . To further investigate the associations between CD8+ T cell subsets and parasitaemia in CCPs , cPCR and qPCR were performed to assess the presence of parasites in peripheral blood as described in the Materials and Methods . Both PCR methods permitted the identification of 12 out of 30 ( 40% ) CCPs with detectable parasitaemia ( 7 by cPCR and 9 by qPCR ) ( Table 2 ) . Notably , we detected parasitaemia in 3 of 9 patients from group C; in contrast , in patients from groups A , B and D , the parasitaemia load was below the detection limit of qPCR . Due to the low numbers of individuals , there were no associations between CD8+ T cell subsets and parasitaemia; however , this finding demonstrated the presence of circulating parasites in all CCP groups , even those with the most severe forms of disease .
Immunity aimed at antigen clearance or the control of disease progression has been shown to be directly related to the quality of the memory T cell response [20] . However , the study of memory T cells depends on technical approaches to describe the complex T cell compartment . In viral chronic infections , CD8+ T cell subsets have been extensively studied , but they have rarely been studied in infections caused by intracellular protozoans such as T . cruzi . In this study , we compared the total and antigen-specific circulating CD8+ T cell subsets among CCPs demonstrating different degrees of disease severity . Changes in the distribution of CD8+ T cell subsets could highlight the behaviour of cellular immunity during the natural history of infection and the pathogenesis of CD . We observed a decreased frequency in total TSCM cells along with an increased frequency of TTE cells in CCPs with severe forms of disease . Interestingly , IFN-γ- , TNF-α- and IL-2-producing antigen-specific TSCM cells were not detectable in CCPs with severe forms of the disease . These changes observed for the TSCM and TTE subsets indicated a negative correlation both in the frequency and the absolute numbers of CD8+ T cells in all CCPs analysed . Conversely , when we studied the functional profiles of CD8+ T cell subsets among CCPs , a higher frequency of monofunctional antigen-specific CD8+ T cells was observed in CCPs with severe forms of disease . However , in SEB-stimulated cells , we observed cells with three functions among CD8+ T cell subsets from CCPs with all degrees of disease severity , similar to that observed in HDs ( Fig . 6 ) . The T cell compartment is composed of different memory cells subsets , which are generated in response to antigen recognition . The recently identified TSCM cells are characterised by their high proliferative and self-renewing capacities and their ability to differentiate into TCM and TEM cells [2] . In this study , we described the absence of antigen-specific TSCM cells in CCPs with severe disease . Although TSCM cells have not previously been described in CCPs , our findings are consistent with previous reports showing that the frequency of early differentiated CD8+ T cells decreases as the disease becomes more severe and the proportion of fully differentiated memory CD8+ T cells increases [9] , [21] . Additionally , in a mouse model , it was observed that antigen-specific CD8+ T cells maintain a TEM phenotype during persistent T . cruzi infection [22] . Recently , TSCM cells have been associated with an improved prognosis in chronic HIV-infected patients because the frequency of CD8+ TSCM cells decreased in all individuals with chronic HIV infection , but the frequency of these cells was restored in treated HIV-infected patients . Interestingly , these findings are in accordance with our results demonstrating that the frequency of CD8+ TSCM cells decreased in CCPs with severe forms of disease [23] . In addition , TSCM cells appeared to be progenitors for TTE cells as these two compartments were inversely correlated . Taking into account that disease progression may last several years , it is difficult to know if this progress is due to the loss of TSCM cells or if the loss of this subset is the result of a more severe infection . However , due to their capacities to generate all memory and effector T cell subsets , we hypothesized that the absence or a very low frequency of T . cruzi-specific TSCM cells may be a contributing factor to failure to control the parasite during the symptomatic phase of the disease in T . cruzi chronic infection . In Listeria monocytogenes-infected mice , the adoptive transfer of cells expressing high amounts of IL-7R α-chain ( CD127 ) helped to control infection due to the expansion of effector cell populations responsible for the rapid clearance of bacteria from the spleen and liver [24] . It is noteworthy that TSCM cells express high levels of CD127 and other molecules associated with early differentiation [2] . The adoptive transfer of TSCM cells in murine tumour models was shown to mediate potent tumour regression even better than TCM and TEM cells , and thus it has been proposed that these cells may be used for adoptive immunotherapy [2] , [25] . However , TSCM cell adoptive transfer has not been studied in chronic infections; thus , it would be interesting to evaluate the role of this cell population in the outcome of chronic infection . In the present study , we included healthy donors from non-endemic areas as uninfected controls . However , it may be a different scenario when compared with healthy individuals exposed to T . cruzi in areas of endemic Chagas disease , who can have T . cruzi-specific T cells capable of producing IFN-γ even with negative conventional antibody testing for T . cruzi [26] . It is possible that repetitive exposure to T . cruzi in endemic areas may provide a persistent source of antigens that affect the proportion or function of CD8+ memory T cell subsets in healthy individuals , or some of the immune responses elicited by T . cruzi antigens in vitro may be induced by other protozoan parasites circulating in endemic areas [27] , [28] . The polyfunctionality of CD8+ T cells has been proposed to be an immune correlate of protection in viral chronic infections [20] . Non-progressor patients chronically infected with human immunodeficiency virus ( HIV ) demonstrated a high frequency of polyfunctional CD8+ T cells compared with progressor patients [29] . Antigenic persistence in chronic infections is implicated in the impaired cellular immune response due to excessive activation of the immune system [30] . In the peripheral blood of CCPs , a high frequency of TTE cells in patients with severe disease and a low frequency or absence of antigen-specific cells with one , two or three functions as assessed by cytokine production were observed . This phenomenon could be attributable to the high frequency of cells with a late differentiated stage , as these cells have decreased polyfunctional capacities . Indeed , we observed a significant decrease in the frequency of polyfunctional CD8+ T cells in patients with severe disease and an increase in inhibitory receptors on CD8+ T cells from CCPs ( Lasso , P , et al , manuscript in preparation ) . However , it has been shown in T . cruzi-infected mice that perforin-producing cells may contribute to cardiomyocyte lesions and heart dysfunction during chronic T . cruzi infection [31] . We observed an increased frequency of TTE cells and a higher frequency of perforin-producing cells in CCPs with severe forms of the disease ( Lasso , P , et al , manuscript in preparation ) . Because TTE cells have a greater cytotoxic capacity than early differentiated cells , we suggest that a high frequency of TTE cells may be associated with the heart damage observed in CCPs due to a high frequency of perforin-producing cells observed in these patients . Based on functional CD8+ T cell responses , it was previously reported that T . cruzi-specific CD8+ T cells from chronic T . cruzi-infected individuals display a functional profile with T cells secreting IFN-γ alone as the predominant pattern and a very low prevalence of single IL-2-secreting cells [32] . However , in T . cruzi-infected mice , parasite persistence has been associated with the phenotype of CD8+ T cells because the absence of antigenic load is correlated with an increased frequency of early differentiated cells [33] . The findings of the present study with parasite persistence in CCPs could be associated with the changes observed both in the frequencies and functional activities of CD8+ T cell subsets , as has been described for other chronic infections [34] , [35] . In chronic T . cruzi infection , the parasitaemia load is low [36] . In our study , parasite DNA was detectable in peripheral blood in 40% of patients . This result is similar to findings in other studies that identified parasitaemia in CCPs , where detection by PCR is between 40–50% of confirmed positive samples [37] , [38] . However , in positive blood donors detected by other tests for T . cruzi infection , it has been reported that PCR is positive in only 13% [39] . In addition , a lack of association between blood-based detection of parasite DNA and cardiac damage has been reported [40] . We find parasite DNA even in the most severe forms of disease but the lack of association is most likely due to low parasitism in blood and tissue , the anatomical location of parasites in CCPs and parasite variability [40] . In summary , IFN-γ- , TNF-α- and IL-2-producing T . cruzi-specific CD8+ TSCM cells and polyfunctionality were not detectable in CCPs with severe forms of disease . In the context of chronic T . cruzi infection , we hypothesized that CCPs with mild disease would be able to control infection and prevent the progression of the disease via TSCM cells and polyfunctional cells at early differentiation stages . For unknown reasons , some infected individuals lost the ability to regenerate antigen-specific CD8+ TSCM cells to create enough polyfunctional cells to control the infection . Memory cells can differentiate into TTE cells with only one function and probably with high expression of inhibitory receptors . Overall , these findings related to changes in the distribution of CD8+ T cell subsets , functional activity of CD8+ T cells and parasite persistence in CCPs may be associated with the immune response to and outcome of CD . However , it is still important to evaluate the role of CD8+ TSCM cells in parasite control of the infection . | Chagas disease is caused by the intracellular parasite Trypanosoma cruzi . After the onset of acute infection , all individuals enter the chronic phase and approximately 70% of them never have symptoms . However , nearly 30% of infected individuals develop symptoms , mainly of heart disease , even decades after the initial infection . Currently , it is unclear how the immune response controls infection and prevents the development of heart disease in some infected people . We have characterised the memory CD8+ T cell subsets in chronic chagasic patients , including a newly described population of cells called memory stem cells . This T cell subset seems important for replenishing the other T cell populations . The findings in this manuscript show that chronic chagasic patients with severe disease have the following: a ) a low frequency of memory stem cells , b ) no antigen-specific memory stem cells , and c ) CD8+ T cells with less effector function compared with asymptomatic patients . These results indicate that the lack of T cell population renewal and the decrease in cells with multiple effector functions may be associated with the clinical outcome of chronic Chagas disease . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"chagas",
"disease",
"neglected",
"tropical",
"diseases",
"biology",
"and",
"life",
"sciences",
"immunology",
"tropical",
"diseases",
"protozoan",
"infections",
"parasitic",
"diseases",
"immune",
"response"
] | 2015 | Low Frequency of Circulating CD8+ T Stem Cell Memory Cells in Chronic Chagasic Patients with Severe Forms of the Disease |
Trypanosomes show an intriguing organization of their mitochondrial DNA into a catenated network , the kinetoplast DNA ( kDNA ) . While more than 30 proteins involved in kDNA replication have been described , only few components of kDNA segregation machinery are currently known . Electron microscopy studies identified a high-order structure , the tripartite attachment complex ( TAC ) , linking the basal body of the flagellum via the mitochondrial membranes to the kDNA . Here we describe TAC102 , a novel core component of the TAC , which is essential for proper kDNA segregation during cell division . Loss of TAC102 leads to mitochondrial genome missegregation but has no impact on proper organelle biogenesis and segregation . The protein is present throughout the cell cycle and is assembled into the newly developing TAC only after the pro-basal body has matured indicating a hierarchy in the assembly process . Furthermore , we provide evidence that the TAC is replicated de novo rather than using a semi-conservative mechanism . Lastly , we demonstrate that TAC102 lacks an N-terminal mitochondrial targeting sequence and requires sequences in the C-terminal part of the protein for its proper localization .
Elegant electron microscopy analysis revealed a structure that connects the basal body with the kDNA disk , the tripartite attachment complex ( TAC ) [20] . The TAC consists of ( i ) the exclusion zone filaments , a region between the basal body and the outer mitochondrial membrane devoid of ribosomes; ( ii ) the differentiated mitochondrial membranes , which are inert to detergent extraction; and ( iii ) the unilateral filaments that connect the inner mitochondrial membrane with the kDNA spanning a region that has been described as the kinetoflagellar zone ( KFZ ) [1 , 2] . Although the basal body does not directly belong to the TAC structure , it is a key organizer in the T . brucei cell and the posterior anchoring point of the TAC [1 , 2 , 21] . A few markers for the basal body and the TAC have been described . Basal body markers include YL1/2 that recognizes the aggregation of non-polymerized tyrosinated tubulin in the transitional fibers of the mature flagellum [22] , and BBA4 that recognizes an unknown protein in the pro- and mature basal bodies [23] . Furthermore , two components of the exclusion zone filaments have been described . The monoclonal antibody MAB22 recognizes a cytoskeletal component of the exclusion zone filaments ranging from the proximal end of the basal body to the outer mitochondrial membrane [24] . The unidentified structure recognized by MAB22 seems to be insensitive to extraction by high concentrations of non-ionic detergents , which is consistent with the earlier descriptions of the TAC . The other known component of the exclusion zone filaments is a ~197 kDa protein ( p197 ) , which was shown to localize in the same region as MAB22 by immunofluorescence microscopy [25] . Depletion of p197 leads to a kDNA segregation phenotype where most cells are devoid of kDNA and a small number of cells accumulate a huge amount of kDNA [25] . For the differentiated membranes , a recently described beta barrel protein ( TAC40 ) of the outer mitochondrial membrane ( OM ) has been demonstrated to be a TAC component [26] . Depletion of TAC40 leads to a phenotype similar to that described for p197 . Additionally , electron microscopy studies demonstrated that the overall ultrastructure of the kDNA remains intact but daughter networks are not separated in cells depleted of TAC40 [26] . Another protein of the differentiated membrane is p166 , historically the first TAC component to be described , which localizes to the inner mitochondrial membrane and its depletion leads to a phenotype similar to that described above for p197 and TAC40 [27] . Additionally , it was shown that the kDNA loss phenotype upon p166 RNAi is indeed a consequence of asymmetrical segregation rather than improper replication of the mitochondrial genome [27] . Potential candidates for anchoring the kDNA to the intermediate filaments are the two universal minicircle sequence binding proteins ( UMSBP1 and UMSBP2 ) [28 , 29] . A homologue of these proteins in Crithidia fasciculata has been shown to specifically bind to two conserved sequences in the minicircles [30] . While the main function of the UMSBPs seems to be the initiation of minicircle replication , they might have additional functions in kDNA segregation . Currently it is unclear , if the kDNA segregation phenotype that is seen upon the loss of both UMSBPs is a consequence of the loss of minicircle replication or the proteins are directly involved in segregating the kDNA in trypanosomes . Aside from the UMSBPs , there are currently no other candidates for intermediate filament proteins . AEP-1 is a mitochondrially encoded protein produced from the alternatively edited cytochrome c oxidase III mRNA . Overexpression of a recoded nuclear version of the C-terminally truncated AEP-1 ( ΔC-AEP-1 ) led to a kDNA loss phenotype reminiscent of the phenotypes described above . Very likely this protein localizes to the inner mitochondrial membrane with a clear enrichment in the KFZ [31 , 32] . Recently , the alpha-ketoglutarate dehydrogenase E2 ( α-KDE2 ) subunit , which is an essential Krebs cycle enzyme in the insect form trypanosomes , has been shown to be also important for proper kDNA segregation [33] . α-KDE2 seems to be localized throughout the mitochondrion and at the inner mitochondrial membrane and loss of this enzyme leads to a growth defect and a kDNA segregation phenotype . In this study , we report a novel mitochondrial TAC protein ( TAC102 ) , which is likely part of the unilateral filaments . We show its localization throughout the cell cycle and characterize the phenotype which occurs during RNAi-induced loss of TAC102 and suggest a mechanism for the replication of the TAC structure . Additionally , we demonstrate that TAC102 has sequences in its C-terminus that are required for proper localization to the mitochondrial organelle and the TAC .
Using RNAi which targets the ORF of TAC102 , we depleted the transcript in bloodstream form ( BSF ) parasites and detected slower cell growth after three days of RNAi induction ( Fig 1A ) . Seven days post RNAi induction the cells stopped growing entirely . The effectiveness of RNAi was shown by probing for the TAC102 mRNA and protein on northern and western blots , respectively ( Fig 1A ) . Analysis of the DNA content of BSF cells by DAPI staining and fluorescence microscopy showed that approximately 75% of the cells had lost the kDNA two days after induction of RNAi ( Fig 1B and 1C ) , while a small number of cells contained very large or “tiny” kDNAs ( Fig 1B and 1D ) . The median intensity of the remaining kDNAs in the population increased by more than two fold 48 hours post RNAi induction based on DAPI fluorescence intensity ( Fig 1D ) . Overall the amount of kDNA in the population decreased to ~45% of the wild type situation after two days of RNAi as measured by probing for the kDNA minicircles on Southern blots ( Fig 1E ) . Also the number of cells properly segregating their kDNA ( 2k1n cells ) dropped from ~15% in the uninduced population to less than 2% after two days of RNAi induction . Together with the appearance of very large kDNAs this suggests that loss of TAC102 has an impact on mitochondrial genome segregation rather than replication . In order to test if the loss of TAC102 also had influence on mitochondrial morphology , we stained the cells with antibody against the mitochondrial heat shock protein 70 ( mtHSP70 ) . During the first two days of RNAi induction we did not detect any changes in mitochondrial morphology ( Fig 1B ) , except in cells that accumulated very large kDNAs; here an increase in organelle volume at the site of the kDNA was observed . Also the segregation of the mitochondrial organelle during cell division even in the absence of kDNA seemed to be unaffected ( Fig 1B ) . Essentially the same RNAi phenotype was observed in the insect form parasites ( procyclic form , PCF; S2 Fig ) . Here we also carefully characterized the cells ( <1% ) that showed unequal segregation of the kDNA ( S2I Fig ) . In the majority of these cells the enlarged kDNA was associated with the old basal body and flagellum , and it was positioned in most cases between the two nuclei . Thus , in both life cycle forms of T . brucei loss of TAC102 leads to kDNA missegregation rather than kDNA replication defect , which in turn leads to the loss of the mitochondrial genome in most cells . Based on our observations , we assume that in most cases the remaining kDNA is associated with the old basal body . In order to investigate the ultrastructure of the enlarged kDNA networks in BSF cells we employed transmission electron microscopy . The kDNA is organized in a disk-shaped structure which is situated close to the inner mitochondrial membrane . When the kDNA disk is viewed from the “side” , the basal body can often be seen juxtaposed on the other side of the mitochondrial membranes ( Fig 2A ) . Just prior to kDNA segregation , the kDNA assumes a kinked conformation ( Fig 2B ) , which can be explained through the connection of the kDNA to the basal bodies and the movement of the new basal body around the old one . In BSF cells with enlarged kDNA networks two days after the induction of RNAi against TAC102 the kDNA generally maintains the striated ultrastructure of the disk ( Fig 2C and 2D ) , however it does not assume a clear kinked structure ( Fig 2C ) and often the kDNA seems folded upon itself with additional smaller kDNA disks ( Fig 2D ) . The median diameter of the sum of the striated kDNA disks from the enlarged networks was around 680 nm compared to 440 nm in the parental cell line ( Fig 2G ) . In some cells , instead of a properly structured kDNA , we detected small patches of electron dense material ( edm ) lacking the typical striated appearance with a median diameter of 320 nm ( Fig 2E ) . In the case of complete loss of the kDNA , the mitochondrial membranes remain in close proximity to the basal bodies and appear intact ( Fig 2F ) . It also seems that the exclusion zone is unaffected since few or no ribosomes are present in this area ( Fig 2F ) . Loss of kDNA is a frequent phenotype observed in trypanosomes when mitochondrial functions are affected . To test if loss of TAC102 has a direct or indirect effect on kDNA segregation , we made use of a recently described T . brucei BSF cell line that contains a single point mutation in the γ-subunit of the ATP-synthase ( γL262P ) and is able to shed its kDNA without any detectable growth defect [37] . RNAi targeting the ORF of TAC102 mRNA in this cell line led to the same phenotype as described above , i . e . loss of kDNA in the majority of cells and large/tiny kDNA networks in few cells; however , in this case the cells lacking kDNA continued to grow at wild type rates eventually leading to an akinetoplastic population without any defects in basal body or flagellar biogenesis . These experiments demonstrate that TAC102 is only essential in cells that require kDNA for proper growth ( Fig 3A–3C ) . In order to localize TAC102 in cells , the protein was tagged in situ at the N-terminus using a dual affinity tag PTP ( ProtC-TEV-ProtA; [38] ) . Super-resolution confocal microscopy using a STimulated Emission Depletion ( STED ) instrument showed co-localization of the N-terminally tagged TAC102 with MitoTracker in BSF cells ( Fig 4A ) . Immunofluorescence microscopy detected the protein in the posterior part of the mitochondrial organelle between the basal body of the flagellum and the kDNA disk , in both BSF ( Fig 4B ) and PCF ( Fig 4D ) cells . The localization pattern of the tagged TAC102 was also confirmed in BSF and PCF cells by immunostaining with polyclonal and monoclonal antibodies against TAC102 ( S2M and S2N Fig ) . Furthermore , biochemical fractionation using digitonin and differential centrifugation steps supported the mitochondrial localization of TAC102 ( Fig 4C ) . We attempted to investigate the localization in more detail using different concentrations of digitonin to extract cells ( Fig 4E ) . TAC102 is observed in the soluble fraction at 0 . 1% detergent , resembling the behavior of lipid dehydrogenase ( LipDH ) , a mitochondrial matrix protein . However , the archaic translocase of the outer mitochondrial membrane ( ATOM ) and the inner mitochondrial membrane protein cytochrome oxidase subunit 4 ( COXIV ) , known to strongly associate with membranes , are solubilized only at 0 . 3% digitonin . From this we conclude that TAC102 is likely to localize in the mitochondrial matrix ( Fig 4E ) . In order to test if TAC102 is indeed part of the TAC structure , we analyzed flagella isolated from BSF cells for the presence of the basal body , TAC102 and the kDNA using immunofluorescence microscopy ( Fig 5 ) . Under these conditions , >90% of the flagella showed a signal for the basal body and about 50% had kDNA attached to their posterior end as demonstrated by DAPI staining . Of the kDNA-positive flagella , >90% had a signal for TAC102 in close proximity to the basal body and the kDNA , indicating that TAC102 , similarly to the previously described TAC40 and p166 , is a component of the TAC . We also analyzed flagella isolated from PCF cells using immunofluorescence microscopy and western blotting ( S3 Fig ) . Similarly to BSF cells , we observed that most flagella were positive for the presence of the TAC102 signal and the kDNA . In order to resolve the flagellar extract by SDS-PAGE and probe for TAC102 by western blotting , we had to treat the extracted flagella with DNAse I . TAC102 was found in both the flagellar and the soluble fractions , indicating that some part of the protein is solubilized by Triton X-100 used for flagella isolation , whereas a significant portion is retained at the flagella . A more detailed analysis of the localization of TAC102 using immunofluorescence microscopy shows that the protein is present throughout the cell cycle in whole cells ( Fig 6 ) . In the G1 phase , prior to kDNA replication , the TAC102 signal occupies a region between the basal body and the kDNA that is smaller than the kDNA structure as seen by DAPI staining ( Fig 6A ) . Furthermore , based on staining of TAC102 and the mature basal body ( YL1/2 antibody ) , the new TAC102 signal only appears after the new basal body matures ( Fig 6B and 6C ) . During the nuclear S phase , the new basal body moves to its posterior position and TAC102 is present at both the old and the new basal body ( Fig 6E and 6F ) . In G2/M , after kDNA segregation , TAC102 remains between the kDNA and the basal body as described for the situation in G1 ( Fig 6G ) . In order to test if the TAC is replicated de novo or by a semi-conservative mechanism , we induced RNAi against TAC102 in BSF cells for a short period ( 18 hours ) and stained the cells with antibodies against TAC102 and the basal body ( YL1/2 ) ( S4 Fig ) . As expected , in non-induced cells each basal body is associated with a TAC102 signal and a kDNA . However , when TAC102 was depleted by RNAi for 18 hours , we observed cells that had two basal bodies , but just one TAC102 signal and one kDNA associated with it . The percentage of such cells in the population is 9 . 4% . Of such cells , more than 90% have the TAC102 signal/kDNA at the more anterior , old basal body . These experiments suggest that TAC102 is assembled de novo into the TAC . Since TAC102 does not contain a detectable mitochondrial targeting signal at the N-terminus we aimed to characterize the region of the protein necessary for its proper localization to the TAC . For this we overexpressed inducible ectopic copies of N-terminally myc-tagged TAC102 in PCF parasites that were ( i ) truncated at the N-terminus ( deletion of the first 200 aa , mycΔN-TAC102 ) or ( ii ) the C-terminus ( deletion of aa 650−951 , mycΔC-TAC102 ) , as well as ( iii ) the full-length TAC102 ( myc-TAC102 ) and followed their localization in the cell by fluorescence microscopy and biochemical digitonin fractionation . The truncations cover most of the conserved regions of TAC102 in the N- and C-terminus . Myc-TAC102 localizes to the position of the endogenous protein as determined by immunofluorescence microscopy and western blotting of digitonin fractionations ( Fig 7 , S5 Fig ) . Furthermore , overexpression of the N-terminally tagged TAC102 does not lead to any detectable growth or cell cycle phenotype ( Fig 7 , S5 and S6A Figs ) . The N-terminally truncated TAC102 ( mycΔN-TAC102 ) localizes to the TAC , however after five days of overexpression 25% of the cells show additional foci of mycΔN-TAC102 that are in the mitochondrial organelle as confirmed by biochemical fractionation; some of the TAC102 foci co-localize with small ancillary kinetoplasts ( Fig 7 , S5 Fig ) . Most of the cells with ancillary kDNAs , >80% , have one “extra” kDNA per cell ( S6B Fig ) . After five days of induction we also could detect a reduction in the number of cells in G1 ( 1k1n ) as well as an increase in cells without kDNA ( 10%; S6A Fig ) . Furthermore , mycΔN-TAC102 cells displayed a very weak growth defect that starts six days post induction of overexpression of the protein . Thus the N-terminus of TAC102 is not required for import into the mitochondrion and targeting to the TAC but prolonged overexpression of the N-terminally truncated version leads to additional TAC102 foci , some of which associate with ancillary kinetoplasts . The C-terminally truncated TAC102 ( mycΔC-TAC102 ) , on the other hand , does not localize to the mitochondrion or the TAC but accumulates in what is likely the cytoplasm ( Fig 7 , S5 Fig ) , indicating that the C-terminus may be important for proper localization to the organelle . This was also supported by the biochemical digitonin fractionations that showed the majority of TAC102 in the supernatant fraction ( Fig 7 ) . Since the mycΔN-TAC102 localizes to the TAC , we wanted to test if the mutant protein can complement the depletion of the endogenous TAC102 . For this we used RNAi targeting the 3’-UTR of the endogenous TAC102 mRNA ( S2D–S2I Fig ) . In this experiment the ectopically expressed mycΔN-TAC102 or myc-TAC102 mRNAs contained the aldolase 3’-UTR and thus were not affected by RNAi targeting the native TAC102 3’-UTR ( Fig 8 ) . Myc-TAC102 is able to partially rescue the TAC102 RNAi phenotype . The growth defect is delayed by two days , when compared to the cells without complementation ( Fig 8 and S2D Fig ) . Most importantly , even after five days of RNAi induction and simultaneous overexpression the majority of cells ( >98% ) still contain kDNA , albeit in many cases additionally to the proper posterior location also at non-conventional positions within the mitochondrion ( Fig 8 and S6 Fig ) . These ancillary kinetoplasts co-localize with additional TAC102 punctae ( Fig 8 ) . In this cell line , more than 60% of the cells with ancillary kinetoplasts had just one “extra” kDNA structure per cell after five days of induction ( S6B Fig ) . The N-terminally truncated TAC102 ( mycΔN-TAC102 ) , on the other hand , is unable to rescue the kDNA loss phenotype induced by RNAi against the endogenous TAC102 . On day five post induction >60% of cells have lost their kDNA , while 14% show ancillary kDNAs ( Fig 8 and S6 Fig ) . In the cells that have “extra” kDNA , we observe one , two or three ancillary kinetoplasts per cell appearing in the population with similar frequencies , and we also sometimes detect cells with even more “extra” kDNA structures ( S6B Fig ) . In summary , we have shown that the N-terminus of TAC102 is not required for its proper localization to the mitochondrion and the TAC , however it is required for proper function since the N-terminal deletion mutant of TAC102 is unable to rescue the loss of the endogenous protein . Furthermore , even though the full length N-terminally tagged TAC102 partially rescues the TAC102 RNAi phenotype , it leads to extra TAC102 and kDNA foci , indicating that the proper function of the protein is compromised by the myc-tag . Since the C-terminal deletion mutant of TAC102 ( mycΔC-TAC102 ) mislocalized to the cytoplasm , we hypothesized that the targeting signal for mitochondrial import of TAC102 is in the C-terminal 301 aa of the protein . This was supported by bioinformatics analysis that predicted the last 18 and 36 aa of TAC102 to form amphipathic helices ( S7 Fig ) , a hallmark of N-terminal targeting sequences and one C-terminally targeted protein in yeast [39–41] . Furthermore , the last 116 aa of TAC102 are highly conserved among Kinetoplastea supporting the hypothesis that this region might be important for the function of the protein . Thus in an attempt to investigate the potential role of the C-terminal part of TAC102 in mitochondrial import , we created four PCF cell lines expressing different parts of the C-terminus of TAC102 that were C-terminally fused to GFP . As a positive control , we used GFP with a known N-terminal targeting sequence . We induced and followed the expression of the chimeric proteins by fluorescence microscopy and western blotting after digitonin fractionation ( S8 Fig ) . While the GFP containing the N-terminal mitochondrial targeting signal was imported into the mitochondrion , we found mostly cytosolic localization when we added 18 , 36 , or 116 of the TAC102 C-terminus to the GFP and induced expression of the protein overnight . However , when we added the last 301 aa of the C-terminus of TAC102 to GFP , we noticed ( i ) co-localization with the mitochondrial marker ATOM and ( ii ) that the mitochondrial network morphology was compromised . In order to further investigate this observation , we expressed this GFP construct for a shorter period of time ( Fig 9 ) . After 1 and 2 hours of induction , individual cells started to express the GFP chimera that co-localized with the mitochondrial marker but no change in mitochondrial morphology could be detected . However , as early as 3 and 4 hours post expression of the GFP chimera the organelle morphology started to change . We were unable to corroborate immunofluorescence microscopy data by digitonin fractionations , as only few cells expressed GFP-301aa at these early time points ( S9 Fig ) and the overall level of expression was below detection by western blotting . Thus overexpression of the 301 C-terminal amino acids from TAC102 fused with GFP eventually leads to changes in the morphology of the mitochondrial organelle , while the protein localizes , at least partially , to the organelle . From these experiments we conclude that the C-terminus is important for localization to the mitochondrial organelle but is unable to target the protein to the TAC .
RNAi against TAC102 in BSF and PCF cells ( Fig 1 , S2 Fig ) leads to missegregation of the kDNA and eventually to kDNA loss in a large part of the population . While DAPI staining reveals a dramatic loss of kDNA already two days post induction of RNAi in BSF cells ( >75% 0k1n cells ) , the detection of minicircles by Southern blotting shows a much less dramatic loss at this time . The apparent discrepancy can be explained by ( i ) the increase of kDNA in some cells that retain the kinetoplast ( Fig 1D ) and ( ii ) the appearance of small kDNA structures ( Figs 1B and 2 ) , which sometimes may be too small to be identified on microscopic images of DAPI stained cells but still can be detected by Southern blotting . The effect of RNAi against TAC102 on kDNA segregation is further illustrated by our TEM experiments ( Fig 2 ) . Upon loss of TAC102 , we observe enlarged kDNAs , retaining the striated structure and the overall kinetoplast morphology seen in wild type cells . Sometimes additional small but well structured kDNAs were attached to the non-segregated networks , similarly to what has been described for p166 and TAC40 [26 , 27] . Interestingly , we also observed small patches of electron dense material ( edm ) in some cells . We assume that these could be the “small” kinetoplasts , which we also detect in epifluorescence microscopy upon the knockdown of TAC102 ( Fig 1B ) . The edm does not have the typical appearance and the structure of kDNA and is never seen together with regular , well-structured kDNAs in the same cell . We speculate that , when the kDNA size is reduced past a certain threshold after improper segregation , the network loses some structural proteins and is not able to condense properly . We have shown that TAC102 , similarly to p166 and TAC40 , is associated with isolated flagella from BSF and PCF cells , indicating that the protein is tightly bound to the TAC ( Fig 5 , S3 Fig ) . However , in biochemical fractionations with increasing concentrations of digitonin the protein is readily soluble , similarly to a typical matrix protein ( Fig 4E ) . So how can TAC102 remain at flagella upon treatment with a strong detergent such as Triton X-100 , used for flagellar isolation , and be relatively soluble upon digitonin fractionation ? We speculate that TAC102 forms a multimeric structure in the TAC which is partially soluble , probably due to a different degree of association with other components . Alternatively , the difference in the solubilization by digitonin and Triton X-100 might also be explained by the difference in the chemical nature of the detergents . Based on the RNAi studies , TAC102 is a component of the mitochondrial genome segregation machinery and loss of TAC102 does not alter mitochondrial morphology ( Fig 1B ) or the ability to properly segregate the organelle during cell division , similarly to what we recently described for TAC40 [26] , an outer mitochondrial membrane component of the TAC . Thus , mitochondrial genome segregation and organelle segregation are two independent processes and failure to properly segregate the mitochondrial genome does not directly impact cell division . The apparently exclusive function of TAC102 in kDNA segregation is supported by our experiments in the γL262P mutant trypanosomes ( Fig 3 ) , a cell line that is able to compensate for mitochondrial genome loss through a mutation in the γ-subunit of the ATP synthase [37] . When depleted of TAC102 , the γL262P cell line shows the kDNA loss phenotype as described above , however , without any growth defect . This demonstrates that in the γL262P TAC102 RNAi cells no essential function is compromised . The exclusive function of TAC102 and the recently described protein TAC40 [26] in kDNA segregation is surprising since in yeast , for example , all known segregation factors are also involved in other functions , including maintenance of ER-mitochondrial contact sites and organelle morphology [17–19] . This makes trypanosomes a very attractive model system to study mitochondrial DNA segregation , as components of the kDNA segregation machinery are unlikely to be implicated in other processes . Interestingly , the phenotypes that are observed upon TAC102 , p166 , TAC40 or p197 depletion are very similar in kinetics and extent of kDNA loss [25–27] . Additionally , for TAC102 and TAC40 it has now been demonstrated that their function is exclusively associated with mitochondrial genome maintenance . Based on the similarities in loss of function phenotypes we speculate that the same is true for p197 and p166 . Thus these four components of the TAC behave differently from the previously described TAC-associated components ACP , AEP-1 or α-KDE2 that are also required for proper kDNA maintenance but either have additional functions , like α-KDE2 [33] , or are more indirectly involved in the segregation , like ACP that is crucial for proper lipid biogenesis [42] . AEP-1 is a special case since it is the only mitochondrially encoded component reported so far and , while it showed enriched localization at the TAC , it was also detected throughout the mitochondrial organelle , hinting at other potential functions of this protein [31 , 32] . Thus we consider TAC102 , TAC40 , p166 and p197 to be the “core” components of the TAC . The TAC102 RNAi experiments show that enlarged kDNAs mostly associate with the old basal body and cells that retain two kinetoplasts ( 2k2n cells ) , although very rare , mostly show unequal segregation of the mitochondrial genomes ( S2I Fig ) , a phenomenon that was previously also observed in cells depleted of p166 [27] . The specific loss of the kDNA–new basal body connection argues that the old basal body−kDNA connection is , at least initially , not affected by RNAi targeting TAC102 or p166 . Thus the TAC that is associated with the new basal body is likely assembled as a de novo structure rather than replicated in a semi-conservative mode , in which case we would have expected a random loss of the basal body−kDNA connection . This is further corroborated by our experiments demonstrating that the loss of TAC102 upon RNAi preferentially occurs at the new basal body−kDNA connection , leaving the old basal body−kDNA connection intact ( S4 Fig ) . A peculiar observation was the small ancillary kDNAs that appeared upon expression of several tagged TAC102 constructs ( myc:ΔN-TAC102; myc:ΔN-TAC102 and myc-tagged full length TAC102 both in the absence of the endogenous protein ) . These results suggest that the N-terminus of TAC102 is important for the connection to upstream components of the TAC ( closer to the basal body ) and its deletion or obstruction by a tag leads to a partial loss of function of the protein , namely the proper localization at the TAC . Furthermore , the results indicate that the C-terminal part of the protein in the absence of the N-terminus is sufficient to either directly connect to kDNA or initiate the assembly of the downstream components that connect to the kDNA and thus initiate the appearance of the ancillary kinetoplasts that have lost the connection to the upstream components of the TAC and , consequently , the basal body . Ancillary kinetoplasts naturally occur in several Kinetoplastea species but the frequency of their appearance in T . brucei is very low and under normal culture conditions these “extra” kDNAs are rarely detected [43] . However , there are examples where depletion or overexpression of mitochondrial proteins leads to additional kDNA structures in the mitochondrion . In Tim17 RNAi cells up to 10% of the population accumulate extra kDNA [44] , but since Tim17 is a protein import component , this effect is likely indirect and occurs due to the loss of kDNA segregation/replication factors , e . g . POLIB and POLIC , which , if depleted by RNAi , also lead to the appearance of ancillary kDNA structures [45] . The depletion of the mitochondrial acyl carrier protein ( ACP ) , a key component of the fatty acid synthesis pathway , was also shown to produce “extra” kDNA structures , which could be due to the loss of the “special” membrane structures in the TAC [42] . On the other hand , overexpression of PUF9 target 1 ( PNT1 ) , a mitochondrial protein of unknown function that localizes to the kDNA , is also able to induce ancillary kDNA appearance [46] . Based on biochemical fractionations and fluorescence microscopy , including super-resolution microscopy ( STED ) , as well as previously published proteomics data [47] , TAC102 is a mitochondrial protein that localizes to the posterior region of the mitochondrial organelle between the basal body and the kDNA ( Figs 4 , 5 and 6; S3 and S4 Figs ) . Interestingly , TAC102 does not contain a classical N-terminal mitochondrial targeting signal but , based on the ΔC-mutant analysis , rather a region within the C-terminus seems to be involved in proper localization to the mitochondrial organelle in vivo ( Figs 7 , 8 and 9; S5 and S8 Figs ) . This was further supported by the bioinformatics analysis that predicted the presence of amphipathic helices at the C-terminus of TAC102 ( S7 Fig ) , similarly to the yeast helicase Hmi1p that was shown to be imported into mitochondria via a C-terminal targeting sequence [39] . To test the hypothesis , we designed a series of cell lines expressing GFP C-terminally fused with C-terminal sequences of TAC102 of different length . We reasoned that , if there is a targeting signal within the C-terminus , we should be able to target the GFP chimeras to the mitochondrion . However , the 18 and 36 aa of the C-terminus of TAC102 are not sufficient to target chimeric GFP proteins to the mitochondrial organelle . Their failure to properly localize could be explained by misfolding of the GFP chimeras . Nonetheless , they are detectable in the cytoplasm without any apparent degradation ( S8 Fig ) , which we would not expect for misfolded proteins . Furthermore , amphipathic helices , in general , are readily transferable to the N-terminus of GFP without any deleterious effect ( for example , see positive control , Ntarget-GFP , in S8 Fig ) . Thus the more likely explanation is that the predicted amphipathic helix is not sufficient for proper import in vivo and additional internal regions of TAC102 are required for this function . This is supported by the fact that the C-terminal 301 amino acids C-terminally fused to GFP are able to target the chimera to the mitochondrion ( Fig 9 , S8 Fig ) , although the expression of GFP-301aa leads to strong vesiculation of the mitochondrial network when expressed for more than two hours ( Fig 9 , S8 Fig ) . We hypothesize that GFP-301aa interacts with the mitochondrial outer membrane and either blocks mitochondrial protein import or alternatively impacts one or several of the mitochondrial morphology factors , which are not well characterized in trypanosomes . The only mitochondrial protein in trypanosomes that has been described to harbor an internal targeting sequence is the Trypanosome Alternative Oxidase ( TAO ) [48] . TAO is a mitochondrial inner membrane protein equipped with an N-terminal as well as an internal targeting sequence both of which are sufficient to target the protein . In other systems , e . g . yeast , most mitochondrial inner membrane proteins do not possess a classical N-terminal targeting sequence , and several , like the cytochrome c heme lyase or BCS1 , have been described to harbor internal targeting signals [49–51] . This raises the question if TAC102 is localized in the inner mitochondrial membrane or even the intermembrane space ? Based on bioinformatics analysis there is no predicted transmembrane domain in TAC102 and biochemical digitonin fractionations performed with different concentrations of the detergent demonstrate that the protein behaves like LipDH , a mitochondrial matrix protein , whereas membrane-associated proteins like the integral outer membrane protein ATOM or the inner membrane protein COXIV require more stringent extraction conditions for their solubilization ( Fig 4E ) . Thus from our current data we conclude that TAC102 is not an inner mitochondrial membrane protein but rather localizes to the unilateral filaments between the kDNA and the mitochondrial inner membrane .
Bloodstream form T . brucei cells were cultured in HMI-9 medium with 10% FCS at 37°C and 5% CO2 , and procyclic trypanosomes were maintained in SDM-79 medium with 10% FCS at 27°C . For transfections , the New York single marker ( NYsm ) or the γL262P strains of BSF T . brucei and the 29–13 strain of PCF T . brucei were used . Cells were transfected with NotI-linearized plasmids by electroporation and then selected with appropriate antibiotics , by limiting dilutions . NYsm BSF and 29–13 PCF trypanosomes were obtained from the established collection of the Institute of Cell Biology , University of Bern , Bern , Switzerland . The γL262P strain of BSF cells is a kind gift of A . Schnaufer . The TAC102 RNAi constructs were targeted against the 451−1021 bp of the ORF of the gene Tb927 . 7 . 2390 or against the 147−646 bp of the 3’-UTR of this gene . Briefly , a PCR fragment with adaptor sequences was amplified from genomic DNA of NYsm BSF cells , and cloned in two steps into the pTrypRNAiGate vector by Gateway cloning . For this the full-length sequence of the gene ( 1−951 bp ) or the ΔN sequence ( 200−951 bp ) or the ΔC sequence ( 1−650 bp ) were amplified by PCR from genomic DNA of NYsm BSF trypanosomes and cloned into the pJM-2 vector ( gift of A . Schneider ) . The final plasmids were used for transfection as described above . Expression was induced by addition of 1 μg/ml tetracycline . For N-terminal PTP-tagging of TAC102 , the ORF positions 4 to 707 were amplified from genomic DNA and cloned between ApaI/NotI sites of pN-PURO-PTP vector . The resulting plasmid was linearized with XbaI prior to transfection . This construct was recombined into the endogenous locus to substitute for one of the TAC102 alleles and thus was constantly expressed . For GFP-301aa , GFP-116aa and GFP-36aa constructs , the respective parts of TAC102 ( PCR products ) and GFP ( PCR product ) were fused together by fusion PCR and ligated between HindIII/BamHI sites into the pFS-3 expression plasmid ( gift of A . Schneider ) . For the GFP-18aa construct , GFP was PCR amplified with a reverse primer that contained the sequence of the last 18 aa of TAC102 and cloned into pFS-3 . For the Ntarget-GFP construct , the N-terminal mitochondrial targeting sequence of the Rieske iron-sulfur protein ( Tb927 . 9 . 14160 , 1−72 bp ) was ligated between XhoI/AgeI sites into the pG-EGFP-ΔLII vector [52] , then the obtained Ntarget-GFP sequence was cut out by HindIII/BamHI and cloned into pFS-3 . All GFP constructs were linearized with NotI and transfected into 29–13 PCF cells as described above . Expression was induced by addition of 1 μg/ml tetracycline . The recombinant version of TAC102 was expressed in E . coli BL21 strain as a fusion with the maltose-binding protein ( MBP ) at the N-terminus of TAC102 , using the pMAL system ( New England Biolabs ) . The fusion protein was purified on amylose resin and analyzed by mass-spectrometry . The purified product was used to generate polyclonal antibodies in rats ( Eurogentec , Belgium ) . The monoclonal antibody was produced in mice ( GenScript , USA ) against a synthetic peptide that represented the 500−660 aa of TAC102 . Specificity of both antibodies was confirmed by western blotting and immunofluorescence microscopy , in PCF and BSF cells ( S2 Fig ) . SDS-PAGE was carried out as described elsewhere , in 8% , 10% or 15% SDS-polyacrylamide gels . The gels were either stained with Coomassie blue R250 or , for western analysis , transferred onto PVDF membranes . Blocking was performed in 5% or 10% skimmed milk solution in PBS or PBST ( PBS + 0 . 1% TWEEN-20 ) . Primary antibodies were: mouse monoclonal anti-TAC102 ( 1:1000 , GenScript ) , rat polyclonal anti-TAC102 ( 1:1000 , Eurogentec ) , rabbit anti-Protein A ( 1:5000 , Sigma ) , rabbit anti-myc ( 1:1000 , Sigma ) , mouse anti-myc ( 1:1000 , Sigma ) , mouse anti-EF1α ( 1:10000 , SantaCruz ) , rabbit anti-ALBA3 ( 1:1000 , [53] ) , rabbit anti-ATOM ( 1:10000 , [54] ) , mouse anti-GFP ( 1:1000 , Sigma ) , rabbit anti-GFP ( 1:1000 , Sigma ) , rabbit anti-COXIV ( 1:1000 ) , rabbit anti-LipDH ( 1:10000 , [55] ) . Secondary antibodies were: mouse anti-rabbit HRP-conjugate ( 1:10000 , Dako ) , rabbit anti-mouse HRP-conjugate ( 1:10000 , Dako ) , swine anti-rabbit HRP-conjugate ( 1:10000 , Dako ) , goat anti-rat 680 LT ( 1:10000 , LI-COR ) , goat anti-mouse 800 CW ( 1:10000 , LI-COR ) , goat anti-rabbit 680 LT ( 1:10000 , LI-COR ) . Cells were collected by centrifugation at 2500 rcf for 8 min at room temperature , washed once with PBS , re-suspended in SoTE buffer ( 0 . 6 M sorbitol , 2 mM EDTA , 20 mM Tris-HCl , pH 7 . 5 ) such that 107 cells were in 25 μl of the buffer , and an equal volume of 0 . 05% digitonin solution in SoTE buffer was added . Alternatively , to use other digitonin concentrations ( Fig 4E ) , the PBS-washed cells were directly re-suspended in SoTE containing the necessary amount of digitonin , to the final volume . The cells were then incubated on ice for 5 min and then centrifuged at 8000 rcf for 5 min at 4°C . The supernatant ( cytosolic fraction ) was separated from the pellet ( mitochondria ) and both fractions were lysed in Laemmli buffer . Total RNA was extracted from trypanosomes with RiboZol ( Amresco ) and separated in 1 . 4% agarose gels with 6% formaldehyde and transferred onto nylon membranes in 10×SSC . The probe for TAC102 ORF was generated from a PCR fragment that had been used for creation of the RNAi construct , by incorporation of α-P32-dCTP using RadPrime DNA Labeling System ( Invitrogen ) . Blots were re-probed for 18S rRNA to ensure equal loading of samples . 18S rRNA probe was generated by T4 PNK labelling of an oligonucleotide ( which is complementary to 18S rRNA ) with γ-P32-ATP . Total DNA was extracted from trypanosomes with phenol/chloroform as described elsewhere , and digested overnight at 37°C with HindIII and XbaI . Reaction mixtures were separated in 1% agarose gels in 1×TAE buffer . After this , the gels were washed twice for 10 min in depurination solution ( 0 . 25 M HCl ) , once for 30 min in denaturation solution ( 1 . 5 M NaCl , 0 . 5 M NaOH ) , twice for 15 min in neutralization solution ( 3 M NaCl , 0 . 5 M Tris-HCl , pH 7 . 5 ) , twice for 15 min in 20×SSC and then transferred onto nylon membranes in 20×SSC . The probe for minicircles was generated from a PCR fragment ( approx . 100 bp of the conserved minicircle sequence ) amplified from total DNA of NYsm BSF T . brucei , by incorporation of α-P32-dCTP using RadPrime DNA Labeling System ( Invitrogen ) . Blots were re-probed for the intergenic region between α- and β-tubulin for normalization . The tubulin probe was generated from a corresponding PCR fragment amplified from total DNA of NYsm BSF T . brucei , by incorporation of α-P32-dCTP and α-P32-dATP using RadPrime DNA Labeling System ( Invitrogen ) . Southern analysis was repeated three times . BSF or PCF cells were fixed on slides with 4% PFA in PBS , permeabilized for 5 min with 0 . 2% TritonX-100 in PBS and blocked for 30 min with 4% BSA in PBS . Primary and secondary antibodies were diluted in 4% BSA in PBS . Primary antibodies were: mouse monoclonal anti-TAC102 ( 1:1000 ) , rat anti-TAC102 ( 1:1000 ) , rabbit anti-Protein A ( 1:1000 , Sigma ) , rat YL1/2 ( 1:2000 ) , mouse Mab22 ( 1:10 ) , rat anti-PFR ( 1:1000 , [56] ) , mouse BBA4 ( 1:100 ) , rabbit anti-myc ( 1:1000 , Sigma ) , mouse anti-myc ( 1:1000 , Sigma ) , mouse anti-GFP ( 1:100 , Sigma ) , rabbit anti-GFP ( 1:1000 , Sigma ) , mouse anti-mtHSP70 ( 1:2000 , [57] ) . The following secondary antibodies ( 1:1000 , Invitrogen ) were used: goat anti-rabbit IgG , goat anti-rat IgG , goat anti-mouse IgG conjugated with fluorophores Alexa Fluor 488 , Alexa Fluor 594 , Alexa Fluor 647 . Cells were mounted with VECTASHIELD Mounting Media with DAPI ( Vector Laboratories ) or ProLong Gold Antifade Mountant with DAPI ( Invitrogen ) . Images were acquired with the Leica DM 5500 fluorescent light microscope and deconvolved by the Leica LAS AF software . For evaluation of kDNA intensities , ImageJ software was used . Trypanosomes in medium with 5 mM EDTA were centrifuged and re-suspended in extraction buffer ( 10 mM NaH2PO4 , 150 mM NaCl , 1 mM MgCl2 ) containing 0 . 5% TritonX-100 , on ice . After one washing step with extraction buffer , cells were incubated on ice for 45 min in extraction buffer containing 1 mM CaCl2 and then subjected to immunofluorescence analysis ( IFA ) . For super-resolution microscopy cells were stained with 200 nM MitoTracker Red CMXRos ( Thermo Fisher ) for 20 min at 37°C . The following staining procedure was performed as described above . The primary antibody was rabbit anti-Protein A ( 1:1000 , Sigma ) and the secondary antibody was goat anti-rabbit IgG Oregon Green 488 ( 1:100 , Invitrogen ) . Cells were mounted with ProLong Gold Antifade mounting solution ( Invitrogen ) . The images were acquired with Leica SP8 Confocal Microscope System with STED and deconvolved with Huygens professional software . Trypanosomes were grown as described above , harvested and centrifuged at 3345 rcf for 5 min and the pellets were submerged with fixative which was prepared as follows: 2 . 5% glutaraldehyde ( Agar Scientific , UK ) in 0 . 15 M HEPES ( Fluka , Switzerland ) with osmolarity of 684 mOsm and adjusted to pH 7 . 41 . The cells remained in the fixative at 4°C for at least 24 hours before further processing . They were then washed with 0 . 15 M HEPES two times for 5 min , post-fixed with 1% OsO4 ( SPI Supplies , West Chester , USA ) in 0 . 1 M Na-cacodylate buffer ( Merck , Germany ) at 4°C for 1 h , washed with 0 . 05 M maleate-NaOH buffer ( Merck , Germany ) three times for 5 min , and then block-stained in 0 . 5% uranyl acetate ( Fluka , Switzerland ) in 0 . 05 M maleate-NaOH buffer at 4°C for 1 h . Then the cells were washed in 0 . 05 M maleate-NaOH buffer three times for 5 min and dehydrated in 70 , 80 , and 96% ethanol ( Alcosuisse , Switzerland ) for 15 min each at room temperature . Subsequently , the cells were immersed in 100% ethanol ( Merck , Germany ) three times for 10 min , in acetone ( Merck , Darmstadt , Germany ) two times for 10 min , and finally in acetone-Epon ( 1:1 ) overnight at room temperature . The next day , cells were embedded in Epon ( Fluka , Switzerland ) and left to harden at 60°C for five days . Sections were produced with an ultramicrotome UC6 ( Leica Microsystems , Vienna , Austria ) , first–semi-thin sections ( 1 μm ) for light microscopy , which were stained with solution of 0 . 5% toluidine blue O ( Merck , Darmstadt , Germany ) , and then–ultrathin sections ( 70−80 nm ) for electron microscopy . The sections , mounted on 200 mesh copper grids , were stained with uranyl acetate and lead citrate with an ultrostainer ( Leica Microsystems , Austria ) . Sections were then examined with a transmission electron microscope ( CM12 , Philips , Eindhoven ) equipped with a digital camera ( Morada , Soft Imaging System , Germany ) . | Proper segregation of the mitochondrial genome during cell division is a prerequisite of healthy eukaryotic cells . However , the mechanism underlying the segregation process is only poorly understood . We use the single celled parasite Trypanosoma brucei , which , unlike most model organisms , harbors a single large mitochondrion with a single mitochondrial genome , also called kinetoplast DNA ( kDNA ) , to study this question . In trypanosomes , kDNA replication and segregation are tightly integrated into the cell cycle and thus can be studied alongside cell cycle markers . Furthermore , previous studies using electron microscopy have characterized the tripartite attachment complex ( TAC ) as a structural element of the mitochondrial genome segregation machinery . Here , we characterize TAC102 , a novel trypanosome protein localized to the TAC . The protein is essential for proper kDNA segregation and cell growth . We analyze the presence of this protein using super resolution microscopy and show that TAC102 is a mitochondrial protein localized between the kDNA and the basal body of the cell’s flagellum . In addition , we characterize different parts of the protein and show that the C-terminus of TAC102 is important for its proper localization . The data and resources presented will allow a more detailed characterization of the dynamics and hierarchy of the TAC in the future and might open new avenues for drug discovery targeting this structure . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"rna",
"interference",
"nuclear",
"staining",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"mitochondria",
"epigenetics",
"bioenergetics",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
... | 2016 | TAC102 Is a Novel Component of the Mitochondrial Genome Segregation Machinery in Trypanosomes |
Cell fate choices are tightly controlled by the interplay between intrinsic and extrinsic signals , and gene regulatory networks . In Saccharomyces cerevisiae , the decision to enter into gametogenesis or sporulation is dictated by mating type and nutrient availability . These signals regulate the expression of the master regulator of gametogenesis , IME1 . Here we describe how nutrients control IME1 expression . We find that protein kinase A ( PKA ) and target of rapamycin complex I ( TORC1 ) signalling mediate nutrient regulation of IME1 expression . Inhibiting both pathways is sufficient to induce IME1 expression and complete sporulation in nutrient-rich conditions . Our ability to induce sporulation under nutrient rich conditions allowed us to show that respiration and fermentation are interchangeable energy sources for IME1 transcription . Furthermore , we find that TORC1 can both promote and inhibit gametogenesis . Down-regulation of TORC1 is required to activate IME1 . However , complete inactivation of TORC1 inhibits IME1 induction , indicating that an intermediate level of TORC1 signalling is required for entry into sporulation . Finally , we show that the transcriptional repressor Tup1 binds and represses the IME1 promoter when nutrients are ample , but is released from the IME1 promoter when both PKA and TORC1 are inhibited . Collectively our data demonstrate that nutrient control of entry into sporulation is mediated by a combination of energy availability , TORC1 and PKA activities that converge on the IME1 promoter .
Cellular differentiation programs are controlled by environmental and cell intrinsic events . How cells integrate multiple stimuli to regulate cell fate choice is poorly understood . The yeast Saccharomyces cerevisiae is an ideal model to study this problem . In response to multiple , well-defined signals , yeast cells induce a differentiation program to form four haploid gametes or spores [1 , 2] . Gametogenesis or sporulation is characterized by a specialized cell division called meiosis . During sporulation diploid cells undergo a single round of DNA replication followed by two consecutive nuclear divisions , meiosis , to generate progeny containing half the number of chromosomes of the diploid parent cell . The initiation of gametogenesis is controlled by cell-intrinsic and cell-extrinsic signals , which together regulate a single master transcription factor called inducer of meiosis I , IME1 [3 , 4] . In cells expressing a single mating type , MATa or MATα , IME1 is repressed by transcription coupled repression of the IME1 promoter involving the long noncoding RNA IRT1 [5] . In MATa/α diploid cells Rme1 , the transcriptional activator of IRT1 , is repressed . As a consequence these cells express IME1 upon nutrient deprivation [6] . For efficient IME1 induction a fermentable carbon source and nitrogen need to be absent from the growth medium . Under these conditions cells produce ATP via oxidative phosphorylation to facilitate IME1 expression [7 , 8] . Two conserved signalling pathways have been implicated in nutrient regulation of IME1 expression . First , the presence of glucose in the growth medium activates the Ras/cAMP-dependent Protein Kinase A ( PKA ) pathway , which in turn inhibits IME1 and entry into sporulation [9 , 10] . The second regulator of IME1 is the target of rapamycin complex I ( TORC1 ) . TORC1 promotes macromolecule biosynthesis in response to nitrogen and amino acid availability [11] . When nitrogen sources/amino acids are ample , TORC1 is active and inhibits IME1 and sporulation [7 , 12] . Whether PKA and TORC1 are the main mediators of nutrient control of IME1 , and how the two pathways control entry into sporulation is not well understood . Here we describe how nutrients control IME1 expression . We find that PKA and TORC1 signalling account for the majority of IME1 regulation by nutrients . Inhibition of PKA and TORC1 activity is sufficient to induce IME1 expression even in the presence of high levels of nutrients . Under these conditions , cells induce IME1 , complete meiosis , and generate spores with kinetics that are highly reminiscent of those observed under starvation conditions . The ability to induce sporulation in the presence of ample nutrients further allowed us to investigate the importance of respiration and TORC1 activity for the induction of gametogenesis . We find that respiration and fermentation are interchangeable for IME1 induction . Both metabolic pathways can serve as energy providers during entry into sporulation . Our analysis further shows that intermediate levels of TORC1 activity are critical for gametogenesis . When TORC1 is fully active or completely inhibited , IME1 is repressed . Finally , we show that the transcriptional repressor Tup1 binds to and represses the IME1 promoter when TORC1 and/or PKA are active , but not when both pathways are inhibited . Importantly , depletion of Tup1 is sufficient to mimic starvation-induced IME1 expression . Our data demonstrate that nutrient control of sporulation is sensed and orchestrated by TORC1 and PKA signalling pathways and by the availability of energy .
In budding yeast nutrient availability determines whether cells enter sporulation . The PKA and TORC1 pathways as well as respiration have been linked to the regulation of IME1 expression by nutrients and to entry into sporulation ( Fig 1A ) [1] . To determine whether TORC1 and PKA are the major mediators of nutrient sensing in triggering sporulation , we examined how inactivation of either or both pathways affects IME1 expression . TORC1 can be rapidly and efficiently inhibited using the small molecule rapamycin that reduces cell proliferation rate significantly ( S1A Fig ) . Inhibition of the PKA pathway is more complex because budding yeast encodes three redundant genes encoding the catalytic subunits of PKA , TPK1 , TPK2 , and TPK3 [13] . To inhibit the PKA pathway , we generated an ATP analogue sensitive strain of PKA that we define as tpk1-as . The strain contains gene deletions in TPK2 , TPK3 and a point mutation in TPK1 ( tpk1M164G ) that transforms this allele into an ATP analog sensitive ( as ) allele [14] . In the tpk1-as strain , PKA activity can be specifically blocked using the ATP analog 1NM-PP1 , which results in a growth arrest ( Fig 1B and S1A Fig ) . To measure IME1 promoter activity in response to modulating PKA activity , we used an IME1-promoter LacZ reporter fusion ( pIME1-LacZ ) that was integrated at the IME1 locus without disrupting the endogenous IME1 . This fusion serves as an accurate readout for IME1 promoter activity [7] . When we shifted control and tpk1-as diploid cells from rich medium containing glucose ( YPD ) to sporulation medium ( SPO ) , a condition which induces IME1 , β-galactosidase activity increased ( S1B Fig and Fig 1C ) . The β-galactosidase levels were comparable between the two strains suggesting that tpk1-as allele does affect IME1 regulation in SPO ( S1B Fig ) . As expected , IME1 promoter activity did not increase when tpk1-as cells were shifted to fresh YPD ( Fig 1C ) . Using the tpk1-as and pIME1-LacZ system , we first determined whether glucose repression of IME1 is mediated by PKA signalling . Cells were shifted from YPD to SPO , or to SPO medium containing glucose in the presence or absence of the ATP analog 1NM-PP1 ( Fig 1D ) . IME1 promoter activity was strongly reduced in the presence of glucose . In contrast , when PKA was inhibited IME1 promoter activity was comparable with cells grown in the absence of glucose . This result shows that glucose inhibits IME1 expression predominantly via the PKA signalling pathway . The presence of a nitrogen source also prevents IME1 expression [3] . To test whether TORC1 signalling is responsible for IME1 repression by nitrogen sources and amino acids we examined the effects of rapamycin on IME1 expression . To exclude the effects of glucose repression on IME1 , we used a nitrogen and amino acid rich medium containing the non-fermentable carbon source acetate ( YPA ) but lacking a fermentable carbon source . We found that IME1 promoter activity slightly increased in cells shifted from YPD to YPA , and , inhibition of PKA did not further increase IME1 expression ( Fig 1E and S1D Fig ) . This is expected because it is known that Ras/PKA transmits the glucose signal and thus glucose levels control IME1 via PKA [3 , 9 , 10 , 15 , 16] . When we inhibited TORC1 , by treating cells grown in YPA medium with rapamycin , IME1 was rapidly induced . The majority of cells ( 95% ) were also able to form spores within 24 hours ( S1 Table ) . We conclude , as reported previously , that the PKA pathway transmits the glucose signal to the IME1 promoter , and that TORC1 most likely transmits the nitrogen signal [3 , 9] . To examine whether PKA and TORC1 are the major mediators of nutrient control of IME1 expression , we inhibited either or both pathways in cells grown in rich medium containing glucose ( YPD ) ( Fig 1F and S1C Fig ) . Inhibition of TORC1 had no effect on IME1 expression . In contrast , IME1 promoter activity strongly increased between 8 to 12 hours following treatment with PKA inhibitors . Interestingly , when both PKA and TORC1 were inhibited , IME1 induction was already noticeable at 4 hours , and peaked at 8 hours and was remarkably similar to levels seen in cells incubated in SPO medium ( Fig 1F and S1 Fig ) . Similar results were obtained when IME1 mRNA levels were examined ( Fig 1G ) . These data show that the combined inhibition of PKA and TORC1 activities is sufficient to mimic nutrient control of IME1 expression . We conclude that TORC1 and PKA are two major mediators of nutrient regulation of IME1 expression . Our results show that inhibition of PKA leads to some degree of IME1 expression in rich medium ( YPD ) ( approximately 50% of that observed in SPO medium; Fig 1F and 1G ) . One explanation for this observation is that PKA inhibition induces IME1 at low or intermediate levels in all cells . It is also possible that IME1 induction occurs only in a subpopulation of cells when PKA is inhibited . To distinguish between these possibilities , we measured the distribution of IME1 expression in cells by single molecule RNA fluorescence in situ hybridization ( smFISH ) ( Fig 2A ) . The technique can reliably measure absolute transcript levels in single cells [17] . To ensure that the signals were specific and probes entered the cells , we first measured IME1 and ACT1 transcript levels in wild-type and ime1Δ diploid cells that were induced in SPO medium ( S2 Fig ) . While IME1 was expressed in the wild type , no transcripts were detected in ime1Δ cells . As expected , ACT1 levels were comparable between the two strains ( S2B Fig ) . Next , we counted the mean number of IME1 transcripts in cells grown in rich medium shifted to SPO medium , or treated with PKA or PKA/TORC1 inhibitors ( Figs 2A and 1B ) . The IME1 expression pattern matched the RT-PCR experiment ( compare Figs 1G and 2B ) . In cells treated with the PKA inhibitor IME1 levels increased after 8 hours to about 10 copies per cell on average . When both TORC1 and PKA were inhibited , IME1 was transcribed efficiently and cells contained on average 30 copies per cell at 8 hours after treatment , which was comparable to IME1 levels in SPO medium ( Fig 2B ) . It is worth noting that IME1 mRNA levels decline sharply 12 hours after inhibition of the PKA and TORC1 pathways but remained elevated in cells incubated in SPO medium for the same amount of time ( Fig 2B ) . Given that expression of IME1 is known to decline when cells undergo meiotic divisions , a plausible explanation is that progression into meiosis differs between the two conditions [18] . Indeed , when PKA and TORC1 were inhibited the majority of cells underwent meiotic divisions within 12 hours ( see next section for details ) . In contrast , when cells were directly transferred from YPD medium into SPO medium sporulation did not occur efficiently , and it is likely that many ( more than 50% ) of the cells were arrested in intermediate stages of meiosis ( S1 Table ) . Therefore a decline in IME1 was not observed at 12 hours ( Fig 2B ) . Finally , we would like to point out that we did not see a decline in pIME1-LacZ reporter activity even when cells progressed into meiosis ( Fig 1F ) . This can be explained by the long half-life of β-galactosidase ( more than 20 hours ) [19] . Next , we analysed IME1 abundance in single cells . We binned the single cell expression data into five classes of transcript levels ( 0–5 , 6–10 , 11–15 , 16–20 and 21 or more transcripts ) . In untreated cells more than 95 percent of cells had no or low levels of IME1 , whereas ACT1 was expressed strongly in the majority cells ( Fig 2C ) . In line with previous observations , rapamycin treatment had no effect on IME1 expression ( Fig 2D ) . Interestingly , when PKA was inhibited the majority of cells expressed no or low levels of IME1 ( 0–5 copies per cell ) , but approximately 20% of cells expressed high levels of IME1 ( 21 or more copies per cell ) ( Fig 2E ) . When both PKA and TORC1 were inhibited the majority of cells ( more than 75% ) harboured high levels of IME1 RNA , which was comparable to SPO medium ( compare Fig 2F and 2G ) . These data complement our population based assays and show that inhibition of PKA/TORC1 leads to significant IME1 expression in the majority of cells grown in nutrient-rich conditions . Little is known about how nutrient signalling controls other aspects of sporulation . To determine the consequences of PKA and TORC1 inhibition on meiotic progression , we examined subsequent stages of meiosis by measuring the kinetics of meiosis in cells shifted to YPD containing inhibitors of PKA and TORC1 . Interestingly , less than 15% of cells underwent meiotic divisions when only PKA was inhibited . Given that a significant higher portion of cells were positive for intermediate or high levels of IME1 compared to number of cells that underwent meiotic divisions ( Fig 2D and 2E ) , the result suggests that events downstream of IME1 are perhaps not efficiently induced in these cells . More than 90% of cells underwent meiotic divisions when both PKA and TORC1 were inhibited ( Fig 3A ) . Under this condition cells also formed spores , but 70 percent ( 224 out 320 spores ) of spores formed colonies compared to 95 percent ( 304 out of 320 spores ) for the wild-type cells induced to sporulate in SPO medium . This reduced spore viability was not due to the tpk1-as allele being hypomorphic , because the same strain sporulated efficiently and exhibited 94 percent ( 301 out of 320 spores ) spore viability in SPO medium ( Fig 3B ) . Apart from reduced spore viability , we also observed a strong enrichment for triads in cells treated with PKA/TORC1 inhibitors ( Fig 3C ) . Although four DAPI masses formed during the two meiotic divisions , three were packaged into spores and one nucleus was evicted or degraded from cells ( Fig 3C and 3D ) . We conclude that inhibition of TORC1 and PKA is sufficient to drive entry into and progression through the sporulation program . Respiration is needed for IME1 transcription and for entry into sporulation [7 , 8] . However , it is not clear whether IME1 expression is dependent on the energy produced by respiration or whether it requires a signal from active mitochondria . The system we developed to induce sporulation in the presence of ample nutrients allowed us to investigate this question . First , we analysed how IME1 expression is affected in respiratory deficient cells . Pet100 is required for the assembly of cytochrome c oxidase . Yeast cells lacking the PET100 gene cannot respire . In line with previous observations , pet100Δ cells did not induce IME1 in SPO medium ( Fig 4A , compare lanes 1–4 to 5–8 ) [7] . Likewise , cells treated with the drugs antimycin A ( antimycin ) , which inhibits cytochrome c reductase , or oligomycin , which inhibits the Fo subunit of the mitochondrial ATP synthase , did not induce IME1 ( Fig 4B , compare lanes 1–4 to 5–8 and 9–12 ) . Uncoupling of respiration from energy production by treating cells with CCCP which disrupts the proton gradient and thus reduces the ability of the ATP synthase to function led to similar results ( Fig 4B , compare lanes 1–4 and 13–16 ) . Thus , respiration is required for induction of IME1 expression in sporulation medium . Next , we induced sporulation by inhibiting PKA and TORC1 in cells grown in glucose-rich medium , and tested whether respiration is required for IME1 expression . To our surprise , IME1 expression levels were comparable between control cells , and antimycin or oligomycin treated cells ( Fig 4C , compare lanes 1–4 to 5–8 or 9–12 ) . pet100Δ cells grown in YPD strongly induced IME1 when the PKA and TORC1 pathways were inhibited ( Fig 4C , compare lanes 1–4 to 13–16 ) . To further quantify IME1 promoter activity in respiratory deficient cells , we measured pIME1-LacZ reporter expression in oligomycin treated cells . As expected , in SPO medium the IME1 promoter stayed repressed when cells were treated with oligomycin . In YPD plus TORC1 and PKA inhibitors , IME1 expression accumulated with slightly slower kinetics in the presence of oligomycin but peaked to similar levels as control cells ( Fig 4D ) . Finally we examined whether IME1 can be induced from a heterologous promoter in SPO medium when respiration is inhibited ( Fig 4E ) . When we induced IME1 from the GAL1 promoter using a Gal4-estrogen receptor fusion ( GAL4-ER ) that can be activated by the addition of estradiol , IME1 was strongly induced . However , in cells treated with antimycin IME1 stayed repressed . Previous work showed that expression of mRNAs from different genes is also affected under this condition [7] . We propose that the effects are not specific for IME1 , but either transcription or mRNA stability or both are generally affected when cells are starved and cannot respire . Notably , even though cells were able to express IME1 when respiration was inhibited in YPD medium with TORC1 and PKA inhibitors the vast majority of these cells did not complete gametogenesis ( Fig 4F ) indicating that other stages of sporulation require respiration . In conclusion , when sporulation is induced in the presence of ample nutrients , respiration is not required for IME1 expression . This result suggests that either respiration or fermentation can serve as energy providers for induction of IME1 transcription . Our results show that inhibition of PKA and TORC1 activity is sufficient to initiate entry into sporulation . Although it is well established that PKA signalling inhibits sporulation , inhibition of TORC1 by rapamycin treatment has been reported to affect sporulation with different outcomes . We and others have shown that rapamycin can stimulate sporulation by inducing IME1 expression [7] . Moreover , inactivation of TORC1 was shown to stabilize Ime1 and promotes its nuclear localization [20] . However , others have found that rapamycin can also inhibit spore formation in budding and fission yeast when added to the SPO medium [21 , 22] . These seemingly conflicting results prompted us to further examine how rapamycin and TORC1 control sporulation . First , we tested whether there is a concentration dependent effect of rapamycin on cell growth and sporulation . Rapamycin treatment at the concentration which ensures efficient sporulation ( 1000 ng/ml ) diminished , but did not stop cell proliferation ( 190 min versus 90 min in control cells; Fig 5A ) . This observation suggests that TORC1 is still active . When we used 50 fold less rapamycin ( 20 ng/ml ) , cell proliferation was somewhat affected ( 145 min versus 90 min in control cells ) , and cells sporulated efficiently when combined with inhibition of PKA ( Fig 5A , right panel ) . Lower concentrations of rapamycin had no observable effect on growth and sporulation . These results indicate that the TORC1 pathway is not completely blocked upon entry into sporulation and meiosis . In line with our observation that rapamycin does not abolish growth completely , a recent study showed that rapamycin , irrespective of the concentration used , does not fully inhibit TORC1 activity [23] . Depletion or inactivation of the Kog1 subunit of the TORC1 complex , however , causes a complete growth arrest and a starvation response [24 , 25] . We therefore depleted Kog1 using an auxin induced degradation system ( AID ) system [26] and examined the effects on IME1 expression and sporulation . The system relies on Oryza sativa TIR1 ( pTEF1-osTIR1 ) , which interacts with the SCF ubiquitin ligase , and the chemical indole-3-acetic acid ( IAA ) , which allows for the SCF-TIR1 and E2 ubiquitin ligases to come together to polyubiquitinate and subsequently degrade AID by the proteasome [26] . Tagging Kog1 with the AID-tag decreased Kog1 activity as judged by reduced proliferation of cells carrying the KOG1-AID allele ( S3A Fig ) . However , when KOG1-AID cells expressing pTEF1-osTIR1 were treated with IAA to deplete Kog1 , growth and proliferation were completely abolished ( S3A and S3B Fig ) . The AID tag also partially interfered with sporulation as Kog1-AID cells exhibited reduced meiosis efficiency following treatment with PKA inhibitors and rapamycin ( Fig 5B ) . Nonetheless , it was evident that depletion of Kog1 strongly affected IME1 expression and as a result meiosis did not occur ( Fig 5B and 5C ) . In conclusion , when we induce sporulation by inhibiting PKA in nutrient rich conditions , Kog1 is required for entry into sporulation . Given that inactivation of TORC1 by depleting Kog1 abolished the cells’ ability to sporulate , we hypothesized that some TORC1 activity is needed for entry into sporulation . To test this hypothesis we modulated TORC1 activity . A number of TORC1 pathway mutants have been isolated previously and have been shown to reduce basal TORC1 activity [27–29] . Typically , these mutants are hypersensitive to rapamycin and some mutants cannot recover growth after rapamycin treatment . If reduced TORC1 activity is necessary for entry into sporulation , such mutants should sporulate in the presence of a nitrogen source and/or amino acids . To test this , we generated gene deletion mutants in two nonessential subunits of TORC1 , TCO89 and the kinase TOR1 . In addition , we mutated the GTPase GTR1 , an upstream activator of TORC1 and a component of EGO complex . As reported previously , vegetative growth was strongly reduced in gtr1Δ and tor1Δ mutants when treated with rapamycin , and was abolished completely in tco89Δ cells ( S3C Fig ) [27 , 30] . Upon inhibition of PKA more than 80 percent of mutant cells ( tco89Δ , tor1Δ , and gtr1Δ ) completed meiosis compared to approximately 20 percent of control cells ( Fig 5D ) . The ability to undergo meiosis was abolished in tco89Δ and gtr1Δ cells when treated with PKA inhibitor and rapamycin ( Fig 5D ) . Meiosis was not affected in tor1Δ cells treated with rapamycin . This can be explained by the presence of the functionally similar Tor2 kinase in TORC1 , which can compensate for the tor1Δ [27 , 29] . Next , we examined how IME1 promoter activity was affected by tco89Δ , gtr1Δ , or tor1Δ mutations . In cells treated with PKA inhibitor , LacZ activity was significantly higher in all three mutants compared to the control ( 6h after treatment ) ( Fig 5E ) . Moreover , the kinetics and levels of IME1 induction in the mutant cells treated with PKA inhibitor alone closely resembled that of control cells treated with both PKA inhibitor and rapamycin . The tco89Δ and gtr1Δ cells treated with rapamycin and PKA inhibitors did not express IME1 , which is consistent with the observation that these mutants did not induce meiosis under this condition . As expected , in tor1Δ mutant cells , rapamycin only had a minor effect on IME1 promoter activity . In order to compare the tco89Δ mutant to control cells more closely , we monitored IME1 induction and meiosis in a detailed time-course . When PKA was inhibited in tco89Δ cells , IME1 promoter activity increased significantly faster than in control cells ( Fig 5F ) . In contrast , the IME1 promoter was not induced even at later time points ( 12h or 24h ) in tco89Δ cells treated with rapamycin and PKA inhibitors . Moreover , the kinetics of meiosis in tco89Δ cells treated with PKA inhibitor alone closely resembled that of control cells treated with PKA and TORC1 inhibitors and both underwent meiosis efficiently ( Fig 5G ) . These data show that a certain level TORC1 activity is required for IME1 transcription and entry into sporulation . When TORC1 activity is high or completely blocked , IME1 expression and sporulation are repressed . Several mechanisms could be responsible for the observation that complete inactivation of TORC1 prevents pIME1-LacZ activity . Given that TORC1 regulates translation and ribosome biogenesis , one plausible explanation is that the β-galactosidase protein is not produced [11] . Another possibility is that mRNA production or stability is affected by inactive TORC1 . To distinguish between these two possibilities , we measured IME1 mRNA and protein levels in tco89Δ cells , rather than IME1 promoter activity . We found that both IME1 mRNA and protein levels were strongly reduced , but not completely abolished , in tco89Δ cells treated with rapamycin ( Fig 6A and 6B , compare lanes 6–9 to 10–13 ) . Furthermore , we observed a small increase in Ime1 protein that correlated with IME1 mRNA levels following 4 hours of treatment . We conclude IME1 mRNA accumulation is predominantly affected when TORC1 is completely inhibited . The TORC1 complex has multiple effectors that regulate cellular processes such as autophagy , nitrogen and amino acid sensing , as well as ribosome biogenesis [11] . We investigated whether the Sch9 branch of TORC1 is important for IME1 regulation . Sch9 is a serine/threonine kinase that controls ribosome biogenesis , autophagy , and entry into stationary phase [31–33] . It is also directly phosphorylated by Tor1 [34] . First , we quantified IME1 promoter activity in sch9Δ mutant cells shifted from YPD to SPO medium . We found that IME1 promoter activity was only slightly higher in sch9Δ cells compared to control cells ( Fig 7A ) . We next measured IME1 promoter activity in cells grown in rich medium using the tpk1-as allele . Upon inhibition of PKA , IME1 promoter activity was overall higher and accumulated with faster kinetics in sch9Δ cells compared to control cells ( Fig 7B ) . These data indicate that Sch9 is a repressor of IME1 . If Sch9 is the only downstream target of TORC1 that represses IME1 than lowering TORC1 activity in sch9Δ cells should not further affect IME1 expression . Indeed , rapamycin treatment did not further increase IME1 promoter activity in sch9Δ cells ( Fig 7B , compare pink bars with light green bars ) . These results indicate that Sch9 mediates repression of IME1 by TORC1 . To further analyse how Sch9 controls sporulation we measured IME1 levels in single cells by smFISH . We observed that in sch9Δ cells treated with PKA inhibitors the majority ( more than 60% ) of cells expressed more than 21 copies per cell ( Fig 7C ) . Only a small fraction of cells did not induce IME1 . Furthermore , the IME1 mRNA distribution pattern in sch9Δ cells closely resembled that of tco89Δ cells , which supports our finding that TORC1 repression of IME1 is mediated by Sch9 . Finally , we measured how meiosis is affected in sch9Δ cells . In line with the IME1 expression data , the percentage of cells that completed meiosis was significantly higher ( 60% versus less than 20% in the control ) in sch9Δ cells treated with PKA inhibitor ( Fig 7D ) . In addition , in sch9Δ cells shifted from YPD to SPO the percentage of cells that underwent meiosis was substantially higher than in wild-type cells induced to sporulate under these conditions ( 90% versus 20% ) . Given that we only observed a small increase in IME1 under this conditions ( Fig 7A ) , the result suggests that Sch9 may also inhibit meiotic progression downstream of IME1 induction and regulate events such as IME2 and NTD80 induction . In conclusion , TORC1-Sch9 signalling contributes to repressing IME1 expression and entry in sporulation in nutrients-rich conditions . Having established that PKA and TORC1 control IME1 expression , we next determined how both signalling pathways control the association of transcription factors with the IME1 promoter . Several regulators of IME1 have been identified that regulate IME1 including the Tup1-Cyc8 complex [5 , 35] . This transcriptional repressor is recruited to promoters by sequence specific transcription factors and represses transcription by masking and inhibiting the transcriptional activation domains of transcription factors at gene promoters [36–38] . Several lines of evidence indicate that Tup1 contributes to IME1 repression . TUP1 and CYC8 mutants have been identified in a screen for genes that repress IME1 expression [35] . In addition , when we plotted the nucleosome occupancy at the IME1 locus using data from a published genome-wide study , we found that in tup1Δ mutant cells the IME1 promoter is almost completely depleted for nucleosomes suggesting that the promoter is de-repressed ( Fig 8A ) [39] . Finally , ChIP sequencing data indicated that depletion of Tup1 leads to increased binding of RNA polymerase II to the IME1 ORF and up-regulation of IME1 transcription [38] . To examine whether Tup1 directly regulates IME1 , we measured Tup1 binding across the IME1 promoter by ChIP in nutrient rich conditions ( Fig 8B ) . We found that Tup1 binds strongly ( more than 30 fold over background ) to the IME1 promoter in a region around 800 to 1000 base pairs upstream of the translation start side . To determine whether Tup1 binding to the IME1 promoter is regulated by nutrients , we treated tpk1-as diploid cells with 1NM-PP1 or rapamycin for 4 hours . Tup1 binding to the IME1 promoter was not affected . However , when we inhibited PKA and TORC1 , Tup1 binding to the IME1 promoter was lost ( Fig 8C ) . Finally , we tested whether the degree of TORC1 activity affected Tup1 binding to the IME1 promoter . In tco89Δ mutant cells Tup1 was bound to the IME1 promoter in rich medium . Inhibition of PKA in this mutant background was sufficient to disassociate Tup1 from the IME1 promoter ( Fig 8C ) . Interestingly , when we treated tco89Δ cells with rapamycin to inactivate TORC1 , Tup1 binding to the IME1 promoter was also not detectable . We conclude that Tup1 binding to the IME1 promoter is controlled by PKA and TORC1 activity . We next tested whether Tup1-Cyc8 association with the IME1 promoter is important for IME1 repression . We found that Tup1 depletion was sufficient for the activation of the IME1 promoter ( Fig 8D ) . In diploids cells harbouring a TUP1-AID fusion and expressing pTEF1-osTIR1 , IME1 promoter activity ( pIME1-LacZ ) increased after treatment with IAA . In contrast , β-galactosidase expression was not induced in untreated cells or cells that only expressed Tup1-AID . Finally , we compared the level of IME1 induction between Tup1 depleted cells , wild type starved cells ( SPO medium ) , and cells treated with PKA and TORC1 inhibitors grown in rich medium ( Fig 8E ) . Overall IME1 promoter activity was similar under the different conditions , but increased more rapidly in Tup1 depleted cells compared to cells starved in SPO medium or treated with PKA and TORC1 inhibitors . We conclude that Tup1-Cyc8 is a key repressor of the IME1 promoter , and that PKA and TORC1 control Tup1 association with the IME1 promoter .
Previous work implicated both the PKA and TORC1 signalling pathways in regulating IME1 . Constitutively active PKA , as observed in hyperactivated RAS2 and in BCY1 loss of function mutants , inhibits sporulation [9 , 10] . Conversely , when PKA signalling is inhibited or reduced , sporulation occurs in a subpopulation of cells even in nutrient-rich conditions . Furthermore , inhibition of TORC1 with rapamycin leads to IME1 induction and sporulation in saturated YPD cultures [7 , 21] . Our work shows that entry into sporulation can be achieved in nutrient-rich conditions by inhibiting PKA and lowering TORC1 signalling . Inhibition of these pathways leads to sporulation with similar kinetics and efficiency as starvation induced sporulation . It has been shown that multiple other signalling pathways can also contribute to IME1 regulation including G1 cyclins , several MAPK pathways and the Snf1 pathway [40–45] . Given that PKA and TORC1 signalling control the phosphorylation status of a large number of proteins [46 , 47] , we propose that some of the previously described regulators of IME1 act downstream of PKA and TORC1 signalling . Further work is needed to decipher how the different signalling networks are connected to each other and how they control entry into sporulation . Our data suggests that TORC1 and PKA do not only control IME1 , but also downstream events such as meiotic divisions and packaging into spores . For example , in cells with low PKA , inhibition of TORC1 with rapamycin further stimulates IME1 induction but also has a profound effect on progression into meiotic divisions and spore formation ( Fig 3A ) . How PKA and TORC1 control other stages of sporulation is not well understood . Our observation that triad formation is significantly enhanced and spore viability is reduced when PKA and TORC1 are inhibited implicates that the two pathways must be tightly regulated during gametogenesis . Further analyses is needed to dissect how PKA and TORC1 themselves are controlled throughout sporulation . In our efforts to understand how TORC1 and PKA repress IME1 , we identified two factors: Sch9 and Tup1 . We find that Sch9 , a major mediator of TORC1 signalling , negatively regulates IME1 . Interestingly , Sch9 and PKA are genetically redundant and functionally overlap [48] . Global gene expression analyses indicate that Sch9 and PKA regulate a common set of genes [49] . These observations suggest that PKA and Sch9 may share one or multiple downstream effectors to control IME1 and entry into sporulation . Indeed , it is known that Sch9 and PKA phosphorylation inhibits the protein kinase Rim15 , which is required for quiescence , IME1 expression and sporulation [32 , 50] . However , a constitutive active allele of RIM15 cannot de-repress IME1 in the presence of ample nutrients suggesting that Rim15 is not the only target of Sch9 and PKA [51] . PKA and TORC1 could also repress IME1 expression by controlling G1 cyclins . It was previously shown that the G1 cyclins CLN1 , 2 and 3 repress IME1 [40] . Given that TORC1 and PKA control CLN1-3 expression , it is possible that CLN1-3 partially mediate PKA and TORC1 repression of IME1 [52–54] . PKA is also known to phosphorylate the transcription factors Sok2 , Msn2/4 , Sko1 and Com2 , which directly bind and control the IME1 promoter [15 , 42 , 55] . Further efforts are needed to identify downstream effectors of PKA and TORC1 that mediate the regulation of IME1 . Our data show that the Tup1-Cyc8 complex is a direct repressor of IME1 that mediates the signals coming from PKA and TORC1 . Tup1 binds to the IME1 promoter in nutrient rich conditions , but dissociates from the promoter when both PKA and TORC1 are inhibited . The Tup1-Cyc8 complex functions as a global repressor of transcription and is recruited to promoters by sequence specific DNA binding proteins [36–38] . Identifying transcription factors that recruit Tup1 to the IME1 promoter will give important insights into how IME1 is regulated by TORC1 and PKA signalling . Interestingly , Tup1 depleted cells do not enter sporulation , even though these cells strongly induce IME1 . It is possible that other downstream factors , which control entry into sporulation , are not activated under these conditions . For example it has been known that Ime1 translation , phosphorylation , and localization are also affected by nutrients [20 , 56–58] . In addition , Tup1 is also required for sporulation . Starving Tup1 depleted cells to induce sporulation , did not result in spore formation . We hypothesize that Tup1 also regulates the transcription of genes that are important for preventing sporulation . Our analyses revealed a positive role for TORC1 in inducing sporulation . When TORC1 is completely inactive , IME1 is not induced and entry into sporulation does not occur ( Figs 5 and 9 ) . We propose that downstream effectors of TORC1 must have opposite effects on IME1 expression and entry into sporulation . Reduced levels of TORC1 activity are required to inactivate Sch9 ( discussed in previous section ) . Some TORC1 signalling however is needed to induce IME1 expression via as yet unidentified downstream mediators . Our findings also reconcile two previous contradictory observations regarding the effect of rapamycin on sporulation . Rapamycin treatment was shown to induce IME1 and sporulation [7 , 12 , 21] but when rapamycin was combined with nutrient starvation , sporulation was reduced [21] . The observation that intermediate levels of TORC1 are needed for IME1 induction also implies that there is a defined window of activity to induce sporulation . Given that sporulation is energy consuming , perhaps TORC1 senses whether there are sufficient nutrients available for cells to induce IME1 and undergo sporulation . To facilitate the energy and metabolic needs throughout sporulation , metabolism is finely and dynamically controlled [59 , 60] . Mitochondrial respiration activity is essential for both IME1 expression and sporulation in starvation medium [7 , 8] . Previous work showed that inhibition of TORC1 in rich medium induces IME1 in respiration competent , but not in respiration deficient cells [7] . However , in this study TORC1 activity was inhibited in cells grown to a relatively high density ( OD600 = 5 . 5 ) . It is likely that glucose was already consumed from the medium for some extent and not abundant enough to support IME1 expression when respiration is blocked . Our system enabled us to challenge the role of respiration during IME1 induction in rich medium plus glucose . By inhibiting both PKA and TORC1 pathways , we demonstrate that IME1 can be expressed in respiration deficient cells when a fermentable carbon source such as glucose is available . Thus ATP/energy production via either respiration or fermentation is required for IME1 activation . It is interesting to speculate that IME1 functions as an energy sensor that ensures that sporulation is induced by the lack of nutrients and only occurs when the energy source is sufficient for cells to complete sporulation . We further note that the kinetics of IME1 induction in respiration deficient cells is somewhat slower than in wild-type cells ( Fig 4C ) . Even though we cannot exclude the possibility that other functions of mitochondria are contributing to IME1 expression , we favour the idea that in wild-type cells glycolysis is simply not sufficient to produce the energy needed for rapid activation of IME1 expression due to a reduced glucose uptake from the medium . Taken together , we propose that respiration is an essential provider of ATP during starvation induced sporulation . Signal sensing and signal integration are key determinants of cell fate specification and development . In mammalian cells multiple signals often integrate at master regulatory genes to control cell specialization . The IME1 promoter serves as a model for signal integration at complex promoters because it can sense multiple nutrient signals and is regulated by transcription of long noncoding RNAs . Understanding the regulation of yeast entry into gametogenesis may shed light on how complex cell fate choices are made in mammalian cells during development .
SK1 strain background was used for the experiments throughout this manuscript and the genotypes are listed in S2 Table . Gene deletions were generated using one-step deletion protocol as described in [61] . The tpk1-as allele was realized by deleting tpk2 , tpk3 , and by making a point mutation in tpk1M164G as described in [14 , 33] . The C-terminal auxin induced degron ( AID ) tag ( used for the KOG1-AID and TUP1-AID alleles ) , which also includes three copies of the V5 epitope , was generated using one step PCR integration ( the plasmid was a gift from Leon Chan , UC Berkeley ) . A plasmid expressing Oryza sativa TIR1 ( osTIR1 ) ubiquitin E3 ligase from the TEF1 promoter was integrated at the HIS3 locus by digestion with Pme1 ( the pTEF1-osTIR1 plasmid was a gift from Leon Chan , UC Berkeley ) . Indole-3-acetic acid ( IAA ) was used to induce depletion of Kog1-AID and Tup1-AID [26] . In general , cells were grown overnight in YPD to saturation ( 1% yeast extract , 2% peptone , 2% glucose ) at 30°C , then diluted to fresh YPD ( OD600 = 1 ) and treated with different drugs or shifted to sporulation medium ( SPO , 0 . 3% potassium acetate , 0 . 002% raffinose , pH 7 . 0 ) . In some experiments cells were grown overnight in YPD , diluted to pre-sporulation medium ( BYTA , 1% yeast extract , 2% tryptone , 1% potassium acetate , 50 mM potassium phtlhalate ) for 16 hours , and subsequently shifted to YPD or SPO [62] . Rapamycin was added to cells in a final concentration of 1000 ng/μl unless written otherwise . 1NM-PP1 was added to cells in a final concentration of 3 μM . The northern blot protocol for IME1 was described previously [5] . The RT-PCR protocol was described previously [5] . In short , total RNA was treated with DNAse and purified . 750 ng of total RNA was used for the reverse transcription reaction , and single stranded cDNA were quantified by real-time PCR using a SYBR green mix ( Life Technologies ) on a 7500 Fast Real-Time PCR system ( Life Technologies ) . Signals were normalized to ACT1 transcripts levels . The primer sequences used are included in S3 Table . A tricarboxylic acid ( TCA ) extraction protocol was used to make total protein extracts . Samples were separated by SDS page , blotted onto PVDF blotting membrane , and subsequently incubated with anti V5/1:2000 dilution ( Life Technologies ) and anti-hexokinase/1:8000 dilution ( Stratech Scientific ) antibodies . As secondary antibodies IRDye800CV/1:15000 dilution ( anti-mouse , LI-COR Biosciences ) and IRDye680RD/1:15000 dilution ( anti-rabbit , LI-COR Biosciences ) were used . Western blot images generated using the Odyssey system ( LI-COR Biosciences ) . Cells were fixed overnight in 80% ethanol , and stained with 0 . 05 μg/ml 4′ , 6-diamidino-2-phenylindole ( DAPI ) solution in 100 mM phosphate buffer ( pH 7 ) . The number of DAPI masses in at least one hundred cells ( n = 100 ) was counted . Liquid ortho-Nitrophenyl-β-galactoside ( ONPG ) assay was performed as described previously [5] . In short , 2 ml of OD600 = 1 cell pellets were washed with buffer Z ( Phopsphate buffer pH 7 , KCl 10 mM , MgCl 1mM ) and were snap frozen in liquid nitrogen . Samples for each biological replicate were collected on different days , but ONGP assays were performed together at the same time . Cells were chemically disrupted using Y-PER buffer ( Thermo Scientific ) . Subsequently cells were incubated with ONPG ( Sigma ) ( 1 mg/ml in Z buffer plus 50 mM β-mercaptoethanol ) till yellow colouring occurred . The reaction was quenches using sodium carbonate ( 1 mM ) and cell debris was cleared by centrifugation . Absorption of each sample was measured at OD420 using a 96 well plate reader . Miller Units were calculated according to a standard formula [63]: Miller Unit = ( signal from plate reader ( OD420 ) x 1000 ) / ( cell density ( OD600 ) x time of incubation with ONPG ( min ) ) . The data from the experiments represents the standard error of the mean of at least two biological experiments . Chromatin immunoprecipitation ( ChIP ) experiments were performed as described previously [5] . Cells were fixed with 1% formaldehyde for 20 min , the reaction was quenched with 125 mM glycine . Cells were disrupted using mini beadbeater ( BioSpec ) , and crosslinked chromatin was sheered by sonication using Bioruptor ( Diagenode , 6 cycles of 30 sec on/off ) . Chromatin extracts were then incubated with anti V5 agarose beads ( Sigma ) for 2 hours at room temperature , and beads were washed accordingly . To measure Tup1 binding , input and ChIP samples were quantified by real-time PCR using SYBR green mix ( Life Technologies ) and primers corresponding to the IME1 promoter on a 7500 Fast Real-Time PCR system ( Life Technologies ) . The mating type locus ( HMR ) was used as a non-binding negative control . The primer sequences used are included in S3 Table . The single molecule RNA fish was performed as described previously [5] . In short , cells were fixed with formaldehyde overnight , treated with zymolyase and further fixed in 80% ethanol . Subsequently cells were hybridized with fluorophore labelled probes directed to IME1 ( AF594 ) and the internal control ACT1 ( Cy5 ) . Cells were imaged using a 100x oil objective , NA 1 . 4 , on a Nikon TI-E imaging system ( Nikon ) . DIC , DAPI , AF594 ( IME1 ) , Cy5 ( ACT1 ) images were collected every 0 . 3 micron ( 20 stacks ) using an ORCA-FLASH 4 . 0 camera ( Hamamatsu ) and NIS-element software ( Nikon ) . ImageJ software was used to make maximum intensity Z projections of the images [64] . Subsequently , StarSearch software ( http://rajlab . seas . upenn . edu/StarSearch/launch . html , Raj laboratory , University of Pennsylvania ) was used to determine number transcripts in single cells . Comparable thresholds were used to count the RNA foci in single cells . Only cells positive for the internal control ACT1 were used for the analysis . At least a total n = 60 cells were counted for each experiment . | The cell-fate controlling gametogenesis is essential for all sexual reproducing organisms . In Saccharomyces cerevisiae , entry into gametogenesis or sporulation is dictated by mating type and nutrient availability . These signals regulate the expression of the master regulator of entry into sporulation , IME1 . In this manuscript we describe how nutrients control IME1 . We show that inhibiting two highly conserved nutrient sensing pathways ( PKA and TORC1 ) mimics starvation-induced sporulation and drives cells to induce IME1 and complete meiosis in nutrient-rich conditions . In addition , we show that respiration and fermentation are interchangeable energy providers for entry into gametogenesis . Finally , we have uncovered a critical role for TORC1 during entry into gametogenesis . In addition to the known role of TORC1 in repressing IME1 , we find that intermediate TORC1 activity is required for entry in gametogenesis . Too much or too little TORC1 activity inhibits gametogenesis . Our data explains how two conserved signalling pathways control a developmental decision essential for sexual reproduction , about which remarkably little is known in all eukaryotes . Thus the activities of two nutrient sensing pathways and energy availability determine whether cells enter gametogenesis or not . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"meiosis",
"medicine",
"and",
"health",
"sciences",
"chemical",
"compounds",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"messenger",
"rna",
"carbohydrates",
"organic",
"compounds",
"glucose",
"reproductive",
"physiology",
"glucose",
"signaling",
... | 2016 | Nutrient Control of Yeast Gametogenesis Is Mediated by TORC1, PKA and Energy Availability |
The Adenovirus ( Ad ) genome within the capsid is tightly associated with a virus-encoded , histone-like core protein—protein VII . Two other Ad core proteins , V and X/μ , also are located within the virion and are loosely associated with viral DNA . Core protein VII remains associated with the Ad genome during the early phase of infection . It is not known if naked Ad DNA is packaged into the capsid , as with dsDNA bacteriophage and herpesviruses , followed by the encapsidation of viral core proteins , or if a unique packaging mechanism exists with Ad where a DNA-protein complex is simultaneously packaged into the virion . The latter model would require an entirely new molecular mechanism for packaging compared to known viral packaging motors . We characterized a virus with a conditional knockout of core protein VII . Remarkably , virus particles were assembled efficiently in the absence of protein VII . No changes in protein composition were evident with VII−virus particles , including the abundance of core protein V , but changes in the proteolytic processing of some capsid proteins were evident . Virus particles that lack protein VII enter the cell , but incoming virions did not escape efficiently from endosomes . This greatly diminished all subsequent aspects of the infectious cycle . These results reveal that the Ad major core protein VII is not required to condense viral DNA within the capsid , but rather plays an unexpected role during virus maturation and the early stages of infection . These results establish a new paradigm pertaining to the Ad assembly mechanism and reveal a new and important role of protein VII in early stages of infection .
Adenovirus ( Ad ) infection is generally associated with mild respiratory , ocular , or gastrointestinal diseases , but Ad has been recognized in recent years as a significant pathogen in immunocompromised patients and in the young and elderly [1] . Ad replication involves a number of events that must be temporally and spatially organized within the host cell in order to lead to optimal productive infection . The 36-kbp Ad genome encodes at least 25 early gene products and ~15 late gene products . The early viral proteins alter host cell functions to promote an environment that is conducive to viral replication and to block cellular and host antiviral responses , as well as for the enzymatic replication of the Ad genome [2] . The late gene products comprise structural components of the capsid , as well as proteins involved in virion assembly and maturation and viral genome encapsidation [3] . Ad is an excellent example of a virus that efficiently utilizes limited genetic capacity to maximize viral protein and virion production . Viral nucleic acids are sensed following infection by pathogen recognition receptors , and other cytoplasmic and nuclear effectors , to trigger cellular , antiviral responses [4] . With Ad , viral infection triggers the cGAS/STING pathway and activation of type I interferon ( IFN ) signaling [5] . Recent studies have shown that cGAS/STING activation stimulates the TBK1/IRF3 cascade to promote type I IFN production and the activation of IFN-stimulated genes ( ISG ) [6] . One such ISG that is known to promote both intrinsic and innate responses to viral infection is the product of the promyelocytic leukemia ( PML ) gene [7] . The PML protein nucleates the formation of PML nuclear bodies ( PML-NB ) that exert antiviral effects on a wide range of viruses . Many viruses , including Ad , express proteins that interfere with PML-NB activity . Ad utilizes different mechanisms to counteract IFN signaling pathways . These include E1A inhibition of different aspects of IFN signaling [4 , 8–10] , E1B-55K inhibition of ISG expression and function [11–13] , inhibition of STAT1 phosphorylation and nuclear translocation by the E3-14 . 7K protein [14] , sequestration of nuclear STAT1 into viral replication centers [15] , and Ad VA RNA-I inactivation of PKR [16] . Viruses with linear , dsDNA genomes , such as Ad , also trigger a cellular DNA damage response ( DDR ) [17] . With Ad , the DDR severely inhibits viral DNA replication if unabated [18] . Ad has evolved two redundant mechanisms to inhibit a DDR , involving the E4-ORF3 protein and the E1B-55K/E4-ORF6 protein complex , to allow efficient Ad DNA replication to occur . With both mechanisms , Ad proteins target and inhibit the sensors of DNA damage , the Mre11-Rad50-Nbs1 ( MRN ) complex [18] . Our results support the hypothesis that the major Ad core protein , protein VII , protects the viral genome from recognition by a DDR during the early stages of infection until the E1B and E4 gene products are synthesized to counteract this response [19] . Ad core protein VII also impacts innate cellular signaling mechanisms . A recent study demonstrated that Ad core protein VII binds to and regulates cellular chromatin and sequesters immune danger signals to control immune signaling [20] . The Ad genome within the virion is associated with ~500–800 copies of core protein VII [21 , 22] . Core protein VII condenses DNA in vitro and in vivo and assembles viral DNA into a nucleosome-like structure [23–28] . It has been widely assumed that protein VII is required to condense viral DNA within the Ad capsid [29] . Core protein VII remains associated with the Ad genome during the early phase of virus infection [30–32] and is released from viral DNA coincident with early gene transcription [31] . Human Ads also encode two other basic core proteins , proteins V and X/μ , which are packaged into the virion [33] . Several lines of evidence suggest that the Ad genome is positioned within the capsid in globular domains [34 , 35] , but the structure of the Ad DNA-protein core is unknown [29] . Protein V of the incoming virions remains in the cytoplasm following escape of the Ad core from the endosome [36]; the fate of core protein X/μ during virus disassembly has not been determined . The packaging of the Ad genome into the capsid is thought to follow the paradigm of dsDNA bacteriophage assembly where viral DNA is inserted into a preformed , empty capsid using a packaging motor [37] . Highly sophisticated studies have been conducted on the assembly of dsDNA bacteriophage [38] . A precursor viral capsid , the prohead , is formed that contains a unique portal vertex through which naked viral DNA is packaged . A portal protein complex is found at this vertex which associates with a powerful packaging motor , the terminase . The terminase contains a DNA translocation ATPase and a concatamer-resolving endonuclease . The DNA packaging motor uses the hydrolysis of ATP to translocate DNA into the capsid [39] . Within the capsid , the viral DNA associates with protamines and cations to neutralize the negative charge and allow genome compaction . A key feature of this packaging process is that there is a direct electrostatic interaction between the phosphate backbone of naked viral DNA with the ATPase subunits of the packaging motor to direct DNA packaging in a stepwise process [40] . An Ad assembly intermediate has been identified that corresponds to a prohead and several lines of evidence suggest that this virus particle is a precursor for viral DNA packaging [41 , 42] . The Ad IVa2 protein is present at a unique vertex [43] and contains conserved Walker A and B box motifs that are hallmarks of ATPases . The IVa2 protein binds ATP in vitro and the Walker A and B box motifs are required for ATP binding and for viral DNA encapsidation [44] . It is not known if naked Ad DNA is packaged , as with bacteriophage , followed by the encapsidation of viral core proteins , or if a unique packaging mechanism exists with Ad where a DNA-protein complex is simultaneously packaged into the capsid [37] . The latter model would require an entirely new molecular mechanism for viral DNA packaging compared to known bacteriophage packaging motors . If naked Ad DNA is packaged and core proteins inserted into the capsid following genome encapsidation , then another packaging process must be proposed to account for the insertion of core proteins into the capsid . An alternative view of Ad virion assembly is a concerted model where capsids assemble around a viral DNA-protein core [45] . Here , we studied the role of the Ad major core protein VII in the viral life cycle . Our results demonstrate that protein VII is not required for virion assembly or Ad genome packaging , but rather protein VII plays a critical role in virus escape from the endosome following infection . Thus , Ad core protein VII plays an unexpected role during early stages of virion entry . These results establish new paradigms pertaining to the Ad assembly mechanism and the role of core protein VII during infection .
The Ad5 L2 region ( Fig 1A ) encodes four viral proteins: penton base ( capsid protein III ) and the three core proteins pre-VII , V , and pre-X . The precursors of protein VII ( pre-VII ) and protein X ( pre-X ) are proteolytically processed by the Ad proteinase AVP to mature forms designated VII and μ , respectively [3] . We wanted to analyze the role of the major Ad core protein VII in virus assembly and other aspects of the viral life cycle . A traditional approach to analyze viral mutants is to establish a complementing cell line to allow the propagation of a defective virus . However , it has been difficult to produce cell lines that efficiently complement the growth of Ad late protein mutants , for example owing to the high levels of protein expression that are required for complementation and the potential toxicity of their expression to cells . We established an approach to conditionally knock out the gene encoding pre-VII in the context of virus infection using the Cre-lox system . An Ad5 infectious clone was generated with loxP sites flanking the pre-VII coding sequences ( Fig 1A; loxP sites at -4 relative to the A of the ATG and immediately following the pre-VII termination codon; the virus is termed Ad5-VII-loxP ) . Recombinant Ad5-VII-loxP virus stocks were readily established in cells that do not express Cre recombinase , and the virus replicated with wild-type kinetics as measured by qPCR . Two independent Ad5-VII-loxP virus stocks were analyzed ( isolates 5 and 11 ) . Ad5-pVII-loxP was used to infect cells that express Cre recombinase , and the efficiency of excision of pre-VII coding sequences ( floxing ) was measured by Southern blot ( Fig 1B ) and qPCR . Ad5-VII-loxP floxing efficiency was ≥99% in 293 cells that express Cre recombinase ( lanes 2 , 3 vs . 4 , 5 ) . We examined pre-VII/VII protein levels following infection of cells that do or do not express Cre recombinase with Ad5-VII-loxP by Western blot ( Fig 1C ) . Infection of Cre recombinase-expressing cells with Ad5-pVII-loxP reduced pre-VII/VII protein expression to barely detectable levels in comparison to infection of parental 293 cells . Equivalent levels of L2 proteins III and V were detected in cells that do or not express Cre recombinase demonstrating a specific effect on pre-VII/VII protein expression and not a global deficit in L2 gene expression . These results establish the utility of using Cre recombinase to direct conditional pVII/VII protein expression . We examined the production of virus particles following infection of cells that do or do not express Cre recombinase with Ad5-VII-loxP . Ad5-VII-loxP virus particles were efficiently produced in cells that express Cre recombinase , and the virions produced in Cre-expressing or parental 293 cells banded at the same density in a CsCl equilibrium gradient ( 1 . 34 g/cc ) ( Fig 2A , isolate 5 ) . Virus particles that banded at this density contained the full-length viral genome and Ad5-VII-loxP particles produced in Cre-expressing cells were floxed ≥99% ( S1A and S1B Fig ) . These virus particles were examined by electron microscopy in comparison to wild-type Ad5 ( Fig 2B ) . Virus particles that contained or lacked protein VII ( VII-lox , Cre+ , Cre– , respectively ) were indistinguishable and had the same morphology and electron density as wild-type Ad5 . We examined virion protein composition by SDS-PAGE and Coomassie blue and silver staining and by Western blot . The identical pattern of major and minor viral capsid proteins were observed with the only detectable difference being the lack of protein VII in Ad5-VII-loxP particles produced in Cre-expressing cells ( Figs 2C and S1C ) . By western blot , all Ad late major and minor capsid proteins were equally represented in Ad5-VII-loxP particles produced in Cre-expressing cells compared to parental 293 cells , and in comparison to wild-type Ad5 , with the sole exception of protein VII ( Fig 2D ) . We conclude that Ad5 protein VII is not required for packaging of the viral genome into the virion and that the loss of protein VII in the capsid is not compensated by an increase in the levels of core protein V or other capsid proteins . We did note an effect on the proteolytic processing of Ad late protein pre-VI when protein VII was missing ( Fig 2D , pre-VI , VI; doublet with VII−compared to a single form with WT ) . Pre-VI is proteolytically cleaved by AVP at both the N- and C-termini , whereas pre-VIII is processed at three internal cleavage sites [46] . The processing of proteins pre-VI and pre-VIII was examined further by western blot using specific antisera ( Fig 2E ) . Ad2 temperature-sensitive mutant ts1 is defective for AVP at the restrictive temperature , accumulates precursor forms of Ad late proteins , and was used as a control [23] . Ts1 virions contained the unprocessed form of pre-VI ( indicated by an asterisk in the top panel; pVI/VI ) and pre-VIII ( VIII-C ) . Mature VI , was observed with wild-type Ad5 virions ( indicated by an open circle in the top panel ) . Virions that lacked protein VII contained a partially processed form of pre-VI ( termed intermediate VI , iVI , [47]; indicated by a bullet in the top panel ) that exhibited a faster mobility than ts1 pre-VI and that was processed at the C-terminus ( VI-C ) , but presumably not cleaved by AVP at the N-terminus . Pre-VIII C-terminal processing in virions that lacked protein VII occurred normally ( VIII-C ) . The parent Ad5-VII-loxP virions contained both iVI and mature VI . The basis for this observation is not clear; the genomic pre-VI region was fully sequenced with both infectious clone isolates of Ad5-VII-loxP and the sequences were wild-type . We examined the infectivity of virus particles that lack protein VII using a fluorescent focus assay in HeLa cells and found a >300-fold decrease in infectivity compared to wild-type Ad5 and a >100-fold decrease compared to Ad5-VII-loxP virus grown in 293 cells ( Particle:FFU ratios were 5:1 for Ad5-WT , 13 . 5:1 for Ad5-VII-loxP grown in 293 cells , and 1820:1 for Ad5-VII-loxP grown in Cre-expressing cells ) . We analyzed the formation of viral replication centers in A549 and HeLa cells at low multiplicity of infection with wild-type Ad5 and Ad5-VII-loxP virus grown in 293 cells and at high multiplicity of infection with Ad5-VII-loxP virus grown in 293-Cre cells by immunofluorescence ( IF ) ( Fig 3A , A549 cells , Fig 3B , HeLa cells ) . The presence of circular , DBP-positive viral replication centers in the nucleus directly correlates with active viral DNA replication . Viral replication centers were readily evident in cells infected at low multiplicity of infection with wild-type Ad5 or the VII+ virus . There was a striking reduction in the number of replication foci observed 24 hours after infection with virus that lacks protein VII where very few DBP-positive cells were evident even at high multiplicity of infection . Further , of the few viral replication centers that were observed following infection with VII−virus particles , all were found to also express protein VII ( Fig 3A and 3B ) indicating that these cells were infected with Ad5-VII-loxP virions that had escaped VII floxing during production . We also used IF to visualize protein VII expression in Cre-expressing cells infected with Ad5-VII-loxP . Very few Cre-expressing cells exhibited protein VII expression at late times after infection , whereas the Ad DNA binding protein was readily evident in all infected cells . Thus , we believe that a small percentage ( ≤1% ) of the input Ad5-VII-loxP viral genomes escaped floxing in Cre recombinase-expressing cells , whereas the majority of the viral genomes ( ≥99% ) were fully floxed . Based on this conclusion , the majority of Ad5-VII-loxP virus particles produced in Cre-expressing cells should be fully deficient for protein VII , while a small percentage ( ≤1% ) contain the full complement of protein VII and are equivalent to wild-type Ad5 . The lack of core protein VII in the Ad5 particle could affect virus infectivity at a number of different steps during the viral life cycle . To determine the basis for the defect observed with the VII−mutant virus , we examined input viral DNA levels in infected cells at 2 hours post-infection ( 2 hpi ) and viral DNA replication at 24 hpi by quantitative PCR using total genome and VII gene-specific primer pairs ( Fig 4A ) . At 2 hpi , the level of VII−mutant viral DNA in total cell lysates was comparable to wild-type Ad5 and the unfloxed Ad5-VII-loxP parent virus; efficient floxing of the VII−mutant stock was verified ( 2 hpi , VII– , total vDNA versus VII+ vDNA ) . At 24 hpi , there was a 4-log decrease in the levels of VII−mutant DNA compared to wild-type Ad5 and a 3-log decrease compared to the unfloxed Ad5-VII-loxP parent virus . Further , the few genomes that replicated following infection with the VII−virions contained an intact VII reading frame and , thus , represented the low level of unfloxed viral genomes present in the VII−stock ( 24 hpi , VII– , total vDNA versus VII+ vDNA ) . These results are consistent with the IF data shown in Fig 3 . We tested if the decrease in viral DNA replication observed with the VII−virus resulted from reduced E1A gene expression , and found that the E1A mRNA levels were strongly reduced in cells infected with VII−virions ( Fig 4B ) . In order to determine if this defect in early gene expression with the VII−mutant virus could be rescued if the E1A proteins were provided in trans , 293 cells , which constitutively express Ad5 E1A proteins , were infected with Ad5-VII-loxP produced in 293 cells ( VII+ ) or Cre-expressing cells ( VII– ) ( Fig 4C ) . RNA levels for Ad early regions E2a , E2b , and E4 were reduced >100-fold in 293 cells infected with the protein VII−mutant virus in comparison to the VII+ virus , indicating no rescue of viral early gene expression . Finally , we performed a coinfection experiment with wild-type Ad5 together with the VII+ or VII−viruses and examined viral DNA levels by qPCR at early and late times after infection ( S2 Fig ) . Input levels of the coinfecting viral DNAs were similar early after infection ( WT+VII+ vs WT+VII– , 3 hpi ) . Wild-type Ad5 and the VII+ virus replicated to similar levels by late times after infection , whereas the VII−mutant virus was significantly reduced ( 24 hpi ) . We conclude that the VII−mutant virus is defective in an early step in the infection cycle . Ad5 particles that lack core protein VII show a global defect in early gene expression and viral DNA replication that cannot be complemented in trans . Viral DNAs isolated from VII+ and VII−virus particles were used to transfect cells , and viral DNA and E1A mRNA levels were analyzed by qPCR and RT-qPCR , respectively . An EGFP expression vector was used as a transfection control , and equivalent levels of GFP protein expression was observed in three replicate experiments ( Fig 5A ) . No significant differences were observed between viral DNAs isolated from VII+ ( Ad5-WT , VII-lox-293 ) or VII− ( VII-lox-Cre1 , VII-lox-Cre2 ) virus particles when viral DNA replication ( Fig 5B ) or E1A mRNA levels ( Fig 5C ) were examined . These results demonstrate that virion DNA in VII−particles is functional when used for transfection , but not in the context of viral infection . We analyzed the subcellular localization of Ad5 particles that contain or lack protein VII by IF and transmission electron microscopy ( TEM ) . Wild-type Ad5 and Ad-VII-loxP particles produced in Cre-expressing cells ( VII– ) were prepared containing viral genomes labeled with the ethynyl-modified nucleoside EdC . Copper-catalyzed azide-alkyne cycloaddition ( Click ) reactions allow visualization of single viral genomes within infected cells using Alexa Fluor 488-azide [48] . Cells were infected with EdC-labeled virions , and the major capsid protein hexon was visualized by IF using a specific antibody and viral genomes were visualized following Click reactions ( Fig 6 ) . Infections were performed either for 1 hour at 4°C to visualize cell surface-bound virus ( Fig 6A , top row ) , or for 30 minutes at 37°C , after which unbound virus was washed away and the incubation was continued at 37°C for 0 minutes , 3 hours , or 5 hours to visualize internalized virus ( Fig 6A and 6B ) . Virion attachment at the cell surface was readily detected with wild-type Ad5 using anti-hexon antibody but little viral DNA was evident due to protection by the intact capsid [49] . Immediately after 37°C infection with wild-type Ad5 ( 30min + 0min , HAdV-C5_wt ) , the viral DNA and capsids/capsid remnants were found in both nuclear and cytoplasmic areas . By 3 . 5–5 . 5 hpi ( 30min + 180min , 30min + 300min ) , wild-type Ad5 capsids and viral DNA were concentrated over the nuclear area marked by DAPI staining . In contrast , the VII−mutant virus capsids were largely excluded from within the nuclear area and concentrated in the perinuclear region with time ( Fig 6B , HAdV-C5_ΔVII ) . Due to technical limitations , VII−virions were not well detected with EdC within cells , although these virions contained EdC-labeled genomes , as shown by single virus analyses of heat-disrupted particles bound on polylysine-coated coverslips ( S3 Fig ) . VII−viral genomes that were evident in infected cells were located outside the nuclei . The localization of VII−virions in the perinuclear region is reminiscent of the phenotype observed with ts1 virions defective in AVP activity where virions do not efficiently escape the endosome [50–52] . It is possible that incoming VII- virion DNA was not well detected since the VII- virions did not open up their capsids , akin to ts1 virions which do not lyse the endosmal membrane [49 , 53] . The localization of wild-type Ad5 ( VII+ ) and VII−virions was visualized by TEM at different times after infection ( Fig 7 ) . Cells were infected for 1 hour at 37°C , washed , and further incubated for 5 minutes or 4 hours at 37°C . Incoming virions were visualized by TEM in multiple sections for each condition and scored to be on the plasma membrane , in endosomes , in the cytosol , and on the nuclear membrane . Representative images are shown in Fig 7A , and the data are quantified in Fig 7B . The number of cells ( c ) and virus particles ( v ) analyzed for each condition are indicated in the graphs shown in Fig 7B , left , and these results are normalized to the total number of particles analyzed under each condition in Fig 7B , right ( total set at 1 ) . The most striking difference between infection with wild-type Ad5 ( VII+ ) and VII−virions was the disappearance of virus particles with wild-type Ad5 over time by a factor of ~5 in contrast to a small decrease in the number of VII−virus particles . The second most striking difference between the VII+ and VII−virions was the significant clustering of VII−virions in cytoplasmic vesicles . This is most drastically illustrated at the 5 . 5 hour time point , when 10 wild-type virions were found in endosomes ( vesicle ) , contrasting to 115 VII−virions in vesicles , including multivesicular eno-lysosomal vesicles . At this time , 5 VII−virions were in the cytosol and none were on the plasma membrane or the nuclear membrane . At this time point , 4 wild-type virions were still on the nuclear membrane and 3 were in the cytosol , but most particles were disassembled and could not be recognized as virions . These results illustrate that VII−virions do not efficiently escape from endosomes , and that this likely is the reason for the poor localization on the nuclear membrane and the greatly reduced nuclear activity ( early gene transcription , viral DNA replication ) of VII−virions .
The Ad major core protein VII is a histone-like protein that condenses DNA in vitro and in vivo and assembles viral DNA into a nucleosome-like structure [23–28] , although the exact conformation of Ad chromatin within the virion remains elusive [29] . It has been widely assumed that protein VII is required to condense viral DNA within the capsid [29] . Here , we show that the opposite is true . Ad core protein VII is not required for virus assembly or packaging of viral DNA into virions . Virus particles that lack protein VII are stable and contain the normal composition of virion proteins . No change in the amount of core protein V within VII−virus particles was observed ( Fig 2 ) . These results have important implications for the Ad assembly mechanism and for the role of protein VII during infection . Ad genome packaging is thought to follow the paradigm of dsDNA bacteriophage where viral DNA is inserted into a preformed empty capsid , the prohead [37] . It was not previously known if Ad packages naked viral DNA like dsDNA phage and herpesviruses or a DNA-protein complex containing viral DNA associated with core proteins . It seems very unlikely that Ad evolved a packaging machinery that could accommodate both mechanisms . The process of packaging naked DNA as observed with dsDNA phage involves direct interaction of the packaging motor with the DNA backbone [39] , and the interaction of Ad DNA with basic core proteins would seem likely to preclude such interaction . Since viral DNA was efficiently encapsidated in the absence of protein VII , we believe that this strongly supports the model that Ad packages naked DNA , followed by , or in concert with but separate from , the packaging of core proteins . Ads in the mammalian Atadenovirus family do not encode core protein V [54] virtually eliminating a role for this protein in the packaging mechanism . Very little is known about Ad protein X/μ [55] , and we have not addressed the potential role of this core protein in the Ad assembly process . The virion population from the VII-floxed Ad5 grown in the Cre-expressing cells contained a small proportion of unfloxed particles which gave rise to protein VII expression and infection ( Fig 3 ) . The particles from the VII-floxed Ad5 entered the cell but were largely defective at endosomal escape ( Figs 6 and 7 ) , which resulted in a dramatic decrease in viral early gene expression and viral DNA replication ( Fig 4 ) . This defect could not be complemented in trans by coinfection with wild-type Ad5 ( S2 Fig ) . Viral DNA isolated from VII−virus particles was fully functional when delivered by transfection ( Fig 5 ) , but not infection . The process of Ad entry into the cell and escape from the endosome has been elegantly and extensively studied [56 , 57] . Following engagement of Ad5 fiber protein with the primary coxsackievirus adenovirus receptor CAR and secondary binding of penton base to cellular αvβ3/5 integrins , the Ad capsid is internalized by dynamin- and actin-dependent endocytosis into clathrin-coated endosomes . Acidification of the endosome is not required to promote early steps in virion disassembly . Escape of the partly uncoated Ad particle from the endosome is critically-dependent on Ad protein VI [58] . An amphipathic helix near the N-terminus of protein VI binds to the inner surface of the endosomal membrane , induces positive curvature , and appears to fragment the membrane releasing Ad into the cytoplasm [59] . Protein VI is exposed from incoming virions by mechanical cues from differential movements of the virion receptors CAR and integrins [60] , and leads to the activation of lysosomal secretion and an increase of ceramide lipids which is key for the membrane disrupting activity of protein VI [61] . Membrane rupture by protein VI occurs through an amphipathic helix near the N-terminus of protein VI , and leads to the rupture of the endosomal membrane [59] , and the clearance of the broken vesicles by an autophagic process [62 , 63] . The N-terminal 33 amino acids of pre-VI stably bind to the peri-pentonal hexon proteins on the inner surface of the virion [64 , 65] . During Ad assembly , this would orient the configuration of pre-VI within the immature capsid . During virion maturation , AVP cleaves pre-VI at the C-terminus to release an AVP-activating peptide and after amino acid 33 to release the remainder of the protein [46] . It is possible , although not very likely , that the failure of capsids that lack protein VII to escape from endosomes may reflect the lack of N-terminal processing of pre-VI observed with the VII−mutant virus ( Fig 2 ) . We note that non-processed pre-VI displays full membrane lytic activity using in vitro assays [47 , 59 , 66] . An Ad5 site-specific pre-VI mutant that has reduced N-terminal processing by AVP only displays an ~4-fold decrease in infectivity compared to wild-type Ad5 [47] . Thus , the hypothesis that the pVII−mutant virus displays an endosome escape defect due to lack of pre-VI N-terminal processing remains unlikely . Perhaps the defect of the protein VII−virions relates to physical changes in the Ad capsid that occur during virion maturation , and such changes are altered in the VII mutant virus infections . Biophysical analyses of the Ad5 core using wild-type Ad5 and ts1 grown at the restrictive temperature demonstrated that the Ad core decondenses during proteolytic maturation of the virion resulting in increased internal pressure [67] . This process has been proposed to facilitate virion disassembly during the early stages of infection [67 , 68] . ts1 capsids produced at the restrictive temperature are highly stable compared to wild-type Ad5 and fail to release capsid vertex proteins even in harsh chemical conditions [69] . Remarkably , the shell of ts1 particles is softer than the shell of wild-type virions [69 , 70] . Structural studies have shown that this process relates to altered interactions between core components and the internal faces of viral vertex proteins [71] . We propose that these changes in the physical structure of the Ad capsid may not occur properly in the absence of protein VII . Protein VII represents ~10% of the total mass of the Ad particle , and it is possible that the absence of this protein within the capsid would reduce the internal capsid pressure . It will be interesting to test this hypothesis using biophysical approaches . Finally , we cannot exclude the possibility that patches of positively charged residues on viral proteins in the inner surface of the virion unnaturally interact with viral DNA in the absence of protein VII and this interferes with virion disassembly and endosome escape . VII−capsids displayed a pronounced defect in N-terminal pVI processing . AVP proteolytically cleaves Ad proteins pre-IIIa , pre-VI , pre-VII , pre-VIII , pre-X , L1-52/55K , and pre-TP [46 , 72] . Pre-IIIa cleavage results in a very minor change in mobility in SDS-PAGE and would not have been detected in our assays . Pre-VI and pre-VIII C-terminal cleavages , and L1-52/55K processing , occurred normally with the VII−mutant ( Fig 2 ) . The effect of protein VII loss on pre-VI N-terminal cleavage appears to be specific , although reagents are not available to analyze pre-X processing and pre-TP cleavage was not evaluated . It is not clear how pre-VII/VII may affect the activity of AVP , but both interact with viral DNA , and our results suggest that pre-VII/VII influence cleavage events by AVP . AVP slides along DNA within the virion to locate and process substrates [73 , 74] . These results suggest that protein VII may influence this process and alter AVP activity and/or specificity . Future studies will be required to clarify the effect of pVII on Ad viron maturation . In conclusion , Ad core protein VII is not required for virion assembly or viral genome encapsidation . Thus , pre-VII/VII is not required for condensation of Ad DNA within the capsid . The latter result is unexpected . Relative to the length of the Ad genome , the Ad capsid is large compared to some other dsDNA viruses . For example , the Ad capsid diameter is 90–100 nm to accommodate a genome of ~36 kbp . The herpes simplex virus capsid is 100–120 nm to accommodate a genome of ~150 kbp . Similarly , the dsDNA phage T4 capsid is ~85 x 120 nm to accommodate a genome of ~170 kbp . With herpesviruses and dsDNA phages , naked viral DNA is highly condensed within the capsid [39] . The relatively large Ad capsid , as well as the results presented here , suggests that Ad DNA is not as highly condensed within the virion as in case of herpesviruses or phages , perhaps because of a relatively large virion size:genome size ratio . Our results also support the conclusion that Ad likely packages naked DNA , like larger dsDNA viruses . The packaging size limit for Ad5 is ~105% of the full-length genome [75] showing a restriction to capsid capacity . If viral DNA is packaged separately from core proteins , it is an interesting conundrum how a limit for genome size is imposed if there is still space within the virion to subsequently package core proteins . Ad core proteins are not found in empty virions [76] , although it is not clear if these capsids are a true assembly intermediate [37] . Light , intermediate capsids observed with Ad5 ts369 at the restrictive temperature , which likely do represent a bona fide assembly intermediate , contain part of the Ad genome but lack core proteins [42] , once again indicating the viral DNA encapsidation precedes core protein insertion . An alternative view of Ad assembly is a concerted mechanism by which the capsid is assembled around the viral DNA-protein core [45] . Our results are consistent with this mechanism of Ad assembly and show that protein VII is not required if this assembly process occurs . This study demonstrates that Ad core protein VII plays an entirely unexpected role during Ad infection , and is required for escape of the virion from the endosome and for full processing of capsid proteins by AVP . It will prove very interesting to determine how protein pre-VII/VII affects capsid pressurization during virion maturation , and if changes in this process underlie the observed phenotype with the VII−virions .
HEK-293 ( ATCC ) , HeLa ( ATCC ) , and A549 cells ( ATCC ) were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% bovine calf serum ( HyClone ) , penicillin , and streptomycin . The Cre66 cell line , a Cre recombinase expressing cell line derived from HEK-293 cells ( a gift from Dr . Stefan Kochaneck , University Ulm , Germany ) , was maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% Fetalclone III serum ( HyClone ) , penicillin , streptomycin , and 0 . 25 mg/ml of Geneticin ( Life Technologies ) . Cre-expressing A549 cells were produced by lentivirus transduction using Cre-IRES-PuroR and maintained in the above mentioned medium supplemented with 8 μg/ml puromycin . Cre-IRES-PuroR was a gift from Darrell Kotton ( Addgene plasmid #30205 [77] ) . Wild-type adenovirus 5 ( Ad5-WT ) was derived from the plasmid clone pTG3602 [78] by restriction digestion with PacI and transfection of DNA into cells . Ad2ts1 was previously described [23] . The Ad5-VII-loxP virus was generated in the background of pTG3602 in the following way . A subgenomic clone of Ad5 from nucleotides ( nt ) 12 , 290–22 , 340 served as an intermediate vector for the introduction of loxP sites flanking the pVII open reading frame by conventional PCR cloning . The 5’ loxP site was introduced at nt 15 , 875 , three nucleotides upstream of the pVII initiation codon and three nucleotides downstream of the protein III ( penton ) stop codon . The 3’ loxP site was introduced at nt 16 , 475 , immediately follows the stop codon for pVII . LoxP-containing Ad DNA was recombined with pTG3602 that had been digested with PmeI and AsiSI restriction enzymes , as described [78] . Following confirmation of clones by nucleotide sequence analysis , pTG3602-VII-loxP was digested with PacI , transfected into HEK-293 cells ( ATCC ) , and plaque assays performed . Two independent plaques were amplified and the introduction of the loxP sites was confirmed; these viruses were named Ad5-VII-loxP-5 and Ad5-VII-loxP-11 . Stock lysates were generated , titered by plaque assay , and virus particles purified using cesium chloride equilibrium gradient centrifugation , as described [79] . Virus infections were performed for 1 h at 37°C , as described [79] , unless otherwise noted . The parental Ad5-VII-loxP-5 and Ad5-VII-loxP-11 viruses were used to infect 293 cells at 5 plaque forming units/cell to yield VII+ viruses ( VII gene intact ) or used to infect Cre66 cells ( Stefan Kochanek , University Ulm , Germany ) to yield VII−mutant viruses ( VII gene deleted ) and harvested 2–3 days after infection when full cytopathic effect was evident . Efficiencies of recombination ( floxing ) for VII−virus were determined using viral DNA extracted from CsCl-purified virus particles by qPCR utilizing two sets primer pairs: 1 ) Ad5 nt 44–63 and 280–261 to amplify Ad5 left-end sequences to quantify total viral DNA , and 2 ) Ad5 nt 16 , 155–16 , 173 and 16 , 333–16 , 315 to amplify sequences within the VII reading frame to quantify viral DNAs with intact VII gene sequences ( primer sequences in Supplemental Information ) . The relative amount of VII+ genomes in the VII−virus preparations was calculated as described [80] . Low molecular Hirt DNA and purified viral DNA also was analyzed by Southern blot following digestion with KpnI . Southern blots were probed using a DNA fragment including Ad5 nt 15 , 658–16 , 887 , as described [79] . HeLa cells ( ATCC ) were infected with the Ad5-VII-loxP viruses , VII+ or VII– , at 200 virus particles/cell , unless otherwise noted . The following primary antibodies were rabbit polyclonal unless indicated and were used at the following dilutions: anti-hexon , mouse monoclonal 9C12 ( University of Iowa Developmental Studies Hybridoma Bank , 1:100; rabbit anti-pVII/VII , 1:2000 ( Dr . Daniel Engel , University of Virginia ) ; rabbit anti-V , 1:1000 ( Dr . David Matthews , University of Bristol ) ; rabbit anti-penton and rabbit anti-fiber , 1:1000 ( Dr . Carl Anderson , Brookhaven National Laboratory ) ; rabbit anti-IIIa , 1:1000 , [81]; rabbit anti-VIII , 1:400 ( Drs . Ann Tollefson and William Wold , St . Louis University ) ; rabbit anti-AVP , 1:500 ( Dr . Maxim Balakirev , CEA-Grenoble ) ; rabbit anti-VI , 1:5000 ( Dr . Christopher Wiethoff , Loyola University Chicago ) ; rabbit anti-pVI C-terminal peptide amino acids 240–250 , 1:100 and rabbit anti-VIII C-terminal peptide amino acids 214–227 , 1:1000 , ( Maarit Suomalainen and Urs Greber , University Zurich , Switzerland ) ; rabbit anti-IVa2 and anti-L1-52/55K 1:1000 [82]; and mouse anti-α-tubulin monoclonal antibody , Sigma-Aldrich T5192 . Whole cell extracts ( WCE ) were prepared by washing cells twice with phosphate buffered saline ( PBS ) followed by cell resuspension in 2X SB ( 120mM Tris pH 6 . 8 , 4% sodium dodecyl sulfate ( SDS ) , and 20% glycerol ) and incubation at 100°C for 10 min . Proteins from virus particles were prepared following ethanol precipitation of CsCl-purified virions [83] by resuspension in 2X SB and boiling . Protein concentrations were determined using the bicinchoninic acid ( BCA ) protein assay kit ( Thermo Scientific ) . Either 15 μg of WCE or 1 . 3 μg of protein from purified virus particles were separated on 12 . 5% or 15% SDS-polyacrylamide gels and transferred overnight to nitrocellulose membranes at 4°C . Membranes were blocked in 3% bovine serum albumin in TBS ( 50mM Tris pH 7 . 5 , 150mM NaCl ) for 1 hour . Membranes were treated with primary antibodies for 1 hour at room temperature or overnight at 4°C and washed 5 times with TBS containing 0 . 05% Tween20 for 5 min each at room temperature . Secondary antibodies were diluted 1:5000 in 5% powdered milk in TBS , and membranes were treated for 1 hour at room temperature followed by washes as described above . Two additional washes with TBS minus Tween were done before scanning using an Odyssey system ( LiCor ) . IRDye 800CW-conjugated goat anti-rabbit IgG or IRDye 600CW-conjugated goat anti-mouse IgG ( LiCor ) were used as secondary antibodies for Western blots . SDS-polyacrylamide gels of proteins from particles were stained with silver nitrate using the method of Dr . Darrick Carter ( www . proteinchemist . com ) . At times post-infection indicated in the text , 1-2x106 infected cells were chilled on ice for 10 min , harvested by scraping , washed twice with PBS , and divided for DNA or RNA isolations . Whole cell DNA was isolated using the DNeasy Blood and Tissue Kit ( Qiagen ) . DNAs were quantified by absorbance at 260nm . Whole cell RNA was prepared by lysing cells using the QIAshredder ( Qiagen ) followed by isolation of RNA by RNeasy Plus Mini Kit ( Qiagen ) . cDNAs were generated by priming with oligo ( dT ) ( NEB ) and reverse transcription using SuperScript II RT ( Invitrogen ) following the manufacturer’s instructions . qPCR was performed using the DyNAmo HS SYBER green qPCR kit ( Thermo Scientific ) and amplifying using the 7500 Real Time PCR System ( Applied Biosystems ) . Reactions contained either approximately 40-80ng of purified DNA or 1/10th of the cDNA product . Results were analyzed using the 7300 system software ( Applied Biosystems ) . Primer pairs for quantification of Ad5 DNA , the Ad5 VII gene , and the cellular GAPDH gene are listed in S1 Table . Primer pairs for cDNA quantification of Ad5 E1a , E2a , E2B , E4 , and cellular GAPDH are listed in S1 Table . Standards for absolute quantification were pTG3602 for Ad5-related sequences and subclones of GAPDH specific for either cDNA or gene quantifications . Levels of DNA were determined by dividing the absolute amount of Ad5 DNA by the absolute amount of GAPDH DNA and are graphed on a log scale . Relative levels of Ad early transcripts were calculated using the method of Pfaffl [80] using the absolute amount obtained for viral early transcripts divided by the absolute amount of GAPDH transcripts . For DBP and pVII immunofluorescence microscopy , A549 or HeLa cells grown on glass coverslips were harvested at 24 hours post-infection and processed for immunofluorescence as described [19] using mouse anti-DBP monoclonal antibodies A6-1 and B6-8 [84] and rabbit anti-pVII antibody [31] . Images were captured and analyzed using a Zeiss Axiovert 200M digital deconvolution microscope with AxioVision 4 . 8 . 2 SP3 software . For copper ( I ) -catalyzed azide alkyne cycloaddition staining and immunofluorescence microscopy , cells grown on glass coverslips were fixed at the times indicated in the text with 3% paraformaldehyde for 15 min , quenched with 25 mM ammonium chloride , permeabilized with 0 . 5% Triton X-100 at room temperature for 5 min , and labeled with mouse anti-hexon monoclonal antibody 9C12 ( University of Iowa Developmental Studies Hybridoma Bank ) followed by anti-mouse AlexaFluor 594-conjugated secondary antibody . Samples were stained with freshly prepared click staining mix containing 10 μM AlexaFluor 488-azide , 1 mM CuSO4 , and 10 mM sodium ascorbate in PBS in the presence of 1 mM THPTA , and 10 mM amino-guanidine ( AG ) , for protection against oxidative damage , at RT for 2 hr in the dark . Samples were stained with 4′ , 6-diamidino-2-phenylindole ( DAPI , Molecular Probes , Leiden ) for total DNA , embedded in DAKO medium ( Dako Schweiz AG , Baar ) for imaging by confocal microscopy . Fluorescence images were recorded on Leica SP5 confocal laser scanning microscope or Zeiss LSM510 Meta confocal system . Images shown represent maximum projections of confocal stacks . Samples were processed for transmission electron microscopy ( TEM ) as previously described [85–87] . In brief , cells grown on glass coverslips were fixed in 1 . 5% glutaraldehyde–2% formaldehyde , 0 . 1M sodium cacodylate pH 7 . 4 for 60 min , followed by post-fixation in 1% OsO4 and 1 . 5% potassium-ferricyanide in deionized water at room temperature for 1 hour , several washes in in 0 . 1M sodium cacodylate and contrasting 1% tannic acid in 0 . 05M sodium cacodylate , followed by a 5 min incubation in 1% sodium sulfate in 0 . 05M sodium cacodylate . The samples were rinsed in deionized water for 5 min , stained with 2% uranyl-acetate overnight , dehydrated with acetone , embedded in Epon , ultrathin-sectioned , and analyzed in a Zeiss EM10 equipped with an advanced interline technology CCD camera Erlangshen ES500W-782 ( Gatan GmbH , Munich , Germany ) . The number of viruses at the plasma membrane , in endosomes , in the cytosol , and at the nuclear membrane was determined by manual counting . DNAs from CsCl-purified Ad particles were obtained by ethanol precipitation followed by digestion with proteinase K in the presence of 0 . 5% SDS , several phenol/chloroform extractions , and ethanol precipitation . DNA was isolated from two different VII- virus particle preparations , from VII+ virus particles , and from Ad5-WT virus particles . HeLa cells were transfected with viral DNAs and pcDNA3-EGFP using Lipofectamine 2000 ( Invitrogen ) and cells were harvested 48 hours post-transfection . DNAs and RNAs were isolated as described above . DNAs were digested with DpnI to fragment input , transfected DNA . DNA and RNA were quantified by qPCR and RT-qPCR , respectively . Primer pairs for DpnI-digested total DNA were Ad5 nt 882–901 and 1052–1033 and primer pairs for cDNA were within the E1a gene . All numerical values represent mean ± sd . Each experiment was done in three replicates , and a representative replicate is shown for each blot . Statistical significance of the differences was calculated using student’s t-test . | The Ad major core protein VII protects the viral genome from recognition by a cellular DNA damage response during the early stages of infection and alters cellular chromatin to block innate signaling mechanisms . The packaging of the Ad genome into the capsid is thought to follow the paradigm of dsDNA bacteriophage where viral DNA is inserted into a preassembled capsid using a packaging motor . How this process occurs if Ad packages a DNA-core protein complex is unknown . We analyzed an Ad mutant that lacks core protein VII and demonstrated that virus assembly and DNA packaging takes place normally , but that the mutant is deficient in the maturation of several capsid proteins and displays a defect in the escape of virions from the endosome . These results have profound implications for the Ad assembly mechanism and for the role of protein VII during infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"hek",
"293",
"cells",
"biological",
"cultures",
"microbiology",
"viral",
"structure",
"dna",
"replication",
"dna",
"molecular",
"biology",
"techniques",
"microbial",
"genomics",
"research",
"and",
"analysis",
"methods",
"viral",
"genomics",
"viral",
"core",
"proteins... | 2017 | The adenovirus major core protein VII is dispensable for virion assembly but is essential for lytic infection |
Trypanosoma brucei , the causative agent of Human African Trypanosomiasis ( HAT ) , expresses two proteins with homology to human glycogen synthase kinase 3β ( HsGSK-3 ) designated TbruGSK-3 short and TbruGSK-3 long . TbruGSK-3 short has previously been validated as a potential drug target and since this enzyme has also been pursued as a human drug target , a large number of inhibitors are available for screening against the parasite enzyme . A collaborative industrial/academic partnership facilitated by the World Health Organisation Tropical Diseases Research division ( WHO TDR ) was initiated to stimulate research aimed at identifying new drugs for treating HAT . A subset of over 16 , 000 inhibitors of HsGSK-3 β from the Pfizer compound collection was screened against the shorter of two orthologues of TbruGSK-3 . The resulting active compounds were tested for selectivity versus HsGSK-3β and a panel of human kinases , as well as in vitro anti-trypanosomal activity . Structural analysis of the human and trypanosomal enzymes was also performed . We identified potent and selective compounds representing potential attractive starting points for a drug discovery program . Structural analysis of the human and trypanosomal enzymes also revealed hypotheses for further improving selectivity of the compounds .
Human African trypanosomiasis ( HAT ) and the lack of effective therapy constitute a health concern in 36 countries of sub-Saharan Africa [1] . The disease affects predominantly poor populations and transmission has been attributed to exposure during activities such as agriculture , animal husbandry , or hunting [2] , which are the major means of livelihood in endemic regions . Following acute infection , the disease progresses to a chronic phase ultimately with invasion of the brain . This can happen within a month of initial infection , or alternatively can take years , depending on the parasite sub-species [3] . Four drugs , Eflornithine , Suramin , Pentamidine and Melarsoprol , are currently licensed for the treatment of HAT [4] , [5] . Unfortunately , these are toxic and difficult to administer , limiting therapeutic choices [6] . Thus , new therapies for HAT are urgently needed . Protein kinases , estimated to represent over 30% of all drug discovery programs , remain one of the most studied drug targets for a number of human and animal diseases [7]–[10] . More than 500 protein kinases have thus far been identified , many of which are linked to disease processes [11] . Of particular interest here is a serine/threonine glycogen synthase kinase -3 ( GSK-3 ) , which plays a role in the regulation of glycogen metabolism [12] , [13] , WNT signaling [14] , cell cycle regulation [15] , [16] and other processes . HsGSK-3 has been investigated as a drug target for several diseases including Alzheimer's disease [17] , neurodegeneration and oncogenesis [18] . Two isoforms of GSK-3 exist in human cells , HsGSK-3 alpha and HsGSK-3 beta . These human isoforms display a high degree of sequence identity with only one amino acid difference ( Glu196 in HsGSK-3 alpha and Asp133 in HsGSK-3 beta ) in the ATP binding domain [19] , [20] . Previous studies [21] demonstrated that the causative agent of HAT , Trypanosoma brucei , expresses two proteins ( TbruGSK-3 short and TbruGSK-3 long ) with homology to HsGSK-3 . The shorter protein isoform was shown to be essential for parasite growth and viability and inhibitors of TbruGSK-3 short were found to kill mammalian-stage T . brucei . The authors concluded that evolutionary variations in the ATP binding domain of TbruGSK-3 short , relative to HsGSK-3 beta , might allow for designing parasite selective inhibitors . HAT drug development is challenged by the disproportionately small commercial interest and investment in developing new anti-parasite agents relative to other human diseases like cancer [22] . This study involved a collaborative Public-Private Partnership ( PPP ) facilitated by WHO TDR between researchers at University of Washington , USA , University of Antwerp , Belgium , and Pfizer Global Research Development , Sandwich , UK , to find specific inhibitors of T . brucei using a target-based high throughput screening ( HTS ) approach . This is an example of how drug development for neglected diseases can be stimulated by the PPP approach . A panel of 16 , 540 putative inhibitors previously associated with projects at Pfizer targeting HsGSK-3 was screened against the recombinant TbruGSK-3 short . Selected hits were counter-screened against HsGSK-3β . Kinase panel specificity and anti-parasitic screening were also conducted . Compounds identified in this study provide useful starting points for further chemical optimisation .
Recombinant TbruGSK-3 short ( accession number Tb10 . 161 . 3140 ) was produced at the University of Washington [21] and shipped to Pfizer , Sandwich , UK for testing . Kinase-Glo reagent ( Promega ) was used as previously described [23] . This luciferase coupled assay , which provides a luminescent quantification of ATP consumed during the kinase reaction , was modified to a 384-well plate screening format . A selected library of 16 , 540 compounds comprising known HsGSK-3 beta inhibitors and close structural analogues was screened at 10 µM final assay concentration . Assay plates were prepared by dispensing 0 . 2 µL of compound ( dissolved in 100% DMSO ) from master plates into white 384-well plates ( Greiner bio one ) . Primary screening was conducted in a 20 µL reaction volume . Enzyme was added to each well to a final concentration of 3 . 8 nM in a volume of 10 µl using a Multidrop Combi dispenser ( Thermo Scientific ) and the plates were incubated for 15 minutes at room temperature ( RT ) . Glycogen synthase peptide 2 ( BioGSP2; Sigma ) and ATP , were dissolved in 20% acetonitrile and 1 M Tris-HCl pH 7 . 6 respectively , then diluted in assay buffer to a final concentration of 3 . 2 µM BioGSP2 and 2 µM ATP . The assay buffer consisted of 25 mM Tris-HCl pH 7 . 5 , 10 mM MgCl2 , 5 mM DTT , 0 . 1 mg/mL BSA , 2 U/mL Heparin and 10 µM EDTA . The reaction was initiated by adding 10 µL of the substrate mixture to each well and allowed to proceed at RT for 2 h . Twenty microlitres of Kinase-Glo reagent was added to quench the reaction . Luminescence was measured after 1 h at a 100 millisecond/well integration time using the Acquest Multimode plate reader ( Molecular Devices ) . Each plate included a positive control ( 4 µM GW8510 , Sigma ) and negative control ( 1% DMSO ) . Hit compounds were further titrated using a through-plate IC50 format with a maximum concentration of 25 µM . The data was analysed using Pfizer SIGHTS software and visualised using Spotfire software ( TIBCO ) . Five separate 384-well plates were screened in duplicate to assess the assay reproducibility . Human GSK-3 beta inhibition data ( IC50 ) for many of the compounds were recovered from Pfizer data files . If historical data were not available , the compounds were tested in an assay using 10 nM HsGSK-3 beta ( Invitrogen ) using Omnia Kinase Assay ( Invitrogen ) according to the manufacturer's instructions . The reaction volume was 20 µL and a range of compound concentrations were tested , up to a maximum of 40 µM . Briefly , 5 µl of HsGSK-3 beta was dispensed into black 384 assay plates ( Greiner bio one ) containing 0 . 2 µl of compounds . The enzyme was incubated with the compounds for 15 minutes at 30°C then 15 µl of substrate mixture was added to each well to commence the reaction . The substrate mixture consisted of 2 µL each of 2× kinase reaction buffer , 10 µM Omina peptide substrate , 0 . 2 mM DTT and 10 µM ATP , and 7 µL of ultra pure water . The reaction was allowed to proceed for 30 minutes at 30°C . Increase in fluorescence levels indicating peptide phosphorylation by the enzyme was monitored using an Envision ( PerkinElmer ) with λex 360/λem 485 nm and the data were analysed using Pfizer software SIGHTS and Spotfire ( TIBCO ) . Compounds with TbruGSK-3 short IC50<100 nM were tested for their ability to inhibit the proliferation of T . brucei ( blood stage form ) . Cytotoxicity testing against human fetal lung fibroblast MRC-5 cell line was also performed . Both assays were carried out with compound concentrations up to 64 µM at the Laboratory for Microbiology , Parasitology and Hygiene , University of Antwerp ( www . ua . ac . be ) . Briefly , T . brucei trypomastigotes ( Squib-427 strain , suramin-sensitive ) were cultured in Hirumi-9 medium supplemented with 10% fetal calf serum at 1 . 5×104 trypomastigotes per well . Following 72 hours incubation , parasite growth was assessed fluorimetrically by addition of resazurin . For cytotoxicity evaluation , 104 MRC-5 cells/well were seeded onto the test plates containing the pre-diluted compounds and incubated at 37°C with 5% CO2 for 72 hours . Cell viability was determined fluorimetrically after addition of resazurin [24] , [25] . Single point kinase panel screening was also conducted on selected compounds at 10 µM by Invitrogen ( www . Invitrogen . com ) and University of Dundee , UK ( www . dundee . ac . uk ) . The crystal structure of human GSK-3 beta complexed with staurosporine ( pdb entry 1q3d ) was used as the basis for modelling work . Selected compounds were docked into the crystal structure of HsGSK-3 beta on the basis of binding modes of related known ligands . The binding-site residues were aligned and the residues that differ between human and TbruGSK-3 in the catalytic pocket were highlighted with different colours . Images were created using the Pfizer molecule-modelling package MoViT .
A high throughput 384-well assay was developed for Tbru GSK-3 short which measures ATP depletion following phosphorylation of the peptide substrate BioGSP-2 . The previously identified inhibitor of TbruGSK-3 short , GW8510 [21] , was used as a positive control . The assay yielded Z and Z' scores of 0 . 2 and 0 . 8 , respectively , indicating excellent quality [26] . Assay reproducibility in HTS format was confirmed by duplicate testing of 5 separate 384-well plates which produced an identical number of hits ( Figure 1A ) . A collection of 16 , 540 compounds targeting HsGSK-3 beta were selected from Pfizer compound library and screened against TbruGSK-3 short at a concentration of 10 µM . In order to capture all potential actives , compounds conferring above 40% inhibition were considered hits , giving an overall hit rate of 8 . 6% ( Figure 1B ) . Hits were titrated in the screening assay , revealing 1 , 317 hits with IC50<25 µM . Of these confirmed hits , 362 compounds had IC50<1 µM and 35 compounds had IC50<100 nM . The IC50 data against HsGSK-3 beta were either recovered from Pfizer records or the titration was conducted on selected compounds . A comparative analysis of inhibitor potencies between TbruGSK-3 short and HsGSK-3 beta is presented in Figure 1C . A majority of the compounds exhibited greater potency against the human enzyme which is not surprising , since the initial library was primarily made up of compounds that had been optimized for binding to HsGSK-3 beta . Compounds were clustered with an in-house algorithm that carries out single-linkage clustering , whereby any pair of compounds sharing a Tanimoto similarity value of 0 . 7 ( calculated using Daylight fingerprints ) were placed in the same cluster ( Daylight Chemical Information System Inc ) . These hits were expanded by selecting near neighbour analogues from the Pfizer compound libraries and further titrating them against both HsGSK-3 beta and TbruGSK-3 short . Two compounds , 0181276 and PF-04903528 , were found to show 7-fold selective inhibition of TbruGSK-3 short compared to HsGSK-3 beta ( Table 1 and Figure 2 ) . Seventeen compounds with TbruGSK-3 short IC50 values of <100 nM ( regardless of selectivity ) were tested for their ability to inhibit the proliferation of mammalian-stage T . brucei . Specificity for the parasite was investigated by testing against the human fetal lung fibroblast MRC-5 cell line ( Table 1 and Figure 2 ) . Ten compounds showed in-vitro inhibition of T . brucei proliferation with EC50s of <1 µM and 6 had EC50s of 1–3 µM . Several of the most potent compounds also showed potent inhibition of the MRC5 cell line . However , six compounds showed at least a 5-fold window between T . brucei activity and activity on MRC5 cells , particularly CE-317112 which had 35-fold selectivity ( Table 1 ) . In general , potent inhibition of TbruGSK-3 enzyme activity correlated with potent activity against the whole parasite . However , CE-160042 which was a potent inhibitor of TbruGSK-3 enzyme activity , showed no inhibition of the whole parasite ( EC50 >25 µM ) . We subsequently discovered that this compound showed no detectable cell permeability in a standard CaCo2 cell flux assay used routinely in drug discovery ( data not shown ) and therefore the lack of activity is most likely due to the compound failing to reach the target within the parasite . Human kinase inhibitors often inhibit more than one kinase leading to safety issues . In order to understand the kinase inhibition profile of TbruGSK-3 inhibitors , 13 of the compounds were screened at 10 µM against a panel of approximately 40 human kinases . One of the compounds , CE-160042 , was highly specific and only inhibited HsGSK-3 beta ( Figure 3 ) . PF-4279731 and 0180532 were also relatively specific showing >50% inhibition of only 2 and 4 other kinases , respectively . The remaining compounds were active against more than 10 other kinases . Previous modelling of the Tbru GSK-3 active site identified a number of residues that differ between the human and parasite enzyme that could potentially be exploited to achieve selective inhibition . Using the published enzyme structures [21] , the predicted binding modes of two of our compounds were examined ( Figure 4 ) . This demonstrated that of the previously reported binding site differences , only one , T . bru M101/Hs L132 is in close proximity to the compound binding site and therefore is likely to be the key residue for achieving selectivity . The modelling suggests that greater selectivity could be achieved by making compounds with substituents that have improved interaction with methionine compared to leucine at this position .
We have exploited knowledge of the essentiality of TbruGSK-3 short and the availability of a large number of HsGSK-3β inhibitors to initiate a drug discovery program for Human African Trypanosomiasis . Over 16 , 000 compounds were screened against TbruGSK-3 short isoform and compounds of interest were tested against HsGSK-3 beta , whole parasites and human cells . Specificity against a panel of approximately 40 human kinases was also evaluated . We identified 2 compounds with approximately 7-fold selectivity for TbruGSK-3 short over HsGSK-3 beta: PF-04903528 and 0181276 . One of these , 0181276 was also relatively specific against the wider human kinase panel . CE-160042 was not selective against the parasite enzyme , but was completely selective for HsGSK-3 beta and showed no significant inhibition of any other kinases . In addition , CE-317112 showed a 35-fold safety window relative to the cytotoxicity control . Together , these compounds represent an attractive starting point for medicinal chemistry with a focus on further improving selectivity for a drug discovery program . Using structural modelling , we have shown that improved selectivity may be possible by exploiting the T . bru M101/Hs L132 active site difference . Given that this is a relatively small difference , highly selective compounds may be difficult to obtain , however it is encouraging that our intitial screening has identified compounds with 7-fold selectivity . Previous studies suggest that in vivo inhibition of mammalian GSK-3 causes no significant changes in body weight , food consumption or any associated adverse effects , as judged by histopathology or blood chemistry analyses [27] , [28] . Therefore , low levels of specificity may be tolerated . However , mouse knock-out studies of GSK-3 beta have shown embryonic lethality due to liver degeneration and changes in bone development [29] , [30] . Consequently , non-selective inhibitors would not be safe for use in pregnant women , infants and young children . Therefore , selective inhibitors of the parasite enzyme would be highly desirable and the availability of the GSK-3 structural models provides a powerful tool for structure assisted compound design which could guide synthesis of more selective compounds , based on the initial 7-fold selective compounds we have identified . This early drug discovery collaboration was facilitated by WHO TDR and demonstrates the power of such public private partnerships in bringing together the drug discovery expertise of pharma companies , the detailed target knowledge from academia and access to parasite biological assays from expert screening centers to accelerate drug discovery for neglected tropical diseases . Our most promising compounds are disclosed to accelerate the pace of drug development for HAT . | Over 60 million people in sub-Saharan Africa are at risk of infection with the parasite Trypanosoma brucei which causes Human African Trypanosomiasis ( HAT ) , also known as sleeping sickness . The disease results in systemic and neurological disability to its victims . At present , only four drugs are available for treatment of HAT . However , these drugs are expensive , limited in efficacy and are severely toxic , hence the need to develop new therapies . Previously , the short TbruGSK-3 short has been validated as a potential target for developing new drugs against HAT . Because this enzyme has also been pursued as a drug target for other diseases , several inhibitors are available for screening against the parasite enzyme . Here we present the results of screening over 16 , 000 inhibitors of human GSK-3β ( HsGSK-3 ) from the Pfizer compound collection against TbruGSK-3 short . The resulting active compounds were tested for selectivity versus HsGSK-3β and a panel of human kinases , as well as their ability to inhibit proliferation of the parasite in vitro . We have identified attractive compounds that now form potential starting points for drug discovery against HAT . This is an example of how a tripartite partnership involving pharmaceutical industries , academic institutions and non-government organisations such as WHO TDR , can stimulate research for neglected diseases . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"infectious",
"diseases",
"medical",
"microbiology",
"global",
"health",
"african",
"trypanosomiasis",
"neglected",
"tropical",
"diseases",
"biology",
"microbiology",
"parasitic",
"diseases",
"parasitology"
] | 2011 | Trypanosoma brucei Glycogen Synthase Kinase-3, A Target for Anti-Trypanosomal Drug Development: A Public-Private Partnership to Identify Novel Leads |
Src tyrosine kinases are deregulated in numerous cancers and may favor tumorigenesis and tumor progression . We previously described that Src activation in NIH-3T3 mouse fibroblasts promoted cell resistance to apoptosis . Indeed , Src was found to accelerate the degradation of the pro-apoptotic BH3-only protein Bik and compromised Bax activation as well as subsequent mitochondrial outer membrane permeabilization . The present study undertook a systems biomedicine approach to design optimal anticancer therapeutic strategies using Src-transformed and parental fibroblasts as a biological model . First , a mathematical model of Bik kinetics was designed and fitted to biological data . It guided further experimental investigation that showed that Bik total amount remained constant during staurosporine exposure , and suggested that Bik protein might undergo activation to induce apoptosis . Then , a mathematical model of the mitochondrial pathway of apoptosis was designed and fitted to experimental results . It showed that Src inhibitors could circumvent resistance to apoptosis in Src-transformed cells but gave no specific advantage to parental cells . In addition , it predicted that inhibitors of Bcl-2 antiapoptotic proteins such as ABT-737 should not be used in this biological system in which apoptosis resistance relied on the deficiency of an apoptosis accelerator but not on the overexpression of an apoptosis inhibitor , which was experimentally verified . Finally , we designed theoretically optimal therapeutic strategies using the data-calibrated model . All of them relied on the observed Bax overexpression in Src-transformed cells compared to parental fibroblasts . Indeed , they all involved Bax downregulation such that Bax levels would still be high enough to induce apoptosis in Src-transformed cells but not in parental ones . Efficacy of this counterintuitive therapeutic strategy was further experimentally validated . Thus , the use of Bax inhibitors might be an unexpected way to specifically target cancer cells with deregulated Src tyrosine kinase activity .
Protein tyrosine kinases of the Src family are involved in multiple facets of cell physiology including survival , proliferation , motility and adhesion [1] . Their deregulation has been described in numerous malignancies such as colorectal , breast , melanoma , prostate , lung or pancreatic cancers and is known to favor tumorigenesis and tumor progression [2]–[4] . Modulation of apoptosis sensitivity by Src deregulation is more controversial . We recently described that Src activation promotes resistance to the mitochondrial pathway of apoptosis in mouse and human cancer cell lines [5] . The molecular mechanism underlying such resistance involved the accelerated degradation of the proapoptotic BH3-only protein Bik . Indeed , in Src-transformed NIH 3T3 mouse fibroblasts , Bik was found to be phosphorylated by activated Erk1/2 , which was followed by Bik subsequent polyubiquitylation and proteasomal degradation [5] . Thus in Src-transformed cells , Bik downregulation compromised Bax activation and mitochondrial outer membrane ( MOM ) permeabilization upon an apoptotic stress [5] . That observation might be of importance since MOM permeabilization is the key step that commits cells to apoptosis . Indeed , MOM permeabilization leads to the irreversible release of cytochrome c and other cytotoxic molecules from the mitochondrial inter-membrane space into the cytosol [6] , [7] . Once released , cytochrome c induces the formation of the apoptosome complex , which triggers caspase activation , these molecules being the main executioners of the apoptotic program . MOM permeabilization is triggered by the insertion and oligomerization of the pro-apoptotic effector Bax into the membrane [8]–[11] . Antiapoptotic proteins such as Bcl-2 or Bcl-xL prevent this process , whereas pro-apoptotic BH3-only proteins contribute to Bax activation [6] , [11]–[16] . Using western blotting and specific shRNAs , the respective contribution of the different Bcl-2 family members to the cell response triggered by various death- inducing agents was assessed in parental and Src-transformed NIH-3T3 fibroblasts [5] . Experimentally and mathematically investigating the cell response to death-inducing agents might be of interest since it has long been postulated that restoration of apoptosis might be an effective way to selectively kill cancer cells . The rationale of this assumption is that cancer cells need to counteract the pro-apoptotic effect of oncogenes such as Myc or E2F-1 that stimulate cell proliferation as well [17] . Moreover , Src deregulation has specifically been associated with resistance to treatment in a number of cancers [18] , [19] . Therefore , a critical clinical concern lies in the design of therapeutic strategies that would circumvent resistance to apoptosis of cells with deregulated Src activity . To this end , several classes of therapeutic agents might be a priori considered . Inhibitors of Src tyrosine kinases , such as dasatinib , are currently widely used in the clinic [20]–[23] . Other anticancer therapeutic strategies aim at restoring apoptosis in cancer cells [24] . In particular , inhibitors of antiapoptotic proteins such as ABT-737 or the Oblimersen Bcl-2 antisense oligodeoxyribonucleotide are currently evaluated in clinical trials [25]–[28] . Apoptosis may also be restored by increasing the expression of pro-apoptotic proteins such as Bax , Bik or p53 [29]–[32] . Here we propose a systems biology approach for optimizing potential anticancer therapeutic strategies using parental and Src-transformed NIH 3T3 fibroblasts as a biological model . To this end , molecular mathematical models of Bik kinetics and of the mitochondrial pathway of apoptosis were built and fitted to available experimental data . They guided further experimental investigation in parental and Src-transformed cells which allowed their refinement . Then , those models were used to generate predictions which were validated by subsequent specifically-designed experiments . Finally , we theoretically explored different drug combinations involving the kinase inhibitor staurosporine , Src inhibitors , and activators or inhibitors of the Bcl-2 protein family , in order to design optimal anticancer strategies for this biological system . Optimal strategies were defined as those which maximized the efficacy on Src-transformed cells considered as cancer cells under the constraint of toxicity remaining under a tolerable threshold in parental cells .
We recently provided evidence that Bik , a BH3-only protein , is a key regulator of apoptosis in the considered biological system [5] . Therefore we first built a mathematical model to investigate Bik kinetics in non-apoptotic conditions . Bik concentration temporal variations were assumed to result from two processes: protein formation and protein polyubiquitylation , which eventually leads to its degradation . Let us denote and the intracellular concentration of Bik and polyubiquitylated Bik proteins respectively , expressed in nM . Bik protein was assumed to be synthesized at a constant rate in both Src-transformed and parental cells as suggested by similar Bik mRNA level in both cell types [5] . Concerning Bik ubiquitylation , we considered that it occurred either spontaneously at the rate , or after Bik phosphorylation by activated Erk1/2 downstream of SRC activation , as demonstrated in Src-transformed fibroblasts ( [5] , Figure 1 ) . In those cells , this prior phosphorylation increased Bik ubiquitylation rate and further proteasomal degradation . This Src-dependent pathway was modeled by Michaelis-Menten kinetics with parameters and . In the model , we assumed that spontaneous and Src-mediated ubiquitylation could occur in both transformed and parental cells . Ubiquitin molecules were assumed to be in large excess compared to Bik amount . Therefore ubiquitin concentration was considered as constant and implicitly included in and . Poly-ubiquitylated molecules were then assumed to be degraded by the proteasome at a constant rate in both cell types . was arbitrary set to 1 as it does not influence kinetics , and only acts on . The model of Bik kinetics can be written as follows: ( 1 ) ( 2 ) Parameters were then estimated for parental and Src-transformed cells by fitting experimental results on Bik protein degradation in both cell types ( [5] , reprinted with permission in Figure 2B ) . We assumed that the spontaneous phosphorylation occurred at the same rate in parental and Src-transformed fibroblasts and therefore looked for a unique . Parameters of Src-dependent Bik degradation were denoted and in parental cells and and in Src-transformed 3T3 cells . Inhibition of Src kinase activity by herbimycin was experimentally monitored in Src-transformed cells ( Figure 2A , reprinted with permission from [5] ) . Herbimycin exposure achieved a decrease of 98% in phosphorylated Y416 amount . Therefore , we modeled herbimycin exposure as a decrease of 98% in values . See Text S1 for details on the parameter estimation procedure . The best-fit parameter value for the spontaneous ubiquitylation was . Src-dependent ubiquitylation was predicted to be inactive in normal fibroblasts as , which was in agreement with experimental results [5] . On the contrary , the Src pathway was predominant in transformed cells as and which leads to The dynamical system 1–2 admits a unique steady state:where . For parental cells , in which is equal to zero , steady state becomes . Bik steady-state concentrations in parental cells was assumed to be equal to 50 nM which is in the physiological range of BH3-only protein intracellular levels [33]–[37] . This allowed us to deduce . We then computed Bik steady-state concentrations in Src-transformed cells which was equal to nM . Thus , the simulated ratio of Bik concentration in Src-transformed cells over that in parental cells was equal to 0 . 18 which is similar to the experimentally-observed value quantified to 0 . 2 ( Figure 3 , Table 1 ) . This constitutes a partial validation of the model since the data of Figure 3 was not used in Bik kinetics model design and calibration . In the following , these steady state concentrations were used as Bik initial condition since cells were assumed to be in non-apoptotic conditions prior to the death stimulus . We then investigated Bik kinetics in parental and Src-transformed NIH-3T3 cells in response to an apoptotic stress that consisted of a 8-hour-long exposure to staurosporine ( 2 M ) . As demonstrated by knockdown experiments ( Figure 2b in [5] ) , Bik was required for apoptosis induction . Bik was present in non-transformed cells with no sign of apoptosis in normal conditions , which suggested either that Bik concentration was not large enough to trigger apoptosis in these conditions , or that Bik was activated upon apoptotic stress . The first assumption to be mathematically investigated was that Bik protein amount might increase upon staurosporine treatment as a result of the turning-off of the degradation processes , Bik synthesis rate remaining unchanged under staurosporine treatment . Thus , if Bik ubiquitylation process is turned off in the model , only the formation term remains in equation 1 which is now the same for parental and transformed cells , with different initial conditions . This equation can be solved analytically: ( 3 ) where stands for Bik initial concentration taken equal to Bik steady state concentration in parental and transformed fibroblasts . We did not observe any significant apoptosis either in normal or Src-transformed cells in the first six hours of staurosporine treatment ( data not shown ) . In non-transformed cells , setting t = 360 min in equation 3 gave which meant that Bik concentration would only double in six hours if this hypothesis was right . This was tested by measuring Bik protein level during staurosporine exposure in parental cells . However , no significant increase in Bik levels upon a 6 hour-long staurosporine treatment was observed , which ruled out that the induction of apoptosis could depend on Bik accumulation ( Figure 4 A ) . Therefore , we investigated a second hypothesis that consisted of an activation of Bik upon apoptosis induction . Such a possibility might rely on a release of Bik from a protein complex upon apoptotic stress as observed with other BH3-only proteins such as Bad , Bim or Bmf [38] . To investigate the likelihood of this hypothesis , we performed the immunostaining of endogenous Bik in parental NIH-3T3 cells upon staurosporine exposure . Our data was in agreement with a relocation of Bik from its known location at the ER to the mitochondria within 2 h of treatment ( [39] , [40] , Figure 4 B ) . This relocation might correspond to Bik release from a binding protein at the ER as previously observed [41] . We modeled this relocation by the equations 4 and 5 in which stands for Bik protein that had been activated possibly through this relocation and represents inactive Bik molecules . This translocation occurred at the rate . Colocalization between Bik fluorescence and mitotracker staining showed that 4513% of Bik molecules were located at the mitochondria within 2 h of treatment which led to the estimated value ( Figure 4 B ) . We then investigated the mitochondrial pathway of apoptosis in NIH-3T3 parental and Src-transformed cells . We only considered the Bcl-2 members that were experimentally detected in this biological model [5] . The only pro-apoptotic multidomain effector was Bax , whereas the multidomain antiapoptotic protein family was represented by Bcl-2 , Bcl-xL and Mcl-1 [5] . Five BH3-only proteins were present: three BH3-only activators ( i . e . able to directly bind and activate Bax ) Puma , Bim and tBid and two BH3-only sensitizers ( i . e . able to bind Bcl-2 and related apoptosis inhibitors , but unable to bind and activate Bax ) Bad and Bik . The respective role of present BH3-only proteins in apoptosis induction was assessed by a shRNA-mediated approach . Bim , which was expressed at very low level , could be neglected in the onset of apoptosis , since its downregulation induced no significant increase in apoptosis resistance upon staurosporine , thapsigargin or etoposide . In contrast , PUMA had a prominent role for apoptosis induced by genotoxic stresses ( UV or etoposide ) but displayed no significant role in staurosporine- and thapsigargin-induced apoptosis [5] . As we focused here on staurosporine-induced apoptosis , the only BH3-only activator that we considered was tBid . Concerning BH3-only sensitizers , Bad could be neglected as its silencing by shRNA did not significantly modify cell response to staurosporine . Therefore the only sensitizer to be considered was Bik . Bax , Bik and tBid were described to bind all the antiapoptotic proteins expressed in our biological model , namely Bcl2 , Bcl-xL and Mcl-1 . Thus , for the sake of simplicity , we denoted by the cumulative concentration of those three antiapoptotic proteins . We then modeled Bax activation . In non-apoptotic conditions , Bax spontaneously adopts a closed 3D-conformation that does not bind antiapoptotic proteins [10] . This conformation was denoted . During apoptosis , Bax transforms into an opened 3D-conformation ( ) and inserts strongly into the MOM . molecules can be inhibited by antiapoptotic proteins which trap them into dimers . Moreover , they may spontaneously transform back into their closed conformation [42] . If they are not inhibited , molecules may oligomerize into molecules and create pores in the MOM which correlates with the release into the cytosol of apoptogenic factors , including cytochrome C [8]–[10] . We considered that was inefficient at binding oligomerized Bax [7] . In the model , Bax oligomerization happens either by the oligomerization of two molecules or by a much faster autocatalytic process in which a molecule recruits a molecule to create two molecules . Those two processes occurred at the respective rates and which were chosen such that to account for the preponderance of the autocatalytic pathway . Bax activation from into isoforms was assumed to be catalyzed by the BH3-only activator . We assumed that this reaction occurred in a “kiss and run” manner and therefore follows Michaelis-Menten kinetics . resulted from activation by truncation which occurred at the rate [43] . BH3-only activator can also be inhibited by which trap it into complexes . Those complexes may be dissociated by active Bik molecules which bind to and release [13] . Finally , we also considered that antiapoptotic proteins directly inhibit active Bik molecules and associate into complexes . Above-mentioned chemical reactions that occur spontaneously were assumed to follow the law of mass action . All protein concentrations are expressed in nM in the mathematical model . This mathematical model is recapitulated in Figure 5 and Table 2 . It can be written as follows: ( 4 ) ( 5 ) ( 6 ) ( 7 ) ( 8 ) ( 9 ) ( 10 ) ( 11 ) ( 12 ) ( 13 ) ( 14 ) Bik total protein amount was assumed to be constant during apoptosis as experimentally demonstrated ( Figure 4 A ) . We also assumed that , and total amounts remained constant following the death stimulus . However , the apoptotic stress may induce Bax transcription and repress Bcl2 one , in particular through the activation of p53 [44] . Four conservation laws hold: Only seven from the eleven equations of the mathematical model 4–14 need to be solved as the four remaining variables can be computed using those conservation laws . We subsequently modeled the cell population behavior . Let us denote by the percentage of surviving cells at time t . No cell division was assumed to occur in presence of staurosporine as the very first effect of most cytotoxic drug consists in stopping the cell cycle [45] . Natural cell death was neglected as almost no apoptosis was observed in either parental or Src-transformed cells in the absence of death stimuli [5] . We considered that apoptosis is irreversibly activated when concentration reaches the threshold which corresponds to the minimal amount of oligomerized Bax molecules required to trigger the cytochrome C release into the cytosol . This assumption was modeled in equation 15 by a S-shape function which also ensures that the death rate does not grow to infinity . Below is the equation for the percentage of surviving cells: ( 15 ) Parameters , a and were assumed to be the same for the two cell populations . At the initial time just before the apoptotic stress , cells were assumed to be in steady state conditions . The initial percentage of surviving cells is . Bik initial concentrations were set to steady-state values computed using equations 1–2 . Moreover , we assumed that Bik was entirely under its inactive form so that: , , . All Bax molecules are assumed to be inactive: , and . All existing molecules are trapped in complexes with : and . Initial protein concentration of Bid and Bcl2 can be computed using the conservation laws: and . For the sake of simplicity , we considered that no complexes were present at the initial time ( ) as dimers do not play any part in the overall dynamics since we assumed that they do not dissociate . As previously stated , the considered apoptotic stress consists of an 8-hour-long exposure to staurosporine ( 2 M ) which starts at time t = 0 . It triggers two molecular events: activation into and formation representing truncation into . Mathematically , and are set to non-zero values at the initial time . Parameters of this model of mitochondrial apoptosis were estimated by fitting experimental data in parental and Src-transformed cells from [5] and integrating biological results from literature . First , we assessed quantitative molar values of considered Bcl-2 family proteins in non-apoptotic conditions as follows . We set Bax total concentration in Src-transformed cells to 100 nM according to [46] in which the authors stated that this was a physiological level in tumor cells . This value is also in agreement with concentration ranges found in the literature [33] , [34] , [36] , [47] , [48] . Then , in [46] , they found that anti-apoptotic total concentration had to be 6 times higher than that of Bax in order to prevent apoptosis . Therefore , we set in Src-transformed cells . Concerning Bid total concentration , we assumed which is in agreement with experimental results from the literature [33] , [34] , [36] , [47] . Finally , tBid initial concentration was set to 1 nM since this band was hardly detectable by western-blot ( Figure 3 ) . Moreover , this value was consistent with previous modeling results [47] . We then computed protein ratios between parental and Src-transformed cells using immunoblotting data of Figure 3 . We experimentally determined that there was a 9-fold higher amount of proteins in the cytosol fraction compared to the mitochondria compartment which allowed us to compute protein ratios of total intracellular quantities ( Table 1 ) . As previously stated , Bik protein amount was reduced in Src-transformed cells by a factor of 0 . 2 compared to parental fibroblasts ( Figure 3 ) . This dramatic decrease was the result of the Src-dependent activation of Erk1/2 kinases , leading to Bik phosphorylation , polyubiquitylation and subsequent degradation by the proteasome [5] . Bax steady-state level in non-apoptotic conditions was increased by a factor of 2 . 1 in Src-transformed cells compared to normal ones and that of Bid was decreased by a factor of 0 . 77 . Concerning antiapoptotic molecules , the sum of Bcl2 , Bcl-xL and Mcl-1 quantities was slightly increased in Src-transformed cells by a factor of 1 . 1 compared to parental ones . Those protein ratios were used to compute molar quantities of considered Bcl-2 family protein total concentrations ( Table 1 ) . Then , we estimated the apoptotic threshold as follows . Quantification of Figure S1d in [5] showed that 38% of BAX molecules at the mitochondria were activated during apoptosis . Previously-described quantification of Figure 3 showed that 33% of BAX total amount were located at the mitochondria , the remaining part being in the cytosol . Therefore , the percentage of activated BAX was set to 33% * 38%/10013% . This percentage is in agreement with previous experimental data which suggests that approximately 10–20% of Bax total amount is actually activated during apoptosis [46] . The high intensity of the bands corresponding to Bcl-xL expression in Figure 3 suggested that it might be the predominant antiapoptotic protein in our biological model . Dissociation constant between Bcl-xL and respectively Bik , Bid and Bax was experimentally found to be equal to nM , nM and nM [49] , [50] . Therefore , we set and and only estimated . At this point , 10 kinetic parameters still needed to be estimated . In order to determine those 10 parameters , we fitted experimental data from Figure 6 under constraints inferred from experimental results . We used the three experimental data points of Figure 6 corresponding to exposure to staurosporine as a single agent or combined with herbimycin . We modeled the administration of staurosporine after an exposure to the Src tyrosine kinase inhibitor herbimycin as follows . We assumed that herbimycin was administrated before staurosporine exposure such that the system had time to reach steady state . As previously described , herbimycin exposure was modeled by decreasing ( the maximal velocity of Src-induced Bik ubiquitylation ) of 98% of its original value . Then , we set constraints on state variables as follows . We assumed that did not decrease below 20% ( i . e . the apoptotic threshold ) of its initial value within 6 h of staurosporine exposure as approximately 20% of Bax total quantity is activated during apoptosis [9] , [46] . Moreover , we ensured that reached the apoptotic threshold in parental cells after 6 to 8 h of exposure to staurosporine as biological experiments showed . Moreover , as Bax oligomerization was assumed to be an autocatalytic process , we expected to obtain . Therefore , in the parameter estimation procedure , we set initial search values for and such that . Finally , molecule association rates were searched between and which is a realistic range with respect to the diffusion limit [48] . Estimated parameter values are shown in Table 2 . The data-fitted mathematical model allowed the investigation of the dynamical molecular response to staurosporine exposure ( Figure 7 ) . As expected , higher Bik concentration in normal fibroblasts led to a higher concentration of and of free compared to transformed cells . molecules then activated into which oligomerized until reaching the apoptotic threshold in parental cells . On the contrary , could efficiently be sequestered in complexes with antiapoptotic proteins in Src-transformed cells as a result of the lower level of Bik protein . This perfectly fit the described function of Bik as a sensitizer [51] . Concerning co-administration of staurosporine and herbymicin , the model predicted that this drug combination circumvents the resistance of the cancer cell population in which 99% of cells are apoptotic after 8 hours of exposure to staurosporine ( Figure 6 ) . This model behavior was in agreement with experimental data which showed 98% of apoptotic cells in the Src-transformed population . Moreover , the model predicted that an exposure to staurosporine as a single agent or combined with herbimycin lead to the same activity of 80% of apoptotic cells in the parental fibroblasts population . We intended to determine optimal therapeutic strategies for our particular biological system in which parental and Src-transformed NIH-3T3 fibroblasts stand for healthy and cancer cells respectively . In the following , both cell populations are exposed simultaneously to the same drugs , mimicking the in vivo situation in which healthy and tumor tissues are a priori exposed to the same blood concentrations of chemotherapy agents . From a numerical point of view , identical parameter changes were applied to normal and cancer cells . First , we investigated the combination of staurosporine with ABT-737 , a competitive inhibitor of Bcl-2 and Bcl-xL that were the main antiapoptotic proteins in our cellular model . ABT-737 inhibits free antiapoptotic proteins but also dissociates complexes of anti- and pro-apoptotic proteins . As for herbimycin , we assumed that ABT-737 was administrated before staurosporine such that the system had time to reach steady state . ABT-737 pre-incubation was thus modeled by decreasing Bcl-2 total amount in proportion to ABT-737 concentration and by setting and at the initial time . Interestingly , ABT-737 exposure in the absence of staurosporine ( i . e . ) did not result in cell death induction for any dose of ABT-737 in the mathematical model , as experimentally demonstrated [5] . Indeed , in the absence of staurosporine , Bid was not activated into tBid and the low quantity of tBid present in the cells at steady state was not sufficient to trigger Bax oligomerization , even when ABT-737 inhibited all anti-apoptotic proteins . This confirmed that the mathematical model described correctly this cell model that does not behave as a “primed for death model” in which inhibition of anti-death proteins results in death , even in the absence of apoptosis induction . As a reminder , in the primed for death situation , incubation with ABT-737 led to cell death as a consequence of the release of the BH3-only pro-apoptotic proteins that were therefore able to activate Bax . The main difference between the primed for death situation and our model is that apoptosis resistance in the primed for death model primary comes from the overexpression of anti-apoptotic proteins such as Bcl-xL or Bcl2 that are efficiently inhibited by ABT-737 whereas here it comes from the decrease of a pro-death protein in the Src-transformed model . The combination of staurosporine and ABT-737 at any concentration , i . e . for any decrease in Bcl2 total protein amount , was predicted by the model to induce much more apoptosis in parental cells compared to Src-transformed cells and thus to fail in circumventing cancer cells resistance ( Figure 8 ) . To experimentally confirm this model prediction , we pre-incubated parental and Src-transformed cells with ABT-737 prior to staurosporine exposure . The resulting death-inducing effect on Src-transformed cells was significantly increased compared to staurosporine alone ( Figure 6 ) . However , as anticipated by the model , this drug combination resulted in an extremely high toxicity of 99% of apoptotic cells in the parental fibroblasts population ( Figure 6 ) . Those data points were reproduced by the calibrated mathematical model for a predictive decrease of 182 nM in Bcl2 total concentrations in both cell types . After that , we looked for theoretically optimal therapeutic strategies by applying optimization procedures on the calibrated model of the mitochondrial apoptosis . Optimal strategies were defined as those which maximized efficacy in cancer cells under the toxicity constraint that less than 1% of healthy cells die during drug exposure . We investigated drug combinations that consisted of the exposure to staurosporine after treatment with Src inhibitors , or up- or down-regulators of BCL-2 family proteins . Pre-incubation with inhibitors or activators aimed at modifying the equilibrium of the biological system before exposure to the cytotoxic drug . Src inhibition was simulated by a decrease in value whereas up- or down-regulation of Bcl-2 family proteins were modeled by modifying the total concentration of the targeted proteins . The theoretically-optimal drug combination would consist of administering staurosporine combined with inhibitors of Src , Bax and Bcl2 , together with a upregulator . The concentration of Bax inhibitor should be set such that Bax total concentration decreases below the apoptotic threshold in healthy cells thus protecting them from apoptosis . As Bax total amount was higher in cancer cells , it would remain high enough to allow these cells to undergo apoptosis . Once healthy cells are sheltered from apoptosis , Bcl2 amount could be decreased , using for instance ABT-737 , and amount increased at the same time without risking any severe toxicity . As expected , the optimal therapeutic strategy also included the suppression of the Src-dependent phosphorylation of Bik in cancer cells , using for instance herbimycin . This drug combination led to 99% of apoptotic cells in the cancer cell population and less than 1% in the parental one where Bax was hardly present ( Figure 9 , Text S1 ) . This theoretically optimal strategy involved the administration of a cytotoxic agent combined with four other chemicals , which may not be realistic in the perspective of clinical application . Therefore we hierarchically ranked the considered therapeutic agents by searching for optimal strategies consisting in the combination of staurosporine with only one or two agents . Strategies which satisfied the tolerability constraint ( i . e . less than 1% of apoptotic parental cells ) and reached an efficacy value of 99% of apoptotic cells all involved Bax downregulation in addition to a second agent among Bcl2 downregulator , upregulator and Src inhibitor ( See Text S1 for more details ) . Of note , isolated decrease of Bax total amount fulfilled the tolerability constraint but resulted in less than 1% of apoptotic cancer cells . Finally , we experimentally validated feasibility of this counterintuitive theoretical strategy . We selected two siRNAs that fully downregulated Bax in parental cells but not in Src-transformed ones ( Figure 10 A ) . Bax knockdown protected parental cells from treatment by staurosporine and ABT737 or staurosporine and herbimycin but not Src-transformed cells ( Figure 10 B ) . Therefore by downregulating Bax in our biological model , we were capable of selectively killing Src-transformed cells .
A combined mathematical and experimental approach was undertaken to study the mitochondrial pathway of apoptosis in parental and Src-transformed NIH-3T3 cells . First , a mathematical model for Bik kinetics in normal and apoptotic conditions was built . It took into account Bik ubiquitylation and further proteasomal degradation that Src-dependent Bik phosphorylation stimulated in Src-transformed cells . Then , we designed a mathematical model of the mitochondrial pathway of apoptosis which only involved the proteins that participated in apoptosis induction in the studied biological model . Interestingly , this mathematical model was quite simple , with only one effector , Bax , two BH3-only proteins , Bik ( a sensitizer ) and tBid ( a direct Bax activator ) , and a pool of antiapoptotic proteins which were all described as behaving identically toward Bax , Bik and tBid [38] . Several published works propose mathematical modeling of apoptosis . Some of them model all pathways to apoptosis from the death stimulus to the actual cell death [52]–[54] , other focus on the caspase cascade leading to apoptosis [55] . Molecular modeling of the mitochondrial pathway was achieved in several works [47] , [48] , [56]–[59] . Those models being conceived to address other biological issues , we had to build a new mathematical model that was tailored to our particular problematic and aimed at optimizing anticancer therapies in the specific case of Src transformation . Exploring Bik kinetics upon apoptosis induction led to the interesting prediction that the inhibition of Bik degradation might not allow its accumulation above a threshold that would induce apoptosis in the experimentally-demonstrated time range . This was validated by immunoblotting that established that Bik concentration was not changed upon apoptosis induction by staurosporine . Therefore , we looked for another explanation that might support these observations . A possibility was that Bik might undergo activation upon apoptosis induction . Activation of BH3-only proteins has already been described and can depend on phosphorylation/dephosphorylation as observed for Bad or Bim or on proteolytic activation as for Bid . These post-translational modifications usually result in a change of cell compartment , from cytosol to mitochondria for Bad , from cytoskeleton to mitochondria for Bim . Therefore , we looked for a clue pointing towards Bik activation during apoptosis induction . Indeed , we observed that a significant part of Bik translocated from the ER , which is its normal location , to mitochondria upon staurosporine treatment . This could be due to the release of Bik from a protein complex with the ER protein GRP78 as described in [41] . However , we could not rule out that this relocation might also be linked to the hyperfusion and subsequent mitochondria fission process observed upon a number of apoptosis stresses [60] or more specifically associated with Bik [61] . The mathematical model of apoptosis induction was used to explore anticancer therapies . First , it confirmed that Src inhibitor circumvented resistance to staurosporine exposure of Src-transformed cells , which was experimentally demonstrated . It also showed that this therapeutic strategy did not give any specific advantage to parental cells . It also predicted that inhibitors of antiapoptotic proteins should not be co-administered with staurosporine in our particular biological system as a result of the slightly lower antiapoptotic protein concentration in Src-transformed cells compared to parental ones . This model prediction was experimentally validated . Indeed , here , resistance to apoptosis comes from the decrease in a pro-apoptotic protein in Src-transformed cells and not from the overexpression of antiapoptotic proteins as frequently observed , which explains why a drug inhibiting antiapoptotic proteins could not target specifically transformed cells . We then investigated theoretically-optimal therapeutic strategies . Interestingly , all optimal drug combinations took advantage of the observed Bax overexpression in Src-transformed cells . The optimal therapeutic strategies consisted of the combination of a cytotoxic agent for the induction of apoptosis ( staurosporine in the model ) with Bax downregulator and one agent among Src inhibitor , Bcl2 inhibitor or tBid upregulator . We experimentally validated this counterintuitive prediction by demonstrating in our biological model that Bax knockdown could protect parental cells but not Src-transformed cells from combinations of staurosporine with antiapoptotic protein or Src tyrosine kinase inhibitor . This strategy might be challenging in clinics since siRNA targeting Bax would need to have the exact same pharmacokinetic properties as inhibitors of tyrosine kinase or of antiapoptotic proteins to target all the places where there is expected to be cytotoxicity . Moreover it would be of importance to also downregulate Bak , due to redundancy between Bax and Bak [62] . It would also be of interest to check for Bax/Bak upregulation in various tumors . Therefore , this study supports the need for further development of Bax/Bak inhibitors that might open a therapeutic window for kinase or antiapoptotic protein inhibitors that would otherwise be globally harmful for the patient when combined with a cytotoxic treatment .
Parental and v-src transformed NIH-3T3 cells were cultured as described in [5] . When indicated , cells were exposed to staurosporine ( 2 M ) during 8 hours and pre-incubated overnight with herbimycin ( 1 M ) or ABT-737 ( 1 M ) prior to staurosporine exposure . Immunoblotting was performed as described in [5] . For the analysis of Bik relocalization to mitochondria upon staurosporine exposure , cells were pre-incubated with Mitotracker-red ( 1∶10 , 000 ) for 20 min before a 20 min washing in fresh medium . Then , cells were fixed by addition of paraformaldehyde ( 4% ) . Mitotracker-red background was removed by a 10 min incubation in acetone at −20°C prior further processing according to standard procedures . Endogenous Bik was detected by the specific anti-Bik-BH3 antibody ( 1∶200 ) . The secondary antibody ( Molecular Probe ) was labeled with FITC . Nuclei were stained with Hoechst-33342 dye ( 1∶50 , 000 ) in the mounting medium . Images were acquired by confocal microscopy on a Axiovert 100 M , LSM510 ( Zeiss ) using a Plan Apochromat 63x/1 . 4 Oil DIC objective . Colocalization of Bik fluorescence and mitotracker staining was quantified using ImageJ software in 30 single cells treated with staurosporine for 2 h . siBax1 ( MMC . RNAI . N007527 . 12 . 2 ) and siBax2 ( MMC . RNAI . N007527 . 12 . 4 ) were purchased from Integrated DNA Technologies . 100 . 000 cells were transfected with 10 nM siRNA using Lipofectamine RNAimax ( Life Technologies ) . 48 h later , cells were treated overnight either by ABT737 or herbimycin . The day after , apoptosis was induced by staurosporine treatment ( exposure to 2 M during 8 hours ) . Matlab ode15s solver was used to solve the systems of ordinary differential equations . Parameter estimation for Bik kinetics model consisted of a least-square approach in which minimization tasks were performed by the CMAES algorithm ( [63] , Text S1 ) . Parameter estimation for the mitochondrial apoptosis model involved the design of a cost function which was also minimized by the CMAES algorithm . This cost function was the sum of a Least square term accounting for the fitting of experimental data of Figure 6 and a term for the constraints described in the Results section . This term was equal to zero when all constraints were satisfied and to a thousand times the number of unsatisfied constraints otherwise [64] . Optimization procedures for the design of therapeutic strategies consisted of maximizing efficacy under the constraint of toxicity not exceeding a tolerability threshold . In order to address this issue we minimized a cost function which was the sum of two terms . The first one consisted of the percentage of surviving cells in the cancer cell population which has to be minimized . The second term was equal to zero when the toxicity constraint was satisfied and took a high value ( e . g . 1000 ) otherwise . Minimization tasks were performed by the CMAES algorithm [63] . | Personalizing medicine on a molecular basis has proven its clinical benefits . The molecular study of the patient's tumor and healthy tissues allowed the identification of determinant mutations and the subsequent optimization of healthy and cancer cells specific response to treatments . Here , we propose a combined mathematical and experimental approach for the design of optimal therapeutics strategies tailored to the patient molecular profile . As an in vitro proof of concept , we used parental and Src-transformed NIH-3T3 fibroblasts as a biological model . Experimental study at a molecular level of those two cell populations demonstrated differences in the gene expression of key-controllers of the mitochondrial pathway of apoptosis thus suggesting potential therapeutic targets . Molecular mathematical models were built and fitted to existing experimental data . They guided further experimental investigation of the kinetics of the mitochondrial pathway of apoptosis which allowed their refinement . Finally , optimization procedures were applied to those data-calibrated models to determine theoretically optimal therapeutic strategies that would maximize the anticancer efficacy on Src-transformed cells under the constraint of a maximal allowed toxicity on parental cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"oncology",
"systems",
"biology",
"cell",
"death",
"medicine",
"mathematics",
"theoretical",
"biology",
"applied",
"mathematics",
"cancer",
"treatment",
"chemotherapy",
"and",
"drug",
"treatment",
"signaling",
"networks",
"biology",
"molecular",
"cell",
"biology",
"comp... | 2013 | Data-Driven Modeling of Src Control on the Mitochondrial Pathway of Apoptosis: Implication for Anticancer Therapy Optimization |
Short non-coding transcripts can be transcribed from distant-acting transcriptional enhancer loci , but the prevalence of such enhancer RNAs ( eRNAs ) within the transcriptome , and the association of eRNA expression with tissue-specific enhancer activity in vivo remain poorly understood . Here , we investigated the expression dynamics of tissue-specific non-coding RNAs in embryonic mouse tissues via deep RNA sequencing . Overall , approximately 80% of validated in vivo enhancers show tissue-specific RNA expression that correlates with tissue-specific enhancer activity . Globally , we identified thousands of tissue-specifically transcribed non-coding regions ( TSTRs ) displaying various genomic hallmarks of bona fide enhancers . In transgenic mouse reporter assays , over half of tested TSTRs functioned as enhancers with reproducible activity in the predicted tissue . Together , our results demonstrate that tissue-specific eRNA expression is a common feature of in vivo enhancers , as well as a major source of extragenic transcription , and that eRNA expression signatures can be used to predict tissue-specific enhancers independent of known epigenomic enhancer marks .
Development and function of mammalian tissues rely on the dynamic control of tissue-specific gene expression , a process largely regulated by distant-acting transcriptional enhancers [1]–[3] . Disruption of enhancer sequences can lead to severe phenotypes in mouse models [4]–[9] . Furthermore , population-scale genetic studies indicate that a large proportion of sequence variants associated with human diseases affect non-coding functions in the genome , of which enhancers are a major category [10] . Despite their functional relevance , the genome-scale identification of enhancers that are active in vivo in developmental and disease processes remains challenging . In principle , genome-wide profiling of enhancer-associated epigenomic marks ( e . g . H3K27ac and CBP/p300 ) enables the genome-scale identification of enhancers predicted to be active in a given cell type or tissue [2] , [3] , [11]–[15] . However , none of these marks is unique to enhancer regions or found at all enhancers and ChIP-based technology has well-documented limitations with sensitivity and specificity [3] , [14]–[16] . Recently , expression of short non-coding transcripts has been described as a feature of many enhancers with a possible tight correlation between cell type-specific enhancer activity and eRNA expression levels [17]–[20] . Using cap analysis of gene expression ( CAGE ) in a collection of human tissues and cell type , Andersson et al . [21] identified over 40 , 000 candidate enhancers marked by bidirectional capped RNA expression suggesting that RNA transcription can provide a complementary approach for de novo enhancer discovery . Anecdotal evidence suggests a functional requirement for such eRNAs in enhancer-mediated gene regulation [22] , [23] . Regardless of the molecular mechanisms underlying eRNA-mediated regulatory functions , the prevalence of eRNA transcription at the whole transcriptome level in vivo and whether eRNA expression signatures can potentially be used as an independent mark for in vivo enhancer discovery remain poorly explored . In this study , we compare eRNA expression profiles determined via total RNA sequencing across developmental mouse tissues and demonstrate highly tissue-specific genome-wide expression signatures of eRNAs in vivo . We find that eRNA expression globally correlates with tissue-specific enhancer activity and that RNAs transcribed from in vivo enhancers constitute a major proportion of tissue-specifically expressed non-coding RNAs . Finally , we demonstrate through application of reporter assays in transgenic mice that differential expression of eRNAs can correctly predict tissue-specific in vivo enhancer activities independent of other chromatin-associated marks .
To test the hypothesis that eRNA transcription marks active in vivo enhancers in a tissue-specific manner , we first measured eRNA expression from 15 intergenic enhancers active in mouse embryonic forebrain or limb buds that were randomly selected from a larger collection of previously identified in vivo enhancers [24] . We assessed eRNA expression from each enhancer by quantitative RT-PCR across three different embryonic mouse tissues including forebrain , limb , and heart as a negative control ( Figure 1 ) . While baseline expression of each eRNA was detected in all three tissues , in 80% of cases eRNAs from tissue-specific enhancers showed highest expression in the predicted tissue compared with the other two tissues ( 12/15; p = 0 . 0006 , Fisher's exact test ) , suggesting that eRNAs are commonly expressed from tissue-specific developmental enhancers with a quantitative relationship between eRNA transcription and tissue-specific enhancer activity . To study eRNA expression from in vivo enhancers beyond this small-scale qPCR screen , we examined genome-wide total RNA transcription in embryonic heart and limb , two tissues with different developmental origins and trajectories , and with divergent in vivo enhancer landscapes as assessed by epigenomic marks [25]–[27] . We extracted total RNA from limb and heart tissues microdissected at mouse embryonic day [E] 11 . 5 . Following ribosomal RNA depletion , we used a strand-specific total RNA sequencing protocol to generate more than 200 million sequencing reads from each tissue ( see Methods , Table S1 ) . While the majority of sequencing reads ( 53% in heart , 60% in limb ) mapped to annotated mouse cDNA sequences , a considerable proportion ( 38% in heart , 30% in limb ) mapped to introns as well as intergenic regions , consistent with a possible association with in vivo enhancers . Examination of individual genomic loci containing known enhancers revealed examples of bidirectional tissue-specific eRNA expression from validated intergenic and intragenic enhancers consistent with their in vivo activity ( Figure 2A and Figure S1 ) . These results indicate widespread transcription from non-coding sequences in vivo and anecdotally support correlation of in vivo enhancer activity with tissue-specific eRNA transcription . In order to assess tissue-specific eRNA expression more systematically , we examined eRNA expression associated with a large collection of in vivo-validated tissue-specific enhancers [24] , [26] , [28] ( http://enhancer . lbl . gov ) . To avoid confounding factors arising from the presence of pre-mRNAs , we restricted this analysis to intergenic in vivo enhancers ( see Methods ) . We examined a total of 145 such enhancers that are active in heart or limb . In general , enhancers were substantially enriched in uniquely mapped reads , and they were nine times as likely as random non-coding regions to contain ten or more independent reads within 1 kb of the enhancer midpoint ( p = 5 . 5E-108 based on background distribution; see Table S2 and Methods ) . While 41% of enhancers met this stringent threshold , overall 92% of enhancers showed evidence of at least weak transcription ( ≥1 uniquely mapped reads; p = 2 . 3E-15 based on background distribution , see Table S2 and Methods ) . Consistent with our small-scale sampling of enhancers by quantitative PCR ( Figure 1 ) , 79% of heart enhancers and 83% of limb enhancers showed higher eRNA expression in the tissue where enhancer activity was observed in vivo ( Figure 2B; p<10−8 , Fisher's exact test ) . We next examined tissue-derived RNA signatures at intergenic regions enriched for enhancer-associated p300 and H3K27ac epigenomic marks [27] , [29] from the same tissues ( see Methods ) . Similar to known in vivo enhancers , eRNA transcription was highly enriched around the center regions defined by ChIP-Seq , and tissue-specific eRNA expression patterns correlated with the predicted enhancer activity based on tissue-specific p300 or H3K27ac signature in the same tissues ( Figure 2C–F; See Methods ) . This global correlation between tissue-specific eRNA expression and enhancer activity corroborates previous observations derived from CAGE analysis of human cell types and tissues [21] and supports the possibility that eRNA expression profiling from tissues may provide an effective approach for identifying tissue-specific in vivo enhancers . To explore the potential of eRNA profiling for de novo enhancer discovery , we first used a sliding window approach to identify candidate intergenic regions enriched for RNA expression . Known coding and intronic regions and unannotated transcripts were removed , which led to the identification of 3 , 422 and 3 , 775 intergenic regions in heart and limb , respectively , that showed marked RNA expression at a conservatively chosen threshold of ≥10 uniquely mapped reads ( see Methods; Figure S2A–B and Table S2 ) . These regions included 834 heart-specific and 1 , 078 limb-specific loci ( tissue-specifically transcribed regions , TSTRs ) that were differentially expressed in these two tissues ( Figure 3A–B and Table S3 ) . Most of these ∼2 , 000 TSTRs were located distal to the nearest transcription start site ( Figure S2C ) . There is substantial overlap between TSTRs identified from developing mouse tissues in this study and candidate transcription start sites ( TSSs ) captured by CAGE from mouse cells and tissues [30] . Overall , 45% of heart TSTRs and 55% of limb TSTRs overlap with at least one CAGE-derived TSS candidate . This represents a strong enrichment compared to random control sequences ( 8% and 8 . 3% , respectively; p<4 . 3E-68 , Fisher's exact test , see Methods ) , but also indicates that large numbers of additional enhancer candidates were identified by analysis of ex vivo tissue at relevant developmental stages . Tissue-specific expression of a panel of 22 candidate TSTRs was tested and in all cases confirmed by quantitative RT-PCR ( Figure 3C–D , see Methods ) , demonstrating that these RNA-seq data sets accurately identified non-coding TSTRs across tissues . To assess whether these TSTRs may represent in vivo enhancers , we first examined their evolutionary sequence constraint , a feature associated with many distant-acting enhancers [25] , [31] , [32] . We found that 69% and 73% of TSTRs in heart and limb , respectively , overlap with elements under evolutionary constraint as compared to 28% and 27% of random control sequences ( p<2 . 0E-62 , Fisher's exact test; Figure 4A and Figure S2D ) . Additionally , heart TSTRs are enriched near genes critical for cardiovascular and heart development , whereas limb TSTRs are enriched near genes involved in muscle tissue development and limb development/morphogenesis ( Table 1 ) . Heart and limb TSTRs are also enriched for different sets of transcription factor binding motifs related to development of the respective tissues compared with random genomic sequences ( Table S5 and Table S6 ) . Finally , we compared tissue-specific TSTR expression with mRNA levels of nearby genes in two tissues ( see Methods ) . The strongest correlation was observed between TSTRs and their nearest genes ( Pearson correlation: R = 0 . 68 for heart , R = 0 . 55 for limb ) , and decreased substantially for more distant genes ( Figure 4B ) . These results support that TSTRs may represent regulatory elements coordinating the transcription of nearby genes . To evaluate the overlap of TSTRs with enhancer-associated epigenomic marks , we examined p300 and H3K27ac enrichment ( Figure 4C–D ) . We find that 36% and 46% of heart and limb TSTRs are marked by p300 and/or H3K27ac . TSTRs with and without epigenomic enhancer marks show similar expression level and substantial evolutionary constraint ( Figure 4A and Figure S3A–B ) . However , the transcription of TSTRs with enhancer marks tends to be more balanced in both directions , whereas TSTRs marked by tissue-specific RNAs only are more biased toward one direction ( Figure S3C–D ) . In addition , TSTRs negative for p300 and/or H3K27ac are more distal to the nearest transcription start sites ( Figure S3E ) . These results indicate a substantial overlap of extragenic TSTRs with enhancer-like regions . However , this does not exclude the possibility that subsets of the observed TSTRs represent other classes of regulatory elements or unannotated non-coding loci . To directly assess the potential of TSTRs identified by transcriptome profiling for the de novo discovery of tissue-specific in vivo enhancers , we used a transgenic mouse enhancer assay previously shown to reliably capture in vivo enhancer activity [25] , [27] , [33] . In an initial retrospective comparison , we found that heart- or limb-specific TSTRs overlap with 12 tested elements that had previously been examined due to increased conservation or enhancer associated epigenomic marks [24] . Of these elements , 9/12 ( 75% ) were annotated as tissue-specific positive enhancers in vivo ( Table S7 , http://enhancer . lbl . gov ) . Next , we performed transgenic mouse assays for another set of 19 TSTRs that had not previously been tested ( Table S8 ) and exhibited tissue-specific RNA expression . This panel included elements both with and without detectable p300 and/or H3K27ac signal in ChIP-Seq experiments ( Table S8 ) that were chosen blind to the identity of nearby genes . Mouse genomic DNA for individual TSTRs with up to 2 kb of flanking sequence was cloned upstream of a minimal heat shock promoter fused to a lacZ reporter gene and transgenic mice were assayed by whole-mount staining for the expression of lacZ reporter at E11 . 5 [25] ( see Methods ) . Only elements that drove reproducible reporter gene expression pattern in at least three embryos were considered positive enhancers . In total , 8/19 ( 42% ) candidate enhancers predicted by tissue-specific RNA expression functioned as positive enhancers in vivo ( Figure 5 , Table S8 and Figure S4 ) . In all cases , the observed tissue-specific in vivo enhancer activity was consistent with the tissue specificity of the corresponding TSTR . As representative examples , transgenic whole-mount embryos and transverse sections for elements mm1052 , mm1018 , mm1054 and mm1064 are shown in Figure 5 . In these examples , reproducible LacZ reporter activities were detected in both atrial and ventricular regions of the heart ( Figure 5A–C ) and anterior regions of the fore- and hindlimb ( Figure 5D ) . Combining the results from newly performed enhancer assays and retrospective comparisons with pre-existing in vivo data sets , 17 of 31 TSTRs ( 55% ) represented in vivo enhancers , and for 15 of these 17 enhancers ( 88% ) the tissue specificity of eRNA expression correctly predicted the in vivo enhancer activity patterns . These results support the general utility of eRNA profiling as an informative mark for in vivo enhancer prediction .
Recent large-scale transcriptome studies suggest that up to 80% of mammalian genomes may be actively transcribed [34]–[37] . While many of these transcripts show differential expression signatures across cell types and tissues , the majority of non-coding transcripts have not been associated with in vivo functions . In the present study , we explored the in vivo expression dynamics of tissue-specific non-coding RNAs using a total RNA-Seq strategy that captures both coding and non-coding transcripts [18] . Our results suggest that the majority of enhancers show evidence of tissue-specific eRNA transcription . In addition , de novo identified tissue-specifically transcribed non-coding regions ( TSTRs ) showed major characteristics of canonical enhancers . These results indicate that enhancers are a predominant function associated with differentially expressed non-coding loci across developing tissues . CAGE analysis from human cell lines and tissues showed that incorporating enhancer expression data can increase the validation rate of ENCODE enhancer predictions and that bidirectional capped RNA signatures can in principle be used to identify de novo cell-specific enhancers [21] . However , in the absence of sizable in vivo validation data sets , the quantitative correlation between tissue-specific eRNA expression and in vivo enhancer activity in mammalian developmental processes has remained unclear [21] . We have tested a set of 19 candidate enhancers predicted by tissue-specific RNA expression in transgenic mouse assays and 42% showed reproducible enhancer activity in vivo , demonstrating the general utility of eRNA-based enhancer prediction in a developmental mammalian system . Of note , two of the tissue-specific enhancers reported in this study ( mm1052 and mm1061 ) did not overlap with any CAGE peaks collected from 399 mouse samples [30] despite the scope of the tissue and cell type panels examined in these previous studies . Considering the dynamics of the enhancer landscape in developing tissues and organs [29] , it appears likely that many additional enhancers active during development will be identifiable by whole transcriptome analysis of tissues across different developmental stages . While a substantial proportion of extragenic transcription appears linked to enhancer activity , our observation of several TSTRs that were not active in the transgenic enhancer reporter assays supports the hypothesis that eRNA-like transcripts can also originate from other non-coding elements , such as inactive enhancers . These observations are consistent with recent mechanistic studies on eRNAs showing that eRNA transcription precedes the establishment of H3K4me1/2 [38] , suggesting that eRNA transcription may occur before enhancer activation . TSTRs without supportive p300/H3K27ac marks show significant , though slightly decreased conservation , less bi-directional transcription , and are more distal to the nearest coding genes ( Figure S3 ) , suggesting that they may have different biological functions . Consistent with this observation , a larger proportion of TSTRs with supportive p300/H3K27ac marks were active in vivo compared to TSTRs without such marks , although this difference was not significant at the sample size examined ( p = 0 . 15 , Fisher's exact test; Table S8 ) . While the results of our study do not permit strong conclusions about the functionality of intergenic loci that exhibit transcription but no accompanying enhancer epigenomic signatures , it is possible that these regions are less likely to be active enhancers . Transcription may be occurring due to other processes or at a different class of regulatory element than active enhancers . Together , our data suggest that additional criteria such as bi-directional transcription , conservation and independent enhancer marks may further increase the performance of eRNA-based enhancer predictions . Nonetheless , considering the overall substantial correlation between TSTRs and tissue-specific in vivo enhancer activity , our results corroborate that short non-coding transcripts are commonly associated with the regulation of cell type- and tissue-specific gene expression . Enhancer RNAs may be very unstable and sensitive to exosome degradation [21] , [39] , resulting in low steady-state level in cells . This may explain why eRNAs represent a small proportion of the transcriptome profile ( Figure S2A ) , despite the large number of sites from which they originate . At current sequencing depth , many enhancers may still be missed ( Figure S2B ) , which is consistent with the notion that a great proportion of mammalian genomes may be actively transcribed and cis-regulatory genomic elements may represent major sites of extragenic non-coding transcription [34]–[37] , [39] . Recently , Andersson et al . showed that depletion of a co-factor of the exosome complex resulted in an over 3-fold average increase of eRNA abundance [21] . Thus , a combination of in-depth transcriptome profiling and exosome depletion may provide a more sensitive method for eRNA-based enhancer discovery . Emerging evidence indicates that eRNA transcripts can be required for enhancer-mediated gene activation . Targeted knock-down of specific eRNAs has been shown to affect the expression of enhancer target genes in cell-based assays , providing a potential strategy for altering gene expression in experimental and therapeutic applications [22] , [23] , [40] . Through in-depth transcriptome profiling , we have shown extensive eRNA expression in developing tissues , as well as a global correlation of eRNA expression with tissue-specific in vivo enhancer activity . Our results highlight the widespread and potentially important role of eRNAs in orchestrating gene expression , providing support for the general feasibility of eRNA-based targeting of in vivo gene expression .
Embryonic heart or limb tissue was isolated from CD-1 strain mouse embryos at E11 . 5 by microdissection in cold PBS [27] . A single sample consisting of tissue pooled from multiple embryos was analyzed for either tissue . After washing , about 1 ml TRIzol reagent ( Life Technologies , 15596-026 ) was added to every 100 mg of tissue sample , followed by homogenization using a glass dounce homogenizer . Total RNA from individual tissues were extracted following the manufacturer's instructions . Genomic DNA contamination was removed by using the TURBO DNA-free kit ( Applied Biosystems , AM1907 ) following manufacture's protocol , and the RNA samples were stored at −80°C before further processing . In order to perform the transcriptome analysis by Illumina sequencing , ribosomal RNAs was removed from total RNA ( 5∼10 µg per reaction ) by using two rounds of the RiboMinus Eukaryote Kit for RNA-Seq ( Life Technologies , A10837-08 ) following the manufacturer's instructions . The quality of total RNA after rRNA removal was analyzed on RNA 6000 Pico chip ( Agilent , 5067-1513 ) to assure that rRNA contamination was less than 30% . 100 ng total RNA after rRNA removal were used to construct the individual sequencing libraries for Illumina sequencing . Strand-specific RNA-Seq libraries were created following in-house protocols . Briefly , RNA samples were fragmented with 10×Fragment buffer ( Ambion , AM9938 ) to achieve an average fragment size of 200–300 nt . First strand cDNA synthesis was performed with random hexamer and Superscript II reverse transcriptase ( Life Technologies , 18064-014 ) . During the second strand synthesis , dUTP was used instead of dTTP to introduce strand-specificity . After adaptor ligation and size selection , the second strand containing dUTP was cleaved by AmpErase UNG ( Life Technologies , N8080096 ) . The resulting strand-specific cDNA was subjected to 12 cycles of PCR amplification and sequenced with HiSeq 2000 instrument . 50 sequencing cycles were carried out . Raw Illumina reads ( 50 bp ) were first filtered using the Illumina CASAVA-1 . 8 FASTQ Filter module ( http://cancan . cshl . edu/labmembers/gordon/fastq_illumina_filter/ ) . The remaining sequence tags were mapped back to the mouse genome ( NCBI build 37 , mm9 ) using bowtie2 [41] , and the alignments were extended to 200 bp in the 3′ direction to account for the average length of DNA fragments . Repetitively mapped reads were excluded from the following analysis . For de novo peak calling , a sliding window method EnrichedRegionMaker module from USEQ [42] was employed . For eRNA-based enhancer predictions , a conservative threshold of 10 or more reads ( without considering strand specificity ) was chosen based on the observation that in retrospective comparison with in vivo validated enhancers , 40 . 7% of enhancers met or exceeded this expression threshold , compared to 4 . 5% of random control regions ( p = 5 . 5E-108 , Table S2 ) . Enriched regions overlapping with refGene , mouse mRNA , or ESTs ( mm9 ) were also removed before the downstream analysis . This process was performed individually for heart and limb RNA-Seq data . To generate Figure S2B , 10% to 100% of sequencing reads were randomly selected from the raw sequencing data , and de novo peak calling was individually performed to identify the enriched intergenic regions . Among raw enriched regions , tissue-specifically transcribed regions ( TSTRs ) were defined as non-coding regions with significantly higher expression in this tissue compared with the other tissue ( p<0 . 01 , two-proportion z-test; Figure 3A–B ) [43] with the equation shown below:where ( n represents mappable reads within each TSTR in heart or limb , and N represent the total number of mappable reads excluding ribosomal regions in the corresponding tissue ) and . RPKM<2−9 were arbitrarily set to 2−9 for visualization purposes in Figure 3A–B . Candidate transcription start sites ( TSSs ) marked by CAGE peaks were downloaded from http://fantom . gsc . riken . jp/5/ [30] and extended to 1 kb each side from the peak midpoint . For each TSTR ( 1 kb around the peak center ) , the overlapping candidate TSSs were identified by BEDTools [44] . Random control peaks were also generated using BEDTools with the same number and size of sequences and excluding known genes , mouse mRNAs and ESTs . We compared tissue-derived RNA signatures at intergenic regions to enhancer-associated p300 [27] and H3K27ac marks from the same tissues and time-point . H3K27ac ChIP-Seq datasets are described in more detail in Nord et al . [29] and Attanasio et al . [45] . Candidate tissue-specific intergenic enhancers were predicted by ChIP-Seq of p300 ( 171 in heart , 656 in limb ) or H3K27ac ( 6965 in heart , 2174 in limb ) as described previously [27] . Briefly , uniquely aligned sequencing reads were extended to 300 bp in the 3′ direction . Enriched regions ( peaks ) were identified with MACS [46] ( p≤1E-5 ) using matched input as controls . Peaks overlapping with repetitive regions , known genes , mouse mRNAs and ESTs were removed for further analysis . Summary eRNA coverage plots were generated for p300- and/or H3K27ac-derived intergenic enhancers within a 10 kb window , centering on the maximum ChIP-seq coverage . Using the mapped reads , normalized mean eRNA coverage values were calculated for 25 bp windows across the 10 kb regions scaled by total mapped reads . For mean calculations , only the 5th–95th percentiles were used to reduce the effect of outliers . Coverage was calculated separately for antisense and sense reads , and as a combined value . For the summary plots , a loess best fit line was plotted for each of the eRNA datasets ( limb and heart ) , separating into sense and antisense reads ( Figure 2C–F ) . Pre-computed conservation scores ( phastCons scores ) generated from 30 vertebrate genome alignments were download from the UCSC Genome Browser [47] . For each TSTR ( 1 kb around the peak center ) , the conservation score was defined as the most highly constrained overlapping phastCons element in the mouse mm9 genome . Random control peaks were generated using BEDTools with the same number and size of sequences and excluding known genes , mouse mRNAs and ESTs [44] . The percentages of TSTRs and random control regions overlapping phastCons elements were plotted in Figure 4A . Tissue-specific TSTRs were classified as enriched in p300 and/or H3K27ac if the relative ChIP-seq coverage was equal to or greater than the 95th percentile of experiment background coverage estimated across 1 Mb of unique sequence . After classification , coverage heatmaps were generated for ChIP-seq data using normalized coverage values , with input corrections . Coverage was plotted for 25 bp windows centered on the peak RNA coverage and extending 25 kb on either side . For plotting purposes , coverage was centered and scaled using mean and SD in order to compare signal across datasets . TSTRs were organized as no H3K27ac and p300 signal , enriched in H3K27ac signal only , enriched in p300 signal only and enriched in both marks from the top to the bottom in Figure 4C–D . Known heart or limb enhancers were downloaded from Vista Enhancer Browser ( http://enhancer . lbl . gov ) . For known enhancer regions , the expression level of individual eRNAs was defined as the mapped sequencing reads within a 2 kb window around the center of in vivo tested enhancers . For eRNAs only expressed in one tissue , the mapped number of reads was arbitrarily set to 1 in the other tissue in order to compute the absolute fold change for plotting purposes in Figure 2B . Fold change was defined as higher expression level divided by lower expression of each eRNA in two tissues . For the volcano plot , y axis represents p-value for the expression differences of each known enhancer , which was computed by two-proportion z-test [43] . Coverage of randomly selected control regions ( excluding known genes , mRNA and ESTs ) was also computed and iterated 100 times to estimate the genome-wide background based on normal distribution . The percentages of enhancers or the average percentage of control regions with indicated numbers of uniquely mapped reads in either tissue are listed in Table S2 , as well as associated p-values . After peak calling , for each individual TSTR , normalized RPKM ( Reads Per Kilobase per Million mapped reads ) was calculated in two tissues ( heart and limb ) with the raw mapped RNA-Seq data within a 2 kb window around the center of each TSTR . Then , a tissue-specificity index was computed as ( s−u ) / ( s+u ) , in which s is the expression of TSTR in the matching tissue and u is its expression in the other tissue . The expression of mouse refGene ( mm9 ) was also analyzed in the same way by computing the RPKM across annotated cDNA regions in two tissues . The tissue-specific expression correlation between TSTRs and their nearby genes was computed as described [18] with minor modifications . Briefly , we paired each TSTR with the nearby genes . For each set of genes with the same ranked distance to TSTRs ( the first to the fifth closest genes ) , genes were ranked based on tissue-specificity indices and grouped into 20 genes per bin . Average tissue-specificity indices from each bin were used to compute the correlation . The Pearson correlation between nearby genes and the corresponding TSTRs was conducted with the statistics module in the R package ( http://cran . r-project . org/ ) . Gene ontology analysis for the genes near TSTR regions was performed by GREAT version 2 . 02 [48] . Enriched GO biological processes with a binomial p-value and fold enrichment were listed in Table 1 . For TSTRs in heart and limb , enriched motifs were computed within a 2 kb window around the center of individual TSTRs by the motif finding module of HOMER ( Hypergeometric Optimization of Motif EnRichment ) [49] . Known motifs for transcription factors with a p-value less than 10−2 compared with random genomic sequences were reporter in Table S5 and Table S6 . For directionality analysis , the expression of individual TSTRs in sense and antisense strands was defined as the strand-specific mapped sequencing reads within a 2 kb window around the center of TSTRs in either heart or limb . Then the directionality index was defined as |f−r|/ ( f+r ) , in which f is the expression of TSTR in one strand and r is its expression in the other strand in the same tissue . Total RNA was extracted from independently collected pools of heart or limb tissues with the same method as described before and synthesized into cDNA by reverse transcription using the SuperScript First-Strand Synthesis System ( Invitrogen ) . Candidate TSTRs for RT-PCR validations were randomly selected from the top 30% differentially expressed regions ranked by Z scores . Expression analysis of candidate TSTRs was carried out by real-time PCR using gene-specific primers ( Table S4 ) and KAPA SYBR FAST qPCR Master Mix ( KAPA Biosystems ) on a Roche LightCycler 480 . All primers were designed in silico using Primer3 ( http://primer3 . wi . mit . edu/ ) and tested for amplification efficiency . Target gene expression was calculated with the 2−ΔΔCT method [50] and normalized to the Gapdh housekeeping gene . Candidate enhancers for in vivo testing were selected randomly from TSTRs with a p-value less than 0 . 01 . The tested regions included up to 2 kb genomic DNA flanking the TSTRs on either sides . This general transgenic procedure has been described before [25] , [27] . Briefly , the selected regions were PCR amplified from mouse genomic DNA and cloned into the Hsp68-promoter-LacZ reporter [51] , [52] . Genomic coordinates and the PCR primers for the cloned regions are listed in Table 8 . The transgenic embryos were assayed at E11 . 5 for expression patterns . A positive enhancer is defined as an element with reproducible expression pattern in at least three embryos resulting from independent transgenic integration events [27] . For histological analysis , selected embryos were embedded in paraffin and sectioned using standard methods . RNA-seq data is available through GEO under accession number GSE58157 . In vivo transgenic data is available through the Vista Enhancer Browser under the identifiers used throughout this study ( http://enhancer . lbl . gov ) . | Up to 80% of mammalian genomes are actively transcribed , producing large numbers of non-coding RNAs without known functions . One particularly exciting category of such non-coding transcripts are the recently discovered enhancer RNAs ( eRNAs ) transcribed from distant-acting enhancer elements . Studies in cell-based paradigms suggest a functional requirement for such eRNA in enhancer-mediated gene regulation . In this study , we explored the in vivo expression dynamics of tissue-specific non-coding RNAs in embryonic mouse tissues via in-depth transcriptome profiling . Our results suggest that enhancers may be a predominant function associated with differentially expressed non-coding loci across developing tissues , and that differential eRNA expression signatures from total RNA-Seq can be used to identify uncharacterized tissue-specific in vivo enhancers independent of known epigenomic marks . Our results highlight the widespread and potentially important role of eRNAs in orchestrating gene expression and the necessity for functional studies in interpreting genome-wide enhancer predictions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genetics",
"biology",
"and",
"life",
"sciences",
"developmental",
"biology"
] | 2014 | Tissue-Specific RNA Expression Marks Distant-Acting Developmental Enhancers |
Adaptive radiation is the rapid origination of multiple species from a single ancestor as the result of concurrent adaptation to disparate environments . This fundamental evolutionary process is considered to be responsible for the genesis of a great portion of the diversity of life . Bacteria have evolved enormous biological diversity by exploiting an exceptional range of environments , yet diversification of bacteria via adaptive radiation has been documented in a few cases only and the underlying molecular mechanisms are largely unknown . Here we show a compelling example of adaptive radiation in pathogenic bacteria and reveal their genetic basis . Our evolutionary genomic analyses of the α-proteobacterial genus Bartonella uncover two parallel adaptive radiations within these host-restricted mammalian pathogens . We identify a horizontally-acquired protein secretion system , which has evolved to target specific bacterial effector proteins into host cells as the evolutionary key innovation triggering these parallel adaptive radiations . We show that the functional versatility and adaptive potential of the VirB type IV secretion system ( T4SS ) , and thereby translocated Bartonella effector proteins ( Beps ) , evolved in parallel in the two lineages prior to their radiations . Independent chromosomal fixation of the virB operon and consecutive rounds of lineage-specific bep gene duplications followed by their functional diversification characterize these parallel evolutionary trajectories . Whereas most Beps maintained their ancestral domain constitution , strikingly , a novel type of effector protein emerged convergently in both lineages . This resulted in similar arrays of host cell-targeted effector proteins in the two lineages of Bartonella as the basis of their independent radiation . The parallel molecular evolution of the VirB/Bep system displays a striking example of a key innovation involved in independent adaptive processes and the emergence of bacterial pathogens . Furthermore , our study highlights the remarkable evolvability of T4SSs and their effector proteins , explaining their broad application in bacterial interactions with the environment .
Adaptation to different ecological niches can lead to rapid diversification of a single ancestor into an array of distinct species or ecotypes . This process , called adaptive radiation , typically occurs after the arrival of a founding population in a novel environment with unoccupied ecological niches ( ‘ecological opportunity’ ) and/or by the acquisition of a novel trait ( ‘evolutionary key innovation’ ) allowing the exploitation of so far unapproachable niches [1] . Spectacular examples of adaptive radiation come from different metazoan lineages with the cichlid fishes of the East African Great Lakes and the Darwin finches on Galapagos Islands representing the most prominent examples [2] , [3] . Although known from a few cases only , bacterial lineages also underwent adaptive radiation - as documented in natural settings as well as in evolution experiments [4]–[6] . It remains a fundamental problem to biology to understand why and how certain lineages diversified; adaptive radiations , and in particular the genetic and genomic basis thereof , provide an ideal set-up to address this question [3] , [7] . One of the most fascinating aspects of adaptive radiation is the frequent occurrence of evolutionary parallelism resulting in independent adaptation to same ecological niches [8]–[10] . Such evolutionary parallelisms are excellent examples for the action of similar , yet independent , selective forces and , hence , for the key role of natural selection in evolution [11] . Furthermore , parallel adaptive radiations in a single group indicate the existence of traits conferring a high degree of adaptability allowing the group members to efficiently occupy distinct environments . Therefore , lineages that radiate in parallel are of great value to study the molecular basis of adaptation and their independent evolutionary trajectories [1] , [9] . This is of particular interest in case of host-adapted bacteria differentiating into divergent ecological niches and potentially resulting in the emergence of new pathogens . Species of the α-proteobacterial genus Bartonella are specifically adapted to distinct mammalian reservoir hosts where they cause intra-erythrocytic infections [12] . Different animal models revealed that Bartonella upon reservoir host infection colonizes a primary cellular niche from where the bacteria get seeded into the bloodstream adhering to and invading erythrocytes [13]–[15] . In most cases , infections of the reservoir host do not lead to disease symptoms suggesting a highly specific adaptation to the corresponding host niche . The transmission between host individuals is mediated by blood sucking arthropods . An integrative genome-wide analysis showed that most factors essential for Bartonella to colonize their mammalian reservoir hosts are found within the core genome of this genus [16] . This is not surprising as it reflects the common strategy used by divergently adapted species to colonize their hosts . However , this study also revealed that two type IV secretion systems ( T4SS ) , Trw and VirB , which are essential for host interaction at different stages of the infection cycle represent the few colonization factors exclusively found in the most species rich sub-lineages of bartonellae . It was assumed that the horizontal acquisition of these T4SS substantially refined the infection strategy of Bartonella facilitating concurrent adaptation to a wide range of different hosts [16] . The VirB T4SS translocates a cocktail of evolutionarily related effector proteins into host cells of the primary infection niche where they modulate various cellular processes [17]–[21] . The Trw T4SS is involved in the erythrocyte invasion by binding to the erythrocytic surface with its manifold variants of pilus subunits [22]–[24] . Here , we study the evolutionary relationship of Bartonella species adapted to distinct reservoir hosts and investigate the genetic mechanisms underlying adaptive radiation in different lineages . We uncover two parallel adaptive radiations in the genus Bartonella . Our genome-wide analysis revealed a remarkable evolutionary parallelism in the horizontally acquired VirB T4SS in the two radiating lineages . This parallelism is characterized at the molecular level by the lineage-specific chromosomal integration of the virB loci and the independent origination of versatile sets of effector proteins for the interaction with host cells . Providing an arsenal of host-subverting functions that can be efficiently modulated , the VirB T4SS thus seems to represent an evolutionary key innovation triggering the independent radiations of the two lineages . Our study provides detailed insights into the molecular mechanisms underlying parallel adaptive radiations in a bacterial pathogen . Furthermore , many of the diversified T4SS effector proteins carry a FIC domain recently shown to mediate ‘AMPylation’ , a lately recognized post-translational modification [25] , [26] . FIC domains are highly conserved in evolution and the diversified variants of the Bartonella effector proteins may display a suitable model to study their activity spectrum in the future .
To study the adaptive evolution of Bartonella on a genomic level , we aimed for a set of genome sequences from species adapted to distinct mammalian reservoir hosts . To this end , we included in our analysis the published genome sequences of Bartonella bacilliformis ( Bb ) , Bartonella grahamii ( Bg ) , Bartonella henselae ( Bh ) , Bartonella quintana ( Bq ) , and Bartonella tribocorum ( Bt ) [16] , [27] , [28] . These five species are adapted to human ( Bb and Bq ) , cat ( Bh ) , mouse ( Bg ) , and rat ( Bt ) . Further , we sequenced the complete genome of Bartonella clarridgeiae ( Bc ) and generated draft sequences of Bartonella schoenbuchensis ( Bs ) , Bartonella rochalimae ( Br ) , Bartonella sp . AR 15-3 ( BAR15 ) , and Bartonella sp . 1-1C ( B1-1C ) . Bs was selected as representative of a solely ruminant-infecting clade [16] . Bc , Br , BAR15 , and B1-1C were previously shown to be closely related [29]–[31] . However , they were isolated from different mammalian reservoir hosts and therefore display a suitable set of species to study adaptive processes on the genomic level . BAR15 and B1-1C were recently isolated from American red squirrel and rat , respectively [30] , [31] , whereas Br was predominantly recovered from canidae like dogs or foxes , and Bc from cats [32]–[34] . Genome sequencing by 454-pyrosequencing resulted in an average sequence coverage of >35x . The single chromosome of the completely assembled genome of Bc was found to be 1 , 522 , 743 bp in size and thus belongs to the smaller genomes of Bartonella ( Table 1 ) . The draft genomes of Br , BAR15 , B1-1C , and Bs consist of 13 to 19 contigs with total genome sizes similar to the one of Bc . On average , 99% of all 454-sequencing reads were assembled into the analyzed 13 to 19 contigs indicating that our draft genomes did not miss essential sequence data for subsequent analysis . Genomic features of the strains used in this study are summarized in Table 1 . We inferred a robust species tree of the genus Bartonella based on 478 core genome genes of the ten available Bartonella genomes sequenced ( Figure S1 ) . To exclude that recombination or horizontal gene transfer within this set of core genome genes was affecting our phylogenetic analysis , we reconstructed single gene trees of the entire data set and performed a recombination analysis using the GARD algorithm [35] . 471 of the 478 genes revealed the same overall topology as our genome-wide phylogeny with the two monophyletic clades of lineage 3 and lineage 4 ( see Figure 1 ) . Further , the GARD analysis detected significant recombination breakpoints with a p-value<0 . 01 in only two out of the 478 core genome genes . Together these analyses show that our genomic data set is suitable for inferring a consistent species tree . Based on available sequence information for the housekeeping genes rpoB , gltA , ribC , and groEL , we included most other Bartonella species in the analysis resulting in a so-called supertree phylogeny ( Figure 1 ) [36] . Just as the analysis based on the 478 core genome genes of the ten sequences species alone , this supertree revealed four major clades in the monophyletic bartonellae: ancestral lineage 1 represented by the highly virulent human pathogen Bb [37]; lineage 2 comprising of Bs and three other ruminant-infecting species; lineage 3 consisting of the closely related Bc , Br , BAR15 , and B1-1C; and the most species-rich lineage 4 with 13 species including Bg , Bh , Bq , and Bt ( Figure 1 ) . A phylogeny based on only the four housekeeping genes resulted in the same clustering of these taxa into the four different Bartonella lineages ( Figure S1 ) . In contrast to the ancestral lineage 1 , lineages 2 , 3 , and 4 are ramifying to different degrees comprising species isolated from various hosts . While the species of lineage 2 are limited to infect ruminants and have overlapping host range [38] , the diversification of lineage 3 and 4 seems to result from the specific adaptation to distinct mammalian hosts [12] . To substantiate the ecological divergence within these two lineages , we analyzed the genotype-host correlation of Bartonella isolates sampled from diverse mammals . Based on gltA sequences , this analysis revealed clustering of strains isolated from same or similar hosts in lineage 3 and 4 ( Figure S2 ) . Further support for the host specific adaptation of different Bartonella species comes from recently published laboratory infections [39] , [40] and from our own rat infection experiments with the strains of lineage 3 ( Figure S3 ) . It is to mention that some Bartonella species can incidentally be transmitted to other hosts like humans [12] . These so-called zoonotic Bartonella species do not cause intraerythrocytic bacteremia in the accidental human host reflecting the lack of specific adaptation . However , such accidental transmissions might facilitate the emergence of new specificity resulting in host switches and the origination of new species . In particular , several species of lineage 3 and 4 are known to display such zoonotic pathogens , whereas for lineage 1 or 2 to our knowledge no such case has been reported so far [12] , [29] . In summary , our genome-wide phylogenetic analysis shows that the sister lineages 3 and 4 have evolved by adaptive radiations into same or similar ecological niches ( i . e . hosts ) . Long internal branches separating the two lineages from each other and preceding the radiations are evidence for their independent occurrence ( Figure 1 ) . Due to the lack of calibration time points , the exact timing of these independent radiations cannot be deduced . However , the phylogenetic tree in Figure 1 might suggest that lineage 3 diversified more recently compared to lineage 4 . This is supported by the mean p-distances inferred for the sequenced taxa of these two lineages: lineage 3 = 0 . 07±0 . 0002 , lineage 4 = 0 . 12±0 . 0003 ( see also Table S1 ) . Two alternative explanations for the observed differences in lineage diversification could be ( i ) a sampling bias , i . e . the full diversity of lineage 3 was not captured or ( ii ) smaller population sizes for species of lineage 4 over lineage 3 leading to faster evolution at purifying sites . Significant differences in population size might be rather unlikely as Bartonella species are thought to share a common life style in their respective reservoir host . Whether sampling of Bartonella species in animal populations was exhaustive enough is difficult to assess . However , a newly discovered species would only change the coalescent point of a lineage if it would hold a more ancestral position than the already known species of this lineage . In contrast to these alternative hypothesis , epidemiological studies rather seem to support the scenario of a more ancient onset of the radiation in lineage 4: ( i ) lineage 4 comprises a much wider range of divergently adapted species and ( ii ) they represent the most frequently found Bartonella species in natural host populations [30] , [41] . In contrast , except for Bc , taxa of lineage 3 were only recently detected and so far only sampled at low prevalence [29]–[32] . The evolutionary parallelism of the radiating Bartonella lineages provides an ideal setting to study independent evolutionary processes linked to the adaptation to divergent niches . An important driving force for these adaptive radiations might be the presence of ecological opportunities . The niche of Bartonella in the mammalian reservoir host , the bloodstream , displays a privileged environment in which other resource-competing microbes are typically absent . Thus , the adoption of the characteristic intra-erythrocytic infection strategy together with the vector-borne transmission route might have enabled adaptive radiations of Bartonella by the specialization to different hosts . However , not all Bartonella lineages appear to have diversified to the same extent ( see Figure 1 ) suggesting that the availability of such an ecological opportunity alone is not sufficient to explain the pronounced radiation of lineages 3 and 4 . Supposedly , key innovations , i . e . lineage-specific traits underlying the adaptation to the mammalian host niches are responsible for the adaptive radiations . Potential adaptive traits would have to be involved in species-environment interactions , such as molecular factors responsible for causing bacteremia . Further , in analogy to the modulation of the adaptive traits in metazoan radiations [2] , [3] , any molecular factor used to exploit distinct environments in a specific manner should be divergent among niche-specialized species [1] . Molecular evolutionary analyses provide the means to identify divergent adaptive traits as the genes encoding them are expected to show signs of adaptive evolution , i . e . an excess of non-synonymous ( dn ) over synonymous ( ds ) substitutions as the result of positive selection . We performed a genome-wide natural selection analysis in the two radiating lineages to detect genes ( and therefore traits ) with divergent evolution . To this end , we analyzed all orthologous genes from the available genomes of the radiating lineage 3 ( Bc , Br , BAR15 , and B1-1C ) and lineage 4 ( Bh , Bg , Bq , and Bt ) for signs of adaptive sequence evolution by inferring the natural selection of orthologs by estimation of ω , the ratio of non-synonymous ( dn , amino acid change ) to synonymous ( ds , amino acid conservation ) substitution rates ( ω = dn/ds ) . Generally , ω<1 , ω = 1 , ω>1 represent purifying , neutral , and positive selection ( adaptive evolution ) , respectively [42] . We first calculated “gene-wide” dn/ds for all orthologous genes of the two lineages ( lineage 3: 1 , 097 genes , lineage 4: 1 , 091 genes ) . We excluded one gene from this analysis for each of the two lineages , because the GARD analysis detected statistically significant recombination breakpoints . As adaptive evolution is typically affecting only a few sites of a gene rather than the entire gene sequence [42] , we first looked for genes exhibiting an elevated value of ω≥0 . 25 over the entire gene length . This analysis revealed 133 ( 12% ) and 86 ( 8% ) genes in lineage 3 and lineage 4 , respectively , under relaxed purifying selection indicating signs of adaptive evolution ( Figure 2 ) . To have an additional measurement for adaptive evolution , we subjected our genomic data sets to a maximum likelihood analysis for the detection of site-specific positive selection . To this end , we used the CodeML module implemented in the PAML package . CodeML compares the likelihoods of different evolutionary models for each analyzed gene alignment . When comparing model M2a ( PositiveSelection ) vs . model M1a ( NearlyNeutral ) , we detected 62 or 34 genes for lineage 3 and 26 or 14 genes for lineage 4 harboring sites under positive selection with a p-value of <0 . 05 or <0 . 01 , respectively . A large fraction of these genes ( 29 for lineage 3 and 12 for lineage 4 ) exhibited also gene-wide dn/ds values ≥0 . 25 indicating them as good candidates for encoding adaptive traits ( Table 2 ) . Comprehensive lists of the genes identified to have dn/ds values ≥0 . 25 and/or exhibiting site-specific positive selection in the CodeML analysis are provided in Table S2 and Table S3 for lineage 3 and lineage 4 , respectively . As non-synonymous mutations accumulate over time , the higher number of genes identified for lineage 3 could be a further indication for its more recent radiation , or alternatively , the effect of larger population sizes compared to lineage 4 . Irrespectively , these findings render the dataset derived from lineage 3 to be more sensitive to the detection of adaptive sequence evolution ( Figure 2 , Table 2 ) . Interestingly , there was a marked number of genes with ω≥0 . 25 over the entire gene length that overlapped in both lineages ( Table 3 ) . Among those were genes encoding autotransporters , hemin-binding proteins , and different components of the VirB T4SS . They all constitute important host colonization factors [43]–[45] and thus are likely to display adaptive traits of Bartonella . For many of these genes , also our analysis of site-specific natural selection detected positive selection ( Table 3 ) . Most remarkably , all analyzed Bartonella effector protein ( bep ) genes of the VirB T4SS were among the genes with ω≥0 . 25 . Particularly in lineage 3 , they showed strong signs of adaptive evolution by exhibiting ω>0 . 4 over the entire gene length . Further , in eight out of nine analyzed bep genes of lineage 3 , we detected site-specific positive selection ( Table 3 ) . Being exclusively found in the radiating lineages and showing strong signs of adaptive evolution , the VirB system and its effector proteins , thus , fulfill the criteria of an evolutionary key innovation likely contributing to the parallel adaptive radiations of Bartonella . Autotransporters and hemin-binding proteins could represent further adaptive traits important for radiations in Bartonella . They exhibited strong positive selection in our analysis and are known to be important factors for host colonization . Further , their conservation throughout the genus Bartonella indicates an important role for the life style and infection strategy of this pathogen ( Table S2 , Table S3 ) . However , factors conserved in radiating and non-radiating lineages appear unlikely to represent specific key innovations , unless other factors , such as the absence of ecological opportunities or ecological separation , prevented certain lineages to radiate and to colonize more divergent niches . Importantly , our analysis revealed some lineage-specific colonization factors to carry signs of adaptive evolution . Among others , surface-exposed pilus-components of the Trw T4SS exclusively present in lineage 4 were found to exhibit elevated dn/ds values and sites under positive selection ( Table S3 ) . This is in agreement with previous studies and appears to reflect the adaptation of this putative adhesion factor to the erythrocytic surface of different host species [22] , [24] . As the Trw T4SS is only present in lineage 4 , it might have specifically contributed to the radiation of this most species-rich clade of Bartonella . Factors known to be important for colonization and exclusively present in lineage 3 were not identified by our analysis . We cannot exclude that the selective pressure imposed by the immune-system might have contributed to the adaptive evolution detected in our genome-wide analysis . It was previously reported that the arms race between host and pathogen can drive the diversification of secretion system- and effector protein-encoding genes [46] , [47] . However , in case of the Trw T4SS , recently published in vitro infections with erythrocytes isolated from different mammals demonstrated that the Trw-dependent binding and invasion of Bartonella is host-specific [22] . Although experimental data is not yet available , our data suggest that the VirB T4SS and its effector proteins evolved by similar mechanisms . Together with the previous finding that the VirB T4SSs belong to the few colonization factors specific to the radiating lineages [16] , this analysis reveals these horizontally acquired host interacting systems as potential key innovations facilitating adaptation to new hosts and therefore driving the radiations of Bartonella . To further assess the role of the VirB T4SS for the independent adaptive radiations of Bartonella , we compared the chromosomal organization of the VirB and effector protein-encoding genes of the two lineages . Remarkably , our analysis uncovered independent evolutionary scenarios for the chromosomal incorporation of this horizontally acquired trait . In the genomes of lineage 4 ( Bg , Bh , Bq , and Bt ) , the virB T4SS genes , virB2-virB11 and the coupling protein gene virD4 , are encoded at the same chromosomal location ( Figure 3 , Figure S4 ) . Also , the bep genes are encoded in this region . In contrast , the genome sequences of lineage 3 ( Bc , Br , BAR15 , and B1-1C ) revealed marked differences in organization , copy number , and chromosomal localization of the genes encoding the VirB T4SS . In the completely assembled genome of Bc , we found three copies of the virB2-virB10 genes encoded at two different chromosomal locations ( Figure 3 , Figure S4 ) . Two copies are encoded at the same locus and belong to inverted repeats of ∼10kb . They are separated by several bep genes and the gene virD4 . A third copy of the virB2-virB10 cluster including an additional bep gene is encoded in another genomic region highly conserved across different Bartonella lineages ( Figure 3 ) . The same chromosomal integration and amplification of the virB T4SS genes was found in the other three genomes of lineage 3 . However , in a common ancestor of Br and B1-1C , one of the three copies must have been partially deleted , as only virB2 , virB3 , and a remnant of the virB4 gene were found in the corresponding region of these two genomes ( Figure 3 , Figure S4 ) . Interestingly , the different copies of virB2-virB10 are identical to each other within one species , but divergent across different species indicating the presence of an intra-chromosomal homogenization process . The fact that duplicated components of another T4SS , Trw , also evolved in concert , and the finding of several other identical genes or gene clusters in different Bartonella genomes [24] suggests that sequence homogenization is a common mechanism in Bartonella to conserve paralogous gene copies . The inverted organization of the two virB T4SS gene clusters seems to result from a duplication event subsequent to the integration of a first copy . Evidence comes from a remnant of the glutamine synthetase I gene ( glnA ) flanking the entire locus at its upstream end . The full-length copy of this vertically inherited housekeeping gene is located directly downstream of the integration site ( Figure 3 , Figure S4 ) . In addition to the effector genes adjacently located to the virB genes ( as in lineage 4 ) , we found six additional loci encoding bep genes in lineage 3 ( Figure S4 ) . These effector genes are not entirely conserved throughout lineage 3 , and the existence of gene remnants provides evidence of their deterioration in certain species . Altogether , we identified 12 to 16 bep genes in lineage 3 , whereas only five to seven bep genes are present in lineage 4 . Incomplete synteny in the corresponding regions may hinder comparison between the two different lineages , however , no gene remnants could be found at the different integration sites across the two lineages . We cannot fully exclude that massive genomic recombination events resulted in the different chromosomal locations and the lineage-specific dissemination of the virB and bep genes . Yet , such a scenario appears unlikely , as the overall genomic backbone is largely conserved ( Figure 3 ) and the flanking regions of the virB T4SS integration sites do not encode vertically-inherited orthologs across the two lineages . Furthermore , the absence of mobile elements adjacent to the virB T4SS genes such as recombinases , transposases , or integrases is not supportive of an intra-chromosomal mobilization of this genomic locus . T4SS are ancestrally related to conjugation machineries [48] . Thus , the virB genes might have been transferred from a conjugative plasmid into the chromosome by independent events after the divergence of lineage 3 and lineage 4 . In Bg , a closely related T4SS , the Vbh , is encoded on a plasmid in addition to a chromosomally integrated copy [28] . This indicates that these horizontally acquired elements can be maintained on extra-chromosomal replicons within Bartonella from where they are integrated into the chromosome . Similarly , pathogenic Escherichia coli strains from different phylogenetic clades were shown to have evolved in parallel by the independent incorporation of virulence traits from mobile genetic elements [49] . As the chromosomal organization implicates different evolutionary histories of the VirB T4SS in the two radiating lineages , we investigated the relation among the effector proteins translocated by this secretion system . It was previously shown that the bep genes have evolved from a single ancestor by duplication , diversification , and reshuffling of domains resulting in modular gene architectures [17] . The C-terminal BID ( Bartonella intracellular delivery ) domain is shared by all Beps as it constitutes the secretion signal for the transport via the VirB T4SS . In their N-terminal part , Beps either harbor a FIC ( filamentation-induced by cAMP ) domain or repeats of additional BID domains ( Figure S5 ) . We assessed the evolutionary relationship among the bep genes by inferring phylogenetic trees on the basis of either the BID or the FIC domain , or the entire gene sequence . This revealed that the bep genes of lineage 3 and lineage 4 form two separate clades ( Figure 4 , Figure S6 ) . Apparently , consecutive rounds of lineage-specific duplications of an ancestral effector gene resulted in the parallel emergence of two distinct arsenals of bep genes . These duplication events preceded the adaptive radiation in both lineages as phylogenetic clusters of effector genes ( Bep clades in Figure 4 ) comprise positional orthologs present in all or a subset of the analyzed genomes of the corresponding lineage ( Figure S4 ) . Gene duplications frequently display the primary adaptive response after the acquisition of beneficial factors , because they occur at much higher frequency than other adaptive mutations [50] . This might have been the initial selective pressure for the independent amplification processes . However , in both lineages , the duplicated bep genes subsequently diversified by accumulating mutations as indicated by different branch lengths separating bep genes in Figure 4 . To analyze the sequence evolution during the parallel amplification and diversification processes , we used a branch test for positive selection . Since positive selection is not continuously acting during evolution , this analysis allows the detection of episodic adaptive evolution on single phylogenetic branches . We detected positive selection on many of the internal branches suggesting that subsequent to their duplication different Bep clades have undergone adaptive sequence evolution in both lineages ( Figure 4 , Figure S6 ) . For Bartonella , experimental studies showed that effector proteins exhibit distinct phenotypic properties on host cells indicating that the evolutionary diversification of the duplicated effectors was substantially driven by the acquisition of novel functions [18]–[21] . Not all branches exhibit dn/ds values >1 , though , suggesting episodic changes in the selection pressure acting on different effector gene copies . For example , functional redundancy of paralogous effector copies could have resulted in neutral drift , whereas conservation of an advantageous function might have led to purifying selection on certain branches . The basis for the functional versatility seems to lie in the adaptability of the domains encoded by bep genes . In case of the FIC domain , recently published work showed that this domain mediates a new post-translational modification by transferring an AMP moiety onto a target protein [25] . Proteins ‘AMPylated’ by FIC domains belong to the family of GTPases . The diversity of these targets and their numerous functions in cellular processes might allow the diversified Beps to target and subvert a variety of host cell functions by target-specific ‘AMPylation’ . The high degree of conservation of the FIC domain in different kingdoms of life provides further evidence for the remarkable versatility of this domain [51] . Interestingly , also the BID domain , constituting part of the translocation signal of the effector proteins , seems to be capable of adopting various functions in the host cell [17] . In case of BepA , it was shown that the BID domain is sufficient to mediate the anti-apoptotic property of this effector protein [18] . BID domains of other Beps with the same domain constitution as BepA do not exhibit this phenotype indicating specific adaptive modulation of this domain for BepA . This functional adaptability might also explain why certain effector genes carry more than one BID domain . At last , tandem-repeated tyrosine-phosphorylation motifs found in a subset of effector proteins confer another multifaceted molecular mechanism to modulate cellular processes . Phosphorylated effector proteins are thought to recruit cellular binding partners resulting in the formation of signaling scaffolds that interfere with specific host cell signaling pathways [21] . For several effector proteins of Bh ( lineage 4 ) , tyrosine phosphorylation by host cells has been reported and the targeted host interaction partners studied [17] , [21] . Beside Bartonella , a number of other pathogens , as E . coli ( EPEC ) , Helicobacter pylori , or Chlamydia trachomatis are using tyrosine-phosphorylation of effector proteins to modulate their hosts in very distinct ways demonstrating the versatility of this type of host subversion [21] . In Bartonella , the tyrosine-phosphorylated effector proteins seem to display an important functionality of the VirB-mediated host modulation as we found effector proteins of this type in both radiating lineages ( see below ) . Our analyses suggest that the domain structure of the ancestral effector gene consisted of an N-terminal FIC and a C-terminal BID domain ( FIC-BID ) . In both lineages , the FIC-BID domain structure displays the most abundant effector protein type . In lineage 3 , only effector genes of Bep clade 9 consist of domain architectures different than FIC-BID ( Figure S5 ) . The gene tree in Figure 4 shows that bep genes with the shortest evolutionary distance across the two lineages are the ones harboring the FIC-BID structure ( BepA clade and Bep clade 1 ) . bep genes with different domain architecture constitute more distantly related clades across the two lineages indicating that they derived by independent recombination from the ancestral domain structure . Furthermore , the distantly related Vbh T4SS of Bg and Bs encodes an effector protein consisting of the FIC-BID domain structure [28] . As mentioned above , it was shown for lineage 4 that some of the derived bep genes become phosphorylated by host cell kinases at conserved tandem-repeated tyrosine-phosphorylation motifs leading to the interference with specific host cell pathways [21] . Strikingly , we found that bep genes with derived domain architecture in lineage 3 also harbour regions with tandem-repeated tyrosine motifs ( Figure 5 , Figure S5 ) . In silico predictions of tyrosine-phosphorylation sites with three different programs [52]–[54] consistently revealed a high number of potentially phosphorylated motifs within these repetitive regions ( Figure 5 , Table S4 ) . We ectopically expressed these effector proteins in HEK293T cells and showed that they are indeed phosphorylated within eukaryotic cells by tyrosine kinases implicating their functional importance for host interaction ( Figure 5 ) . Interestingly , the motifs found in lineage 3 are clearly different from the ones present in lineage 4 and are also less conserved as depicted by their consensus sequences in Figure 5 . This suggests that the motifs in lineage 3 may generally be under weaker purifying selection than in lineage 4 , because they target either less conserved or even different pathways in their hosts . Further , the lower degree of conservation and the higher number of motifs per effector found in lineage 3 could also indicate that these proteins and particularly their motifs are under positive selection and evolved more recently than their equivalents in lineage 4 . Together with the fact that tandem-repeated phosphorylation motifs are only found in bep genes with derived domain architecture , our findings , thus , suggest parallel evolution of this class of effector proteins within the two radiating lineages . Whether similar pathways are targeted by these effectors in the two lineages remains unknown . Yet , the striking parallelism in the molecular evolution of this class of effector proteins indicates their central role in the VirB T4SS mediated host modulation by Bartonella . Emerging infectious diseases are frequently caused by zoonotic pathogens which are incidentally transmitted to humans from their reservoir niche ( e . g . other animal hosts ) . Therefore , the understanding of the mechanisms driving diversification of host-adapted bacteria in nature is of relevance for human health . In the present study , we explored the adaptive diversification of host-restricted bartonellae . Our genome-wide phylogeny revealed that two sister clades of this α-proteobacterial pathogen have evolved by parallel adaptive radiations ( lineage 3 and lineage 4 in Figure 1 ) . Both lineages comprise species adapted to same or similar reservoir hosts including zoonotic ( e . g . Bh and Br ) or human specific ( Bq ) pathogens . The more recent diversification of lineage 3 including the recently recognized incidental human pathogen Br [29] underlines the importance to study the molecular basis of such lineage diversifications . In line with the ‘ecological’ parallelism of their radiations , our comparative genomic analyses between lineage 3 and lineage 4 uncover striking evolutionary parallelisms at the molecular level of a likely key innovation - the VirB T4SS – essentially involved in the infection of the mammalian hosts . Chromosomal fixation of this horizontally transferred trait occurred by independent evolutionary events . In both lineages , the arsenal of effector proteins translocated via the VirB T4SS was shaped independently by gene duplications and positive selection of diversified gene copies . This amplification process mostly occurred before the onset of the radiations . Strikingly , beside the diversification of effector proteins encoding the evolutionary conserved ‘AMPylase’ domain ( FIC ) , both lineages have convergently evolved a novel effector class with derived domain structure and tandem-repeated tyrosine-phosphorylation motifs . By these evolutionary processes , large reservoirs of distinct biological functions were invented from a single ancestral effector gene . This functional versatility provides the framework for the adaptive potential of the VirB T4SS . Apparently , the plasticity of the underlying genomic loci seems to have favored the parallel occurrence of these adaptive processes in two distinct lineages , thereby essentially contributing to the parallel radiations of Bartonella .
Animals were handled in strict accordance with good animal practice as defined by the relevant European ( European standards of welfare for animals in research ) , national ( Information and guidelines for animal experiments and alternative methods , Federal Veterinary Office of Switzerland ) and/or local animal welfare bodies . Animal work was approved by the Veterinary Office of the Canton Basel City on June 2003 ( licence no . 1741 ) . Bc strain 73 [34] , B1-1C [31] , and Br strain ATCC BAA-1498 [29] were grown routinely for 3–5 days on tryptic soy agar containing 5% defibrinated sheep-blood in a water-saturated atmosphere with 5% CO2 at 35°C . BAR15 [30] and Bs strain R1 [55] were grown under the same conditions on heart infusion agar and Colombia base agar , respectively . Using the QIAGEN Genomic DNA Isolation kit ( Qiagen ) , DNA was isolated from bacteria grown from single colonies . For 454-sequencing , the DNA was prepared with an appropriate kit supplied by Roche Applied Science and sequenced on a Roche GS-FLX [56] . To assemble the reads , Newbler standard running parameters with ace file output were used . Newbler assemblies were considerably improved by linking overlapping contigs on the basis of the “_to” and “_from” information appended to the read name in the ace files . For the assemblies of Bc , BAR15 , B1-1C , Br , and Bs , we obtained a 454-sequence coverage of 35x , 37x , 39x , 39x , and 29x , respectively ( for details on 454-sequencing see Table S5 ) . Repeats were identified by analyzing the coverage of each Newbler contig . If the link between two contigs was ambiguous , PCR and long-range PCR were used to confirm contig joins . For the complete assembly of the Bc genome , a library of 35 kb inserts was generated using the CopyControl Fosmid Kit ( Epicentre ) . By end-sequencing of library clones with Sanger technology , 983 high-quality reads were obtained and mapped onto the 454-sequencing-based assembly . Remaining sequence gaps were closed by PCR . The final singular contig was fully covered by staggered fosmid clones indicating a correct assembly of the circular chromosome of Bc . Gene predictions of the genome of Bc and the draft genomes of Bs , Br , BAR15 , and B1-1 were performed using AMIGene software [57] . Automated functional gene annotation was conducted with the genome annotation system MaGe [58] . For orthologous genes , the annotation was adopted from the manually annotated genome of Bt [16] . Manual validation of the annotation was performed for the virB and bep genes . By using the “FusionFission” tool of MaGe [58] fragmented genes were identified and the corresponding sequences subsequently examined for 454-sequencing errors . After correcting these errors , the updated sequences were re-annotated as described above . The sequence data of the genome of Bc and the contigs of the draft genomes of Br , BAR15 , B1-1C , and Bs is stored on the web-based interface MaGe ( Bartonella2Scope , https://www . genoscope . cns . fr/agc/mage/bartonella2Scope ) and has been deposited in the EMBL Nucleotide Sequence Database under accession numbers FN645454–FN645524 . Phylogenetic trees were based on nucleotide sequence data . Alignments were generated on protein sequences with ClustalW [59] and back-translated into aligned DNA sequences using MEGA4 [60] . Tree topologies were calculated with maximum likelihood and Bayesian inference methods as implemented in the programs PAUP* [61] and MrBayes [62] , respectively . The genome-wide phylogeny of Bartonella was calculated on the basis of 478 orthologous genes of the ten sequenced Bartonella genomes and the genome of Brucella abortus ( bv . 1 str . 9-941 ) . Orthologs were determined by using the “PhyloProfile Synteny” tool of MaGe [58] with a threshold of 60% protein identity over at least 80% of the length of proteins being directional best hits of each other . The alignments of the 478 identified genes were concatenated resulting in a total of 515 , 751 aligned nucleotide sites . Tree topology and branch lengths were obtained by maximum likelihood analysis using the HKY85 model . Bootstrap support values were calculated for 100 replicates . For Bayesian inference , the program MrBayes [62] was run for one million iterations with standard parameters ( two runs with four heated Monte-Carlo Markov chains in parallel; number of substitutions = 6; burnin = 25% ) . For the Bartonella ingroup , single gene trees were calculated with maximum likelihood and tree topology congruency assessed with PAUP* . 471 of the 478 single gene trees revealed the same monophyletic clustering of the eight taxa into lineage 3 and lineage 4 as the genome-wide phylogeny . Further , we performed a recombination analysis for each of the 478 single gene alignments using the GARD algorithm as implemented in the HYPHY package [35] . The GARD analysis was run with the GTR model using a general discrete distribution with three rate classes . To identify statistical significant recombination breakpoints in our alignments , we used the Kishino-Hasegawa test as implemented in the GARDProcess . bf algorithm of the HYPHY package . To include non-sequenced Bartonella species in the genome-wide phylogeny , we used available sequence data from the gltA , groEL , ribC , and rpoB genes ( 7731 aligned sites ) . Trees were obtained as described above . MrBayes [62] was run for five million iterations . Branch lengths for tip branches of non-sequenced taxa are calculated on the basis of the four housekeeping genes . Branch lengths for tip branches of sequenced taxa and internal branches separating sequenced and non-sequenced taxa are based on the genomic data set . The maximum likelihood tree only based on the gltA , groEL , ribC , and rpoB genes was inferred as described for the genome-wide phylogeny . Bep gene trees were inferred from nucleotide alignments of either the most C-terminal BID domain including the C-terminus ( 948 sites ) , the FIC domain including the N-terminal extension ( 1 , 305 sites ) , or the entire bep sequence of genes harboring FIC domains ( 3 , 972 sites ) . To select an appropriate substitution model , the Akaike information criterion of Modeltest 3 . 7 [63] and MrModeltest 2 . 0 [64] was used for the maximum likelihood and Bayesian inference analysis , respectively . For the alignments based on the BID domain or the entire bep gene sequence , we obtained the GTR+G+I model with both programs . For the alignments based on the FIC domain , the TVM+I+G model ( Modeltest 3 . 7 ) and GTR+G+I model ( MrModeltest 2 . 0 ) were selected . Trees were inferred with the parameters provided by these models as described above . MrBayes [62] was run for one million iterations . The Neighbor-joining phylogeny of different Bartonella isolates in Figure S2 was inferred from a 242 nt segment of the gltA gene with the program MEGA4 [60] . Bootstrap values were calculated for 1 , 000 replicates . Based on the four available genomes , orthologous genes for each of the two lineages 3 and 4 were determined by using the “PhyloProfile Synteny” tool of MaGe [58] . The threshold was set to 30% protein identity over at least 60% of the length of proteins being directional best hits of each other . The same tool was used to detect genes without orthologs . By comparing these automatically identified orthologs and non-orthologs , genes present in neither of the two lists were detected and manually assigned to one of the two lists . Alignments were generated and a GARD recombination analysis conducted as described above . To obtain the average dn/ds value ( ω ) of each ortholog , the arithmetic mean of pair-wise dn/ds values ( calculated by the method of Yang and Nielsen implemented in PAML 4 . 1 [65] ) was used . Site tests of positive selection were performed with PAML 4 . 1 using the CodeML module [65] . To detect positive selection model M1a ( NearlyNeutral ) vs . model M2a ( PositiveSelection ) and model M7 ( beta ) vs . model M8 ( beta+ω ) were analyzed . PAUP* [61] was used to infer maximum likelihood trees for each set of orthologs . For the CodeML control file , standard parameters were used . The relative significance of model M2a ( PositiveSelection ) vs . model M1a ( NearlyNeutral ) and model M8 ( beta+ω ) vs . model M7 ( beta ) was assessed using likelihood-ratio-tests ( two degrees of freedom ) . Genes for which significant positive selection was detected were inspected for alignment errors potentially affecting the results of this analysis . If necessary , the alignments were manually modified and the CodeML analysis repeated . Phylogenetic branches were tested for positive selection by using the TestBranchDNDS . bf module implemented as standard analysis tool in HyPhy [66] . Ten weeks old female WISTAR rats obtained from RCC-Füllinsdorf were housed in an BSL2-animal facility for two weeks prior to infection allowing acclimatization . For inoculation , bacterial strains were grown as described above , harvested in phosphate-buffered saline ( PBS ) , and diluted to OD595 = 1 . Rats were anesthetized with a 2–3% Isuflurane/O2 mixture and infected with 10 µl of the bacterial suspension in the dermis of the right ear . Blood samples were taken at the tail vein and immediately mixed with PBS containing 3 . 8% sodium-citrate to avoid coagulation . After freezing to −70°C and subsequent thawing , undiluted and diluted blood samples were plated on tryptic soy agar and heart infusion agar containing 5% defibrinated sheep-blood . CFUs were counted after 8–12 days of growth . Nucleotide distances were calculated with the program MEGA4 [60] for the alignments based on the genome-wide dataset and the four housekeeping genes . The numbers of base substitutions per site from averaging over all sequence pairs within and between groups were calculated . Codon positions included were 1st , 2nd , and 3rd . All positions containing gaps and missing data were eliminated from the dataset ( Complete deletion option ) . To construct the plasmids pPE2002 and pPE2004 , bep genes BARCL_1034 ( Bc ) and BARRO_80017 ( Br ) were amplified from genomic DNA with primer pairs containing flanking BamHI/NotI sites: prPE453 ( ATAAGAATGCGGCCGCGATGAAAAC-CCATAACACTCCTG ) /prPE454 ( CGGG-ATCCTTAATGTGTTATAACCATCGTTC ) and prPE455 ( ATAAGAATGCGGCCGCG-ATGAATTTTGGAGAAAAGAAAAAAATG ) /prPE456 ( CGGGATCCTTAAATAGC-TACAGCTAACGATTTTTTC ) , respectively . PCR products were digested with the enzymes BamHI and NotI and ligated into the BamHI/NotI sites of the backbone of plasmid pAP013 ( kindly provided by Arto Pulliainen ) . The resulting constructs pPE2001 ( BARCL_1034 ) and pPE2003 ( BARRO_80017 ) were cut with NotI and ligated with a GFP fragment obtained from NotI digested pAP013 . The plasmid pPE2007 was constructed by cutting bepE of B . henselae from plasmid pRO1100 ( kindly provided by Rusudan Okujava ) with NotI and BamHI and ligating it into pAP013 . All plasmid DNA isolations and PCR purifications were performed with Macherey-Nagel and Promega columns according to manufacturer's instructions . The protocol for growth and transfection of HEK293T was performed as described previously [18] . 36 h after transfection , cells were incubated for 10 minutes with 10 ml Pervanadate medium ( 5 ml PBS containing 100 mM orthovanadate and 200 mM H2O2 , incubated for 10 min with 500 µl Catalase [2 mg/ml in PBS] before 45 ml M199 medium were added ) . After washing three times with 7 ml of PBS at room temperature , cells were scraped off and resuspended in 1 ml of ice-cold PBS containing 1 mM EDTA , 0 . 5 mM phenylmethylsulfonyl fluoride ( PMSF ) , 1 mM orthovanadate , 1 mM leupeptin , and 1 mM pepstatin and collected by centrifugation ( 3 , 000g at 4°C for 60 sec ) . The resulting pellet was lysed in 300 µl of ice cold modified RIPA buffer ( 50 mM Tris-HCl [pH 7 . 4] , 75 mM NaCl , 1 mM EDTA , 1 mM orthovanadate , 1 mM leupeptin , 1 mM pepstatin ) for 1 hour at 4°C . The lysate was centrifuged ( 16 , 000g at 4°C for 15 min ) and 12 µl of anti-HA-agarose ( Sigma ) added to the supernatant . After 150 min of incubation at 4°C on a slowly turning rotation shaker , the agarose was washed three times with 300 µl of modified RIPA buffer ( 3 , 000g for 10 sec ) . The affinity-gel pellet was then resuspended in 20 µl of modified RIPA buffer , 20 µl of SDS-sample buffer ( 2× ) were added , and the sample was heated for 5 min at 95°C . Proteins were separated on a 10% SDS-polyacrylamide gel , blotted on a nitrocellulose membrane ( Hybond-C Extra , Amersham Pharmacia ) , and examined for tyrosine phosphorylation by using monoclonal antibody 4G10 ( Millipore ) and anti-mouse IgG-horseradish peroxidase ( HRP ) afterwards . The HRP-conjugated antibody was visualized by enhanced chemiluminescence ( PerkinElmer ) . For visualization of the signal from GFP-fusion proteins , the membrane was subsequently incubated in 4% PBS-Tween containing 0 . 02% NaN3 and anti-GFP antibody ( Invitrogen ) , followed by incubation with anti-mouse IgG-HRP and visualization by enhanced chemiluminescence . | Adaptive radiation is the rapid origination of an array of species by the divergent colonization of disparate ecological niches . In the case of pathogenic bacteria , radiations can lead to the emergence of novel human pathogens . Being divergently adapted to a range of different mammalian hosts , including humans as reservoir or incidental hosts , the genus Bartonella represents a suitable model to study genomic mechanisms underpinning divergent adaptation of pathogens . Here we show that two distinct lineages of Bartonella have radiated in parallel , resulting in two arrays of evolutionary distinct species adapted to overlapping sets of mammalian hosts . Such parallelisms display excellent models to reveal insights into the genetic mechanisms underlying these independent evolutionary processes . Our genome-wide analysis identifies a striking evolutionary parallelism in a horizontally-acquired protein secretion system in the two lineages . The parallel evolutionary trajectory of this system in the two lineages is characterized by the convergent origination of a wide array of adaptive functions dedicated to the cellular interaction within the mammalian hosts . The parallel evolution of the two radiating lineages on the ecological as well as on the molecular level suggests that the horizontal acquisition and the functional diversification of the secretion system display an evolutionary key innovation underlying adaptive evolution . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/genomics",
"genetics",
"and",
"genomics/microbial",
"evolution",
"and",
"genomics",
"evolutionary",
"biology/microbial",
"evolution",
"and",
"genomics",
"infectious",
"diseases",
"genetics",
"and",
"genomics/comparative",
"genomics",
"evolutionary... | 2011 | Parallel Evolution of a Type IV Secretion System in Radiating Lineages of the Host-Restricted Bacterial Pathogen Bartonella |
Recent genome-wide ( GW ) scans have identified several independent loci affecting human stature , but their contribution through the different skeletal components of height is still poorly understood . We carried out a genome-wide scan in 12 , 611 participants , followed by replication in an additional 7 , 187 individuals , and identified 17 genomic regions with GW-significant association with height . Of these , two are entirely novel ( rs11809207 in CATSPER4 , combined P-value = 6 . 1×10−8 and rs910316 in TMED10 , P-value = 1 . 4×10−7 ) and two had previously been described with weak statistical support ( rs10472828 in NPR3 , P-value = 3×10−7 and rs849141 in JAZF1 , P-value = 3 . 2×10−11 ) . One locus ( rs1182188 at GNA12 ) identifies the first height eQTL . We also assessed the contribution of height loci to the upper- ( trunk ) and lower-body ( hip axis and femur ) skeletal components of height . We find evidence for several loci associated with trunk length ( including rs6570507 in GPR126 , P-value = 4×10−5 and rs6817306 in LCORL , P-value = 4×10−4 ) , hip axis length ( including rs6830062 at LCORL , P-value = 4 . 8×10−4 and rs4911494 at UQCC , P-value = 1 . 9×10−4 ) , and femur length ( including rs710841 at PRKG2 , P-value = 2 . 4×10−5 and rs10946808 at HIST1H1D , P-value = 6 . 4×10−6 ) . Finally , we used conditional analyses to explore a possible differential contribution of the height loci to these different skeletal size measurements . In addition to validating four novel loci controlling adult stature , our study represents the first effort to assess the contribution of genetic loci to three skeletal components of height . Further statistical tests in larger numbers of individuals will be required to verify if the height loci affect height preferentially through these subcomponents of height .
Body height is determined by several biological processes that occur throughout the life of an individual and involve both normal and pathological growth . Most skeletal bones are formed by endochondral ossification , the process of replacing hyaline cartilage with bony tissue . Ossification starts shortly after gestation at the diaphysis and extends along the end of the long bones . Growth continues throughout childhood via mitotic division of the cartilage at the distal surface of the epiphysis and the epiphyseal plate , accelerates during the adolescence growth spurt and slows down in the early twenties , when the epiphyseal plate completely ossifies and peak body height is achieved . Subsequent decreases in peak body height throughout life are mainly the consequence of vertebral bone deformities such as crush fractures ( osteoporosis ) and cartilage degeneration ( osteoarthritis ) . Epidemiological studies have revealed marked differences in the growth patterns for the lower and upper portions of the body . Both trunk and lower limb length are associated with parental height , birth weight and weight at age 4 years [1] . Leg length , the principal determinant of final height attainment in adults [2] , is positively associated with advantaged socio-economical circumstances and nutritional intake in childhood [1] , [3] , [4] . Leg length is also largely responsible for the secular increase in height in some populations [5] . Conversely , trunk length is not correlated with nutrient intake in children , and has been shown to be negatively correlated with psychophysical stress [1] . This evidence suggests that independent growth pathways might be in part responsible for final upper and lower body size . Recent , well-powered genome-wide association ( GWA ) scans have already identified 43 independent loci associated with height , revealing a significant overrepresentation of genes controlling DNA replication , intracellular signalling , cell division and mesoderm and skeletal development [6]–[10] . We have carried out an independent meta-analysis of stature to assess the contribution of genetic loci to overall body height . In addition , we have investigated for the first time the contribution of these height loci to leg and trunk length using derived measurements of skeletal frame size and we have used conditional analyses to explore a possible differential contribution of the height loci to the different skeletal measurements .
To search for loci influencing adult height we analyzed genome-wide ( GW ) data for 299 , 216 SNPs from a combined sample of 12 , 611 adults of Caucasian origin from the UK ( EPIC Norfolk , n = 3 , 211; TwinsUK , n = 2 , 224; 1958 Birth Cohort , n = 1 , 430 ) and the Netherlands ( Rotterdam Study , n = 5 , 746 ) genotyped with Human300 and 550 bead arrays ( Illumina ) ( Figure S1 ) . We note that the EPIC Norfolk sample was also included in a recent meta-analysis for height , where the same individuals were genotyped using the 500 K Gene chip ( Affymetrix ) [9] . Although our discovery set was not entirely independent from the Weedon et al study , inclusion of this collection in the discovery set can provide further support to validate other genomic regions in conjunction with other sample collections . In total , 86 SNPs from 35 genomic regions reached significance with P-values<10−5 in the full set of 12 , 611 individuals; in the independent discovery set of 9 , 400 individuals ( excluding EPIC , due to its previous use ) 57 SNPs from 20 genomic regions were significant at this threshold . We selected 52 SNPs from the 35 regions for replication genotyping in an additional 4 , 275 samples from three population cohorts ( TwinsUK , Chingford and Chuvasha ) . In addition , we genotyped a subset of 31 SNPs highly suggestive of true association but that did not reach GW-significance in the first replication set in additional 2 , 912 individuals with height data from the CBR cohort ( Table 1; details of the cohorts are given in Table S1 ) . A total of seventeen independent signals reached GW-significance ( P-value≤5×10−7 ) in the combined dataset , and seven additional loci had suggestive evidence for association with height ( 5×10−7≤P-value<10−5 , Table 1 and Table S2 ) . Two of the seventeen loci that reached GW-significant association are novel . The SNP rs11809207 is located in the third intron of the CATSPER4 gene ( combined P-value = 6 . 1×10−8 ) . The A allele at rs11809207 was associated with an increase in height of 0 . 071 ( 95% C . I . 0 . 044–0 . 098 ) standard deviations , corresponding to an effect size of 0 . 46 cm per copy of the allele . Two nearby SNPs are in high LD with rs11809207 and have marginally higher P-values in the meta-analysis ( rs2783711 near PDIK1L , P-value = 5 . 7×10−5 , r2 = 0 . 61 in the HapMap CEU sample; and rs12069719 , P-value = 3 . 7×10−5 , r2 = 0 . 6; Figure S2A ) . The lead SNP of the second locus rs910316 ( P-value = 1 . 4×10−7 ) is located in the first intron of TMED10 ( Figure S2O ) . The A allele is associated with an increase in height of 0 . 053 ( 95% C . I . 0 . 031–0 . 075 ) standard deviations in the combined sample , corresponding to 0 . 34 cm per allele copy . Two additional loci , NPR3 and JAZF1 , which were previously reported as weakly associated to height ( P-values>10−5 [6] ) reached GW-significance in our study ( rs10472828 in NPR3 , P-value = 3×10−7 and rs849141 in JAZF1 , P-value = 3 . 2×10−11 , Table 1 ) . The remaining thirteen loci reaching GW-significance ( Table 1 ) had been described in one or more of the recent GWA scans for stature , providing strong evidence for independent and widespread replication [6]–[10] . The Weedon et al ( 2008 ) study shared approximately 3 , 200 samples with our discovery cohort ( EPIC cohort ) . Six of the eight loci discovered in both studies reached GW-significance in our replication set when EPIC was excluded from the analysis ( Table 1 ) , indicating independent replication of these loci in the remaining sample collections . The strongest association signals in our combined set were observed at HMGA2 ( rs8756 , P-value 5×10−14 ) and UQCC ( rs6088813 , P-value 9 . 8×10−14 ) both of which have been replicated in multiple studies [6]–[10] . The seventeen GW-significant loci explain 0 . 07%–0 . 18% of total height variance in our sample ( Table 1 ) . Finally , rs1812175 located in HHIP had a nominal P-value in the replication sample but did not reach GW-significance in the combined analysis ( Table 1 ) . A further six loci which did not attain nominal significance in the replication sample have suggestive evidence for association with height in the combined sample ( Table 1 ) . To assess heterogeneity in the height associations among cohorts , we compared regression coefficients for height normalised to z-scores using the Cochran's and I2 statistics , finding little or no evidence for heterogeneity ( Table 1 ) . We used a similar approach to investigate gender-specific effects in height associations . We focused on the Rotterdam study , which is the largest cohort with similar numbers of males and females . We compared normalised height z-scores calculated in 3 , 374 females and 2 , 362 males from the Rotterdam Study using Cochran's and I2 statistics ( Table S3 ) . We observed limited evidence for gender-specific effects . The exception was ADAMTS33 SNPs , where we detected significant heterogeneity in height associations at all three SNPs investigated ( P-value = 0 . 002 , I2 = 89%; Table S3 ) . We tested the association of the 17 GW-significant loci with three different skeletal size measurements , namely spine length , femur and hip axis length , which provide proxies for trunk , leg and skeletal size length respectively . We first investigated skeletal size measurements representing proxies for trunk length . We analysed 6 , 053 samples from three cohorts with available measurements of spine length ( TwinsUK and Chuvasha ) and vertebral body heights ( Rotterdam Study ) . We combined study-specific summary statistics using z-scores and found that nine of the 17 loci were significantly associated with trunk length at the nominal level . The strongest associations with spine were at rs6570507 in GPR126 ( P-value = 4×10−5 ) , rs6817306 in LCORL ( P-value = 4×10−4 ) , rs849141 in JAZF1 ( P-value = 0 . 001 ) and rs10472828 in NPR3 ( P-value = 0 . 0018 ) ( Table 2 ) . We next tested association of the 17 confirmed height loci with hip axis length ( HAL ) in 2 , 341 individuals from the Rotterdam Study ( Table 3 ) . HAL is a highly-heritable measure of femoral geometry that measures the distance from the lateral aspect of the greater trochanter to the inner border of the pelvic rim , passing through the mid-section of the femoral neck . HAL is strongly correlated with total frame size and height [11] and represents a clinically important predictor of hip fracture independent of age and femoral neck bone mineral density [12] . Of the 17 validated height loci , seven had one or more SNPs significantly associated with HAL in the Rotterdam Study , with the strongest statistical associations observed at LCORL ( rs6830062; P-value = 4 . 8×10−4 ) and UQCC ( rs4911494; P-value = 1 . 9×10−4 ) . Finally , we investigated associations of the 17 validated height loci with measurements of lower limb length ( femur ) in 3 , 505 individuals from two cohorts ( TwinsUK , N = 2 , 364 and Chuvasha , N = 1 , 141 ) . The strongest associations with femur length were observed at rs710841 ( PRKG2 , P-value = 2 . 4×10−5 ) and rs10946808 ( HIST1H1D , P-value = 6 . 4×10−6 ) ( Table S4 ) . We used the following qualitative approach to explore a possible differential contribution of height loci to skeletal size measurements . We selected a homogeneous set of measurements ( vertebral heights and HAL in the Rotterdam Study , and femur length in TwinsUK ) to avoid bias deriving from heterogeneous measurements among cohorts . Although this is expected to reduce the power to detect statistically significant associations , the magnitude of the betas is unlikely to be materially affected . Secondly , we re-calculated associations of the 17 GW-significant height loci with each skeletal size measurement as described before , only in this case we restricted the analysis to the homogeneous set of measurements . We then performed an analysis where association with each skeletal size measurement was assessed after adding height as an additional term to the linear regression model . We finally compared qualitatively the magnitude of the association , expressed as regression coefficients and SE , in the two models ( univariate and height-adjusted ) . In cases where the locus acts prevalently through the given skeletal size measurement , we expect the regression coefficients of the height-adjusted analysis to show the least reduction compared to the original unadjusted analysis . For loci showing associations in the height-adjusted analysis , we then recalculated associations with height after adding skeletal size in the linear regression model . A null regression coefficient in this case suggests that the height association may be explained by the prevalent effect of the locus on the skeletal measurement under exam . Collectively , these conditional analyses , with relevant caveats [13] , can provide some indicative information as to whether the association between relevant genetic variants and height may be mediated by specific components of height . Observed association signals in intergenic regions may be due to regulatory variants of nearby genes . For the 17 validated height loci ( Table 1 ) we undertook an eQTL analysis in 2 Mb windows centered on each lead SNP ( see Materials and Methods for details ) using expression data from lymphoblastoid cell lines derived from individuals of the four HapMap population panels [14] . The recombination interval harboring rs1182188 contains a total of 27 SNPs tested for association with both height and expression . We tested 14 genes within ±1 MB of this interval and found significant expression association evidence ( Spearman Rank correlation P-value<0 . 001 ) for 2 of them ( GNA12 and LOC392620 ) in at least one HapMap population . The rankings of the CEU expression and height association p-values at the 27 tested SNPs correlate well , suggesting a potential common functional variant underlying both the height and expression signal ( GNA12: rho = 0 . 678 , P-value = 0 . 0001; LOC392620; rho = 0 . 386 , P-value = 0 . 047 ) . Simulation of random significant eQTLs at this interval ( as significant as the observed ones , maintaining SNP frequencies and haplotype structure ) and comparison of the height-expression correlation strength to the observed data indicates that the probability of obtaining such a relationship between the two phenotypes by chance is very small ( P-value = 0 . 0228 ) . We detected one cis signal that reached significance in the eQTL analysis in a 340 Kb region of chromosome 7 centred on rs1182188 . The rs1182188 SNP is located in the first intron of GNA12 ( Figure 1 ) . GNA12 encodes for Galpha12 , a serine/threonine phosphatase modulating essential signalling pathways , including apoptosis [15] , [16] , regulation of the actin cytoskeleton and cadherin-mediated cell-cell adhesion [17] , [18] . Even though it is known that apoptosis and cadherin-mediated signalling are involved with oncogenic effects during tumor progression , we have no clear indication how they can contribute to the determination of body height .
We carried out a genome-wide scan for stature in 12 , 611 adults of Caucasian origin and identified seventeen independent regions having GW-significant associations at the 5×10−7 threshold [19] , and seven additional loci with suggestive evidence for association ( P-values between 5×10−7 and 10−5 ) . Two of the regions with GW-significant associations with stature are novel and were centred on an intronic SNP in CATSPER4 ( rs11809207 , combined P-value = 6 . 1×10−8 ) , and on the first intron of TMED10 ( rs910316 , P-value = 1 . 4×10−7 ) . CATSPER4 encodes for a receptor membrane ligand ion channel with a role in acrosome reaction and male fertility [20] , whereas TMED10 ( TMP21 ) encodes for a presenilin complex component with a role in vesicular protein trafficking [21] . Both these functions do not have an obvious association with height , which may be mediated by variants in nearby genes . For instance , the lead SNP in TMED10 is in high linkage disequilibrium with three SNPs in NEK9 displaying marginally lower association P-values in the genome-wide scan . NEK9 is a regulator of cellular processes essential for interphase progression [22] , a biological function shown to be overrepresented among height association signals [6] . We also identified two additional loci showing genome-wide statistical association with height , namely rs10472828 in NPR3 ( P-value = 3×10−7 ) and rs849141 in JAZF1 ( P-value = 3 . 2×10−11 ) both of which were identified with low confidence ( P-value>10−5 ) in one recent GWA scan also using Illumina genome chips [6] . NPR3 encodes for natriuretic peptide ( NCP ) , a protein class that elicit a number of vascular , renal , and endocrine effects that are important in the maintenance of blood pressure and extracellular fluid volume [23] . JAZF1 is a transcriptional repressor associated with a role in endometrial stromal tumors [24] . Interestingly , common variants in JAZF1 were also recently implicated in type 2 diabetes and prostate cancer susceptibility [25] , [26] , in line with evidence for pleiotropic effects at many common disease loci . The remaining thirteen loci were previously described [6]–[10] , confirming highly reproducible height associations despite overall small effect sizes for individual loci . Our study confirmed HMGA2 ( rs8756 , P-value 5×10−14 ) and UQCC ( rs6088813 , P-value 9 . 8×10−14 ) as the loci with the strongest overall association with height [6]–[10] . In this study we also attempted , for the first time , to study the contribution of height loci to the components of trunk and leg length using measurements of skeletal size by directly testing associations of the 17 validated height regions with skeletal size parameters of trunk , leg and skeletal frame size , and by conducting an exploratory conditional analysis . Some intrinsic limitations of our study design and analytical approaches need to be taken into consideration in the interpretation of results for skeletal size measurements . Firstly , although we assembled the largest dataset of this kind , the small sample size available for skeletal size associations will have affected our ability to conclusively confirm or rule out associations for some loci . However , the composite nature of height as a phenotype , and an overall higher accuracy and specificity of the skeletal measurements in assessing the contribution of genetic loci to limb length , suggests that smaller sample sizes might be sufficient to detect associations for these intermediate traits with high confidence . For example , the seventeen GW-significant height loci jointly explain approximately 5% of total variance in femur length , which is more than twice the variance in height explained by the same loci combined . In order to increase power , we combined trunk size measurements from different cohorts that , despite being highly correlated , were not identical . Radiographic measurements of vertebral height size ( Rotterdam Study ) provide a more accurate measure of the skeletal component of trunk height compared to the DXA-derived measurements of the total spine ( TwinsUK and Chuvasha ) , which include inter-vertebral disk heights and potential measurement error due to vertebral crush fractures . Measuring skeletal associations requires methods that are relatively expensive and low-throughput compared to height , making it difficult to assemble large homogeneous samples for analysis . To provide more comparable estimates of association , we focused our analyses on a subset of homogeneous standardised measurements of vertebral size and HAL in the largest available cohort , the Rotterdam Study . To help assess whether components of skeletal height mediated the association between relevant genetic loci and height , and to examine the independency of associations between genetic loci and components of height , we conducted an exploratory conditional analysis . In this and related contexts , conditional analysis to infer conditional independence and mediation has limitations . Conditional analysis of highly correlated traits ( for example , height and skeletal subcomponents of height ) can lead to an attenuation of effect sizes for relevant genetic loci even when there is no underlying causal network between genetic variants and mediating traits [13] . These analyses also assume that measurement error is evenly distributed among traits . Differences in measurement error among traits can result in spurious inferences - distorting the magnitude of relevant effect sizes in conditional and unconditional analyses . The latter is also relevant to conditional analyses of genetic variants . Moreover , because of the statistical resolution required to assess differences in unconditional and conditional analyses , and the correlated variance structures of these data ( testing differences in effect estimates using the same sample population ) we opted to use a qualitative assessment of these interrelated associations rather than a quantitative one . With the above caveats in mind , our results provide some interesting first insights into the potential contribution of height loci and possible differential effects on skeletal size measurements . These results should be considered exploratory and will require replication in larger cohorts to better understand their role and address potential sources of heterogeneity including the impact of measurement error . For instance , in a previous study Weedon and colleagues [10] described an association for HMGA2 ( rs1042725 ) with sitting height ( 0 . 2 cm increase for the C allele , 95% C . I . 0 . 1–0 . 3 , P-value = 0 . 0002 ) in a cohort of approximately 2 , 000 children . Although this may suggest a differential effect on trunk length , in our study such association was not replicated in an adult sample of 6 , 509 individuals for a highly correlated SNP in the same locus ( rs8756 , r2 = 0 . 87 in CEU , combined P-value = 0 . 12 ) . Yet , direct spine length measures in adults are likely to be more precise and specific than measures of sitting height since they are less affected by sources of measurement error that may explain such discordant results . For instance , sitting height is prone to measurement errors affected by head dimensions which are disproportionate in children ( having already achieved adult head dimensions around 3 years of age , long before the pubertal growth spurt ) . In addition , radiographic measures of spine length will also be less affected by other artefacts arising from posture differences , age-specific growth patterns and/or possible common age-related effects of inter-vertebral disk degeneration , all of which can play a key role in this discrepancy . The clinical relevance of the effects observed for these height loci is interesting and merits further exploration . Several loci displayed significant association with HAL , a measurement shown to vary between ethnic groups and to have substantial heritability . In addition HAL is highly correlated with long limb growth and represents an important predictor for osteoporotic fracture [12] . In our study , two intronic variants in the recombination interval containing GDF5 and UQCC ( rs4911494 and rs6088813 ) were strongly associated with HAL in the Rotterdam study ( P-values = 1×10−4 and 1 . 32×10−4 respectively ) , but not with femur length ( P-values = 0 . 83 and 0 . 76 respectively ) . GDF5 is a member of the TGF-beta superfamily of growth factors/signalling molecules that act as regulators of cell growth and differentiation in both embryonic and adult tissues . Mutations in this gene are associated with severe skeletal malformations including acromesomelic dysplasia , Hunter-Thompson type , brachydactyly , type C and chondrodysplasia Grebe type [27]–[29] . A common functional SNP in the 5′ UTR of GDF5 ( +104T/C; rs143383 ) has been associated with osteoarthritis ( OA ) , the commonest form of human arthritis characterized by degeneration of articular cartilage and bone remodelling [30] , [31] . OA is also under strong genetic influence , with several shared genetic risk factors with skeletal traits including bone density , bone content , turnover and skeletal alignment [32] . The two low-stature alleles rs4911494-A and rs6088813-A were in high linkage disequilibrium with the risk T allele at rs143383 ( r2 = 0 . 93 ) , indicating a possible role of this gene in cartilage metabolism and/or bone shape and alignment in determining height . In summary , this study extended by four the list of loci with confirmed association to adult height , which now comprises 47 independent regions . The use of fine mapping through genotype imputation and resequencing will be important for refining the association signal in each locus and for identifying the true causative variants . The potential differential effects that we observed of height loci to lower limb and trunk growth are consistent with some genes potentially acting as regulators of long-bone growth , while others appear to be specific to different bone sites or to influence cartilage growth . Further analytical and experimental approaches to assess the contribution of height loci to skeletal measurements and intermediate phenotypes will be important to understand the physiology of human growth , and may lead to the identification of genetic variants relevant to diverse musculoskeletal pathologies in humans . Note: While this manuscript was in review the JAZF1 locus was confirmed by an independent study [33] .
The initial discovery sample included 9 , 400 samples of European origin , including 1 , 430 British individuals ( 710 females and 720 males ) from the British 1958 Birth Cohort , 2 , 224 individuals from the TwinsUK cohort ( all females ) and 5 , 746 individuals ( 3 , 374 females and 2 , 372 males ) from a Dutch cohort ( Rotterdam Study ) . After applying quality filters , 299 , 216 SNPs remained for analysis with data in at least 9 , 000 individuals . Further details of individual cohorts are given below and in Table S1 . In addition to these three cohorts we also had available GWAS data for 3 , 211 samples from the EPIC Norfolk study , genotyped using the Illumina HumanHap300 ( v1 ) SNP panel . These individuals were genotyped using the Affymetrix 500 K SNP panel in a recent height meta-analysis [9] , and therefore do not constitute an independent discovery sample . Nevertheless , as the samples may provide novel loci once combined with different cohorts and a different platform , we included them in a second stage of discovery . Furthermore , these samples provide independent replication for published height signals , except than for those described in the Weedon scan [9] ( Table 1 ) . For this reason in Table 1 we provided association statistics both including and excluding the EPIC collection for all loci already described by Weedon and colleagues . Genotyping of SNPs selected for replication was carried out using mass spectrometry ( Sequenom iPLEX ) at the Wellcome Trust Sanger Institute following standard procedures . Details of genotyping assays are available upon request from the authors . TwinsUK Cohort: http://www . twinsuk . ac . uk/ Chingford Cohort: http://www . chingfordstudy . org . uk/ 1958 Birth Cohort: http://www . b58cgene . sgul . ac . uk/ Rotterdam Study: http://www . epib . nl/ergo . htm R statistical package: http://www . r-project . org METAL: http://www . sph . umich . edu/csg/abecasis/Metal/ | The first genetic association studies of adult height have confirmed a role of many common variants in influencing human height , but to date , the genetic basis of differences between different skeletal components of height have not been addressed . Here , we take advantage of recent technical and methodological advances to examine the role of common genetic variants on both height and skeletal components of height . By examining nearly 20 , 000 individuals from the UK and the Netherlands , we provide statistically significant evidence that 17 genomic regions are associated with height , including four novel regions . We also examine , for the first time , the association of these 17 regions with skeletal size measurements of spine , femur , and hip axis length , a measurement of hip geometry known to influence the risk of osteoporotic fractures . We find that some height loci are also associated with these skeletal components , although further statistical tests will be required to verify if these genetic variants act differentially on the individual skeletal measurements . The knowledge generated by this and other studies will not only inform the genetics of human quantitative variation , but will also lead to the potential discovery of many medically important polymorphisms . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/complex",
"traits",
"rheumatology/bone",
"and",
"mineral",
"metabolism",
"genetics",
"and",
"genomics",
"rheumatology/cartilage",
"biology",
"and",
"osteoarthritis"
] | 2009 | Meta-Analysis of Genome-Wide Scans for Human Adult Stature Identifies Novel Loci and Associations with Measures of Skeletal Frame Size |
Ecological niche models are useful tools to infer potential spatial and temporal distributions in vector species and to measure epidemiological risk for infectious diseases such as the Leishmaniases . The ecological niche of 28 North and Central American sand fly species , including those with epidemiological relevance , can be used to analyze the vector's ecology and its association with transmission risk , and plan integrated regional vector surveillance and control programs . In this study , we model the environmental requirements of the principal North and Central American phlebotomine species and analyze three niche characteristics over future climate change scenarios: i ) potential change in niche breadth , ii ) direction and magnitude of niche centroid shifts , iii ) shifts in elevation range . Niche identity between confirmed or incriminated Leishmania vector sand flies in Mexico , and human cases were analyzed . Niche models were constructed using sand fly occurrence datapoints from Canada , USA , Mexico , Guatemala and Belize . Nine non-correlated bioclimatic and four topographic data layers were used as niche components using GARP in OpenModeller . Both B2 and A2 climate change scenarios were used with two general circulation models for each scenario ( CSIRO and HadCM3 ) , for 2020 , 2050 and 2080 . There was an increase in niche breadth to 2080 in both scenarios for all species with the exception of Lutzomyia vexator . The principal direction of niche centroid displacement was to the northwest ( 64% ) , while the elevation range decreased greatest for tropical , and least for broad-range species . Lutzomyia cruciata is the only epidemiologically important species with high niche identity with that of Leishmania spp . in Mexico . Continued landscape modification in future climate change will provide an increased opportunity for the geographic expansion of NCA sand flys' ENM and human exposure to vectors of Leishmaniases .
Leishmaniases are an increasingly important disease group worldwide , based on case numbers , geographic expansion , socioeconomic implications , psychological impact , and immunosuppression due to HIV infection which re-activates Leishmania spp . patency [1] . There are four main clinical manifestations of Leishmaniases: localized cutaneous leishmaniases ( LCL ) , diffuse cutaneous leishmaniases , mucocutaneous leishmaniases and visceral leishmaniases ( VL ) ; the first ( LCL ) and last ( VL ) cause the greatest disease burden and mortality , respectively , for the disease group [2] . Leishmania spp . are transmitted by female sand flies of the genus Lutzomyia in the New World [2] , [3] . In North and Central America ( NCA ) included in the North American tectonic plate which extends to southern Guatemala , 62 species of sand flies have been recorded [3]–[5] , of which nine have been confirmed or incriminated as vectors of Leishmania [6]–[9] . The sand flies Lutzomyia longipalpis and Lutzomyia evansi are confirmed primary VL vectors in several countries [2] , [10] , whereas Lutzomyia olmeca olmeca is a confirmed vector of LCL in Mexico [6] . Other sand fly species , such as Lutzomyia anthophora [8] , Lutzomyia cruciata [9] , [11] , Lutzomyia diabolica [7] , Lutzomyia ovallesi [12] , Lutzomyia panamensis [9] , Lutzomyia shannoni [7] , [9] , and Lutzomyia ylephiletor [9] , [13] , however , have been found naturally infected or experimentally infected with Leishmania spp . [11] . Four species of Leishmania have been isolated in NCA , and are responsible for all human and canine clinical manifestations: Leishmania mexicana , Leishmania braziliensis , Leishmania panamensis and Leishmania infantum chagasi [2] , [6] , [14] . Clinical symptoms depend on the host species , its immune-competence , parasite species or strain , in addition to other as yet unidentified genetic determinants [15] . Despite early studies on the taxonomy and geographic distribution of sand flies in NCA , knowledge regarding the biology , distribution , and ecology of new collections and species continues to be registered from only a few regions [16] . Knowledge of current and potential sand fly distributions are important to predict the impact of environmental modification , the expansion of human settlements and migration , and climate change ( CC ) or its variation on parasite and vector population dynamics . Hence , there is a need for alternative tools to analyze species' distributions and potential sand fly dispersal areas [17] , [18] . The choice and use of prevention strategies in risk areas for all Leishmaniases will depend on current and potential distributions of epidemiologically relevant species ( ERS ) [19] . Generally , macroclimatic variables influence species distributions at coarse scales , topographic variables at regional scales , and land use and biotic interactions at finer scales [20] . Whereas land use and biotic interactions are more related to demographic dynamics , macroclimatic variables determine the distributional limits while topography delineates physical barriers for dispersal [21] , [22] . Therefore , species' geographic range shifts are predicted in the forthcoming decades , as a result of the accelerated rate of climate change [23] which reduces niche suitability in current locations , while offering new suitable colonization sites [24] . The change of at least two important attributes of a species' ecological niche , the niche breadth ( the expressed geographic coverage of the abiotic niche related to the available geographic space ) and the niche's geographic centroid ( the geometric central point of the specie's geographic range which indicates the latitudinal mid-point of the range ) would dramatically affect the geographic epidemiology of Leishmaniases in North America ( i . e . the emergence of new regions where transmission cycles could be established due to the convergence of mammal hosts , parasite and vectors , and human population exposed to these vectors ) . Ecological niche modeling ( ENM ) has already been used to project the geographic distribution potential of epidemiologically relevant Old World sand fly species: Phlebotomus papatasi [25] , Phlebotomus orientalis , Phlebotomus martini [26] , and Phlebotomus alexandri [25] . ENM have also been generated for a few New World , such as Lutzomyia whitmani , Lutzomyia intermedia and Lutzomyia migonei [17] , and a few NCA species [16] , [18] . Multiple abiotic and biotic factors have been associated with NCA sand fly species' distributions , in particular precipitation , temperature , altitude , latitude , physical barriers , and host distributions and abundance [27] , [28] . While certain sand fly species exhibit local extinctions , others are predicted to adapt successfully and indeed to increase their relative abundance in modified habitats [29] . All of these factors also affect the spatial and temporal distribution of vectors and reservoirs , which in turn affect the epidemiology and dynamics of pathogen transmission to the human population [30] . Analysis of the impact of climate variability on Leishmaniases has focused principally on vector distributional changes due to El Niño [17] , [18] , [31] , or using climate simulations [32] . All tropical and temperate NCA species from Guatemala and Belize to Canada are modeled together in this study , and niche characteristics as well as epidemiological associations of relevant species are analyzed in two contrasting CC scenarios . We have focused on analyzing potential change in species' geographic ranges as predicted by macroclimatic changes at the coarse-grain level , since these provide greater model consistency and accuracy for climate circulation models and their bioclimatic variables [33] . No reliable data layers for future land use changes are available to be incorporated into the niche models , although we use the differential between climate change scenarios to predict the impact of local scale habitat changes .
The study area for model construction and projection includes Canada , USA , Mexico , Guatemala and Belize , limited by 14 . 07°N , 58 . 23°N and −136 . 15°W , −56 . 29°W . The region was divided into 7 , 536 , 074 pixels at a resolution of 30 arc-seconds ( 0 . 008333°≈1 km ) for latitude and longitude . Ecological region categories were assigned using the World Wildlife Fund ( WWF ) shape files based en Terrestrial Eco-regions of the World [34] . A database was constructed from collections reported in published scientific literature , entomological collections housed in several academic institutions in Mexico ( Universidad Autónoma de Yucatán ( UADY ) , El Colegio de la Frontera Sur ( ECOSUR ) , and Universidad Autónoma de Nuevo León ( UANL ) ) , the Instituto Nacional de Diagnóstico y Referencia Epidemiológica ( InDRE ) , and author's unpublished collections ( Table S1 ) . The database included 1 , 478 occurrence data points for 28 sand fly species with ≥10 records in the NCA region: Belize ( N = 230 ) , Canada ( N = 2 ) , USA ( N = 208 ) , Guatemala ( N = 42 ) and Mexico ( N = 996 ) . In order to analyze niche shift trends , all species were assigned to one of three ecological region categories: tropical ( moist and dry forest , n = 22; 1 , 306 data points ) , temperate ( desert , grasslands , steppe , savanna , prairies , mountains forest , scrubland , pine forest , conifer forest , swamps , mangroves and mezquital , n = 4; 103 data points ) and broad-range ( species in both regions; n = 2; 69 data points ) ( Table 1 ) . Thirteen environmental layers were used for the construction of ENM . Nine bioclimatic data layers ( annual mean temperature , temperature seasonality , maximum temperature of warmest month , minimum temperature of coldest month , temperature annual range , annual precipitation , precipitation of wettest month , precipitation of driest quarter and precipitation seasonality ) were obtained from the Worldclim- Global Climate Data ( www . worldclim . org; last accessed Nov , 2011 ) at a resolution of 30 arc-seconds [35] . These bioclimatic variables were selected from 19 by choosing the more meaningful variables hypothesized to limit species distribution at coarse-grain scale , after analysis of multicolinearity in a correlation matrix [18] . The final dataset layer includes variables with relatively low inter-correlation ( r<0 . 75 ) . Additionally , four topographic layers ( aspect , slope , topographic index and elevation ) obtained from the Hydro 1k data set ( Earth Resources Observations and Science- http://eros . usgs . gov/products/elevation/gtopo30/gtopo30 . html; last accessed Dec , 2011 ) were also used for ENM models . ENM based on occurrence data , bioclimatic and topographic layers were constructed using the Genetic Algorithm for Rule-set Prediction ( GARP ) and best subsets implementation [36] , [37] from the OpenModeller desktop ver . 1 . 1 . 0 [38] . In general , the procedure focuses on modeling the set of ecological conditions in which a species can maintain populations without immigration [39] . GARP is the preferred model for datasets which may have heterogeneous occurrence records across a broad geographic range . The software randomly divides occurrence points into training data for model building ( 75% ) and test data for model testing ( 25% ) . One hundred replicate models were developed for each species and a soft omission threshold of 20% of the distribution was used for all [37] . Each ENM was evaluated using two tests: accuracy , a measure of performance , and the AUC ( area under the receiver operating curve [ROC] ) , as a test of predictive ability . Both tests are based on two types of error: commission ( areas of actual absence predicted present ) and omission ( areas of actual presence predicted absent ) [37] . The internal ( training data ) and external ( test data ) accuracy was calculated using the confusion matrix , equivalent to “sensitivity” [a/ ( a+c ) ] . The AUC ( ROC curve ) was calculated using the values of “sensitivity” in the y-axis and the commission error in the x-axis , measuring the maximum inflection point where both errors are minimized . The AUC has a range of 0 . 0 to 1 . 0 ( in general , acceptable models have AUC>0 . 85 ) [18] . We used a minimum presence threshold criterion of 90% in order to generate a binary map ( presence/absence ) of each projection from the 0–100 range of the model output . To do this , we first selected a set of 90% of random records per species and projected them onto the model . Then , we selected a threshold that predicted the presence of all of the 90% datapoints and converted the values≥of that number in “1” ( presence ) and the values<of that threshold in “0” ( absence ) to get a binary map of distribution . The binary maps were tested on training and test datasets , using a binomial test which evaluates the success rate of correct classification of presence data in comparison with random expectation [40] . Since there is no active epidemiological surveillance for Leishmaniases in Mexico , we use an identity test to identify niche overlap of Leishmania spp . ( PEN ) and each vector [41] . ENM were generated for all incriminated vector species: Lu . anthophora , Lu . cruciata , Lu . diabolica , Lu . longipalpis , Lu . olmeca olmeca , Lu . ovallesi , Lu . panamensis , Lu . shannoni and Lu . ylephiletor . Human cases of Leishmaniases from multiple Mexican states , Campeche ( N = 8 ) , Chiapas ( N = 161 ) , Guerrero ( N = 10 ) , Morelos ( N = 2 ) , Oaxaca ( N = 3 ) , Puebla ( N = 4 ) , Quintana Roo ( N = 101 ) , Tabasco ( N = 15 ) and Veracruz ( N = 37 ) were used as proxy to generate the PEN . A maximum-entropy-based algorithm , MaxEnt [42] was used to generate all vectors and PEN ENM using topographic and bioclimatic variables previously mentioned , since this spatially explicit test and corresponding statistical analyses are not available for GARP . The parameters to measure identity were the random test percentage ( 75% ) , replicated run type ( bootstrap ) , maximum iterations ( 500 ) , and the threshold rule ( minimum training presence ) , using ENMtools ver . 13 . 2 ( http://enmtools . com/ , last accessed Mar , 2012; [43] ) . Two climate change scenarios were used: the A2 and B2 scenarios [44] . The A2 scenario assumes a rapid increase in human population , economy , technology , land use change , agriculture and energy consumption , while these parameters are more moderate in the B2 scenario . In the A2 scenario , there is an average of 3 . 4°C temperature increase for the year 2099 , while in the B2 scenario , this increase would not supercede 2 . 4°C [33] , [44] . Two general circulation models were used for both scenarios: CSIRO ( CSIRO Division of Marine and Atmospheric Research , Australia [45] ) and HadCM3 ( Hadley Center for Climate Prediction and Research , England [45] , [46] ) . Both models included four primary characteristics ( atmosphere , ocean , sea ice and land ) and feature a 1% increase to 2×CO2 at time of doubling . The CSIRO model uses an increase in 1 . 21°C , 2 . 05°C and 3 . 07°C for 2020 , 2050 , and 2080 , respectively . The HadCM3 model uses an increase of 1 . 21°C , 2 . 10°C and 3 . 17°C for the same years [44] . Generally , the CSIRO model has better performance at a global level [45] , while the HadCM3 model was chosen according to performance in reproducing regional climate for Mexico , Central America and the Caribbean [33] .
In both models and CC scenarios ( B2 and A2 ) , niche breadth increased over time for all but one species ( Figure 6 ) . In the CSIRO model Lu . vexator's ENM expanded over time as for all other species , while it contracted using the HadCM3 model . The geographic projection of this species' ENM diminished in the HadCM3 model , in northern regions , as well as fragments in the south . Average niche breadth increase was marginally greater in temperate and broad-range as compared with tropical species . Most tropical species had a greater increase in B2 than A2 , while the opposite was observed for temperate species ( Table S3 ) . Species with the greatest ENM increase over time were Lu . bispinosa , Lu . cruciata , Lu , ylephiletor , Lu . diabolica , and Lu . texana , most of which had greatest change in the A2 scenario . Temperate as compared to tropical ERS species have greater breadth increase , specifically in the A2 scenario . Despite variable expansion of ENM in geographic space , the change in niche overlap between current and 2050 was minimal: 93 . 5% in A2 and 98 . 6% in B2 for tropical species , and 95 . 6% for A2 and 97 . 2% for B2 for temperate species ( Table S2 ) . Changes in ENM overlap among sand fly species ( 11 . 41–80 . 20% ) was variable according to distribution category and time period . Overlap between current and 2050 projections was lowest for temperate species ( average 43 . 49% ) and highest for tropical ( 67 . 44% ) species ( Table 3 ) . The ERS have an intermediate average 55 . 3% overlap , as compared with 67 . 7% for non-incriminated vector species . In the A2 scenario , Lu . cruciata had the lowest overlap with other species ( 11 . 4% ) , while Lutzomyia permira ( 80 . 1% ) and Lutzomyia undulata ( 80 . 2% ) had the highest overlap with other species . There is a direction shift in ENM centroids for all species in all time periods and both CC scenarios; the majority of species shifted to the northwest ( 64 . 3% ) , followed by northeast ( 35 . 1% ) , and minimally to the southwest ( 0 . 6% ) . The direction shift was to the northeast for Lu . longipalpis and Lu . panamensis , and to the northwest for all other ERS . The distance shift of ENM centroids was variable ( 47–940 km ) according to species and time periods ( Table 3 ) . As expected , the shift was greater in the A2 than in the B2 scenario . In general , centroid shifts were greatest for temperate species , followed by tropical and broad-range categories . Lutzomyia ovallesi ( tropical ) had the greatest centroid shift , followed by Lutzomyia bispinosa ( tropical ) and Lu . anthophora ( temperate ) . In contrast , Lu . shannoni ( tropical ) has the lowest centroid shift of all 28 species . Although the shift in elevation range is highly variable , the average range for all 28 species decreased in future CC scenarios . In general , the decrease in the A2 scenario was greater than in the B2 ( Table 3 ) . The elevation patterns of ERS did not change substantially , although tropical species such as Lu . cruciata and Lu . longipalpis shift to lower elevations . The elevation range of all broad-range species , as well as Lu . anthophora ( temperate ) and Lutzomyia carpenteri ( tropical ) increased . The combined changes in niche breadth , elevation and centroid range and direction was analyzed focusing on 2050; the CSIRO model was run using the B2 scenario ( Figure 7 ) , and the HadCM3 model with the A2 scenario ( Figure S26 ) . The pattern of niche breadth and centroid shift was similar between scenarios , although elevation range shifts were differentially affected in the combined analysis . The average maximum elevation was higher in the A2 as compared with the B2 scenario ( 1 , 147 , 1 , 380 m , and 1 , 411 m for tropical , temperate and broad-range species , respectively ) . Epidemiologically relevant sand flies had similar patterns in both scenarios . Based on these patterns in the B2 scenario , four categories of ERS were defined based on average elevation range shift: the first group includes only Lu . olmeca olmeca ( average elevation range = 930 m ) , the second group includes Lu . ovallesi , Lu . panamensis , and Lu . ylephiletor ( 1 , 230 m ) , the third group contains only Lu . longipalpis ( 1 , 255 m ) , and the fourth group is composed of Lu . anthophora , Lu . cruciata , Lu . diabolica , and Lu . shannoni ( 1 , 599 m ) . In the A2 scenario , the average shift for the four groups was 1 , 058 m , 1 , 156 m , 1 , 319 m , and 1 , 590 m , respectively ( Table 3 ) . The complete projected vector-exposed Mexican population , based on ENMs of ERS , was calculated separately for urban and rural communities . Lutzomyia diabolica's niche covers the greatest total human population ( 107 , 176 , 279 inhabitants ) , followed by Lu . shannoni ( 71 , 002 , 449 ) , Lu . cruciata ( 57 , 966 , 560 ) , Lu . longipalpis ( 42 , 563 , 408 ) , Lu . ylephiletor ( 32 , 403 , 860 ) , Lu . ovallesi ( 30 , 792 , 955 ) , Lu . anthophora ( 24 , 230 , 744 ) , and Lu . olmeca olmeca ( 24 , 174 , 255 ) . The rural population ( communities <10 , 000 inhabitants ) exposed to sand flies will increase over time in both CC scenarios ( Table 4 ) . Lutzomyia diabolica's niche overlapped with the largest rural human population , while Lu . anthophora's contained the least . Tropical sand flies Lu . shannoni and Lu . cruciata have ENM in areas with the highest proportion ( 20 . 7% and 21 . 0% , respectively ) of exposed population , while Lu . olmeca olmeca overlaps with the lowest ( 8 . 9% ) . The only sand fly with significant niche identity with Leishmania spp . was Lu . cruciata ( Figure 8 and Table 5 ) .
The present study models the niche and potential natural distribution of the most abundant 28 NCA sand fly species , and projects these in two CC scenarios using atmospheric , ocean surface , sea ice land surface and elevation characteristics , and climate models appropriate for the region [33] . Although there are approximately 500 phlebotomine sand fly species described in the Americas , NCA has the lowest sand fly diversity with only 62 species reported to date . We have focused this study on the 28 species which fulfilled minimum abundance collection registries for modeling confidence [3] . Even though the ecology of reservoir hosts and certain abiotic variables [16] , [29] have been associated with sand fly distributions , their association with ecological niche modeled in a broad geographic area has not been analyzed . In addition , the differential impact of current versus accelerated environmental modifications ( local anthropic change ) in future CC scenarios has been explored for the geographic projection of the breadth , centroid location and elevation range shifts of these ENM . Highest sand fly species diversity in NCA , based on ENM , occurs in the sub-tropical region of southern Mexico , Belize and Guatemala , including Florida . The present model confirms highest sand fly species' richness in areas where other taxa ( mammal , reptile , bird , and amphibian ) diversity is highest [49] , [50] , which may be due to the fact that major land-use change is far more advanced in temperate as compared to sub-tropical regions [51] . Greatest vector ENM shifts are projected to occur where historical environmental modifications have occurred in temperate areas ( higher longitudes and lower elevations , [51] ) , and are projected to have greatest impact to 2020 , 2050 , and 2080 , in the extreme A2 scenario in these same areas . More accelerated environmental modification coincides with broader human exposure to these vectors as observed both for overall sand fly distribution , and specifically for ERS , by comparing between CC scenarios [52] . It is interesting to note that the HadCM3 model , currently considered one of the more appropriate to model climate for Mexico [33] , projects greater geographic expansion for all sand fly species , in comparison with CSIRO . The geographic projection of niche breadth increase is uniform surrounding most NCA sand fly ENMs over time in both CC scenarios . This uniform increase depends on specific landscape components , biotic interactions , habitat modification , or other characteristics affecting population growth or the species' fitness , all of which affect the realized niche [32] , [48] , [53] . Temperate and broad-range sand fly species' niche are projected to increase more than that of tropical species , an expansion which may reflect an increase of generalist host species' resources in modified habitats . Jetz et al . [54] projected the impact of climate and land-use change on global bird diversity , and although greatest impact was expected in temperate areas , species at greatest risk are narrow-range species endemic to the tropics , where range reduction is a result of anthropogenic land conversion . In the present study , Lu . shannoni is projected to have the greatest increase in range size over time , perhaps due to the heterogeneous landscape where this species occurs: aquatic mangrove habitats , arid vegetation of spiny forest , desert , grasslands and xerophilous brushland , temperate vegetation of conifer and cloud forest and perennial , deciduous , and sub-deciduous tropical forest [54] . May et al . [55] have already reported a recent increase in Lu . shannoni's relative abundance and habitat adaptation ( conserved and modified habitats ) at least in Quintana Roo , Mexico . Interestingly , Lu . cruciata , which also inhabits a wide variety of landscapes and is also a potential vector of Leishmania [55]–[57] , has the second greatest projected increase in niche breadth due to CC in the present study . The predominant trend for ENM centroid shift to the northwest was consistent over time for most sand fly species . However , the centroid of temperate species , which have the greatest ENM breadth increase , shifted predominately to the south , a trend previously observed in European birds [58] . Generally , the greater the expected environmental modification ( A2 ) , the greater was the distance shift in niche centroid , indicating that highly fragmented and degraded landscapes have a greater impact on sand fly ENM shifts . Loarie et al . [59] observed a trend for diversity shifts northward toward coastal areas in Californian flora; the species' centroids shifted by an average 151 km , to higher elevations . Huntley et al . [58] analyzing six climate scenarios to 2070-99 for 431 European bird species , observed a mean centroid shift of 258 to 882 km in a direction between 341° ( NNW ) and 45° ( NE ) . Although these previous studies involve distant taxa , they represent evidence for species' ENM centroid shifts away from increasing climate values , positively associated with environmental modification . In addition to potential increases in fundamental niche area and latitude/longitude centroid shifts , the elevation range for certain species have been reported , due to shifting or varying thermoclines and precipitation [57] , [60] . The elevation range for all 28 sand fly species decreases over time and the greater the environmental modification ( A2 ) , the greater the decrease in ENM elevation range . Lutzomyia intermedia has an increased body size and greater dispersal capacity at higher elevations [61] . Previous studies of tropical sand fly species' ENM project an elevation range increase due to habitat modification in the southern tropical lowlands of NCA [51] . This trend is not homogeneous for ERS; some epidemiologically relevant species increase elevation range in CC scenarios ( Lu . anthophora , Lu . diabolica , Lu . olmeca olmeca and Lu . panamensis ) , while others shift to a lower range ( Lu . cruciata ) . Interestingly , an increase in Leishmaniases incidence in higher elevations has been reported in various countries , although it is not clear whether this is due to a prevalence shift of vector species , an increase in parasite prevalence only in certain vector species , or increased human exposure and changes in biotic interactions at higher elevations [62] , [63] . Important changes in the geographic projection of ENM of temperate and broad-range sand fly species are also projected , although the dynamics and degree of these projected shifts are species-specific , with ERS having the least overall shift over time . Niche breadth increase and centroid shifts of potential vector species could contribute to an increase in parasite dispersal and hence an increase in human transmission hazard . Even though the fundamental niche is projected to expand , dispersal capacity of the species will depend upon genetic plasticity , the availability of dispersal routes , and host interactions [59] , [62] . The species with the least shift over time in distribution centroid was Lu . shannoni , which is the same species projected to have the greatest increase in total niche breadth . Lutzomyia cruciata , an ERS currently proposed to form a species complex [64] , is the species with the second largest niche breadth increase and centroid shift . The public health Secretariat in Mexico recognizes only 17 , 000 Leishmaniases cases over the last twenty-two years ( CENAPRECE; www . dgepi . salud . gob . mx; last accessed Dec , 2012 ) . However , there is no effective surveillance program and very poor knowledge by medical personnel of the diseases , and hence the total number of officially recognized cases may be much higher . Since rural populations are the principal group exposed to vector contact [65] , there may be as many as 32 million inhabitants at-risk for exposure to transmission of Leishmaniases in Mexico . Population at-risk is projected to increase to 2080 [18] , and based on significant niche identity , this increase corresponds principally to exposure from Lu . diabolica , Lu . shannoni , and Lu . cruciata . Ecological niche modeling of pathogens has been applied to a broad range of infectious and toxicity-related diseases in order to project potential shifts to the end of the present century: dengue fever and Aedes aegypti [66] , malaria transmission in Africa [67] , plague and tularemia [68] , and Loxosceles reclusa in the US [69] . Fundamental biodiversity analyses have used these same methods to model biotic community interactions and the impact of environmental modification , key issues affecting pathogen dispersal: Argentinian ants [70] , Canadian butterfly species [71] , European birds [58] , amphibians in Australia [72] , mammals in Spain [73] , and maize races in Mexico [74] . In the present analysis , only Lu . cruciata has significant niche identity with that of human infection with Leishmania spp . . González et al . [16] reported an association between recurrent Leishmaniases transmission areas and Lu . panamensis ( 91 . 08% ) and Lu . olmeca olmeca ( 84 . 84% ) , based only on geographic overlap . Additional studies will be required to analyze landscape quality and its impact on niche overlap areas where vector , reservoir , and parasite species interact , in order to extend the use of niche identity analysis within fragmented landscapes . The present analysis of the distribution of NCA sand fly species and their ENM shifts in climate change scenarios predicts range shifts which may modify vector-host interactions and relationships associated with habitat and future land use . Temperate sand fly species , and therefore those with least epidemiological importance , project the greatest ENM changes . Those changes projected for certain epidemiologically relevant tropical species support previous evidence for current , and also highlight future importance of Lu . cruciata as an important vector of Leishmania spp . in México . | The present study models the niche of the most abundant sand fly species in North and Central America , including all proven and incriminated vectors of Leishmaniases , an important neglected tropical disease of the region . The expansion and elevation or centroid shifts of the species' niche are modeled for extreme ( A2 ) and conservative ( B2 ) climate change scenarios to 2020 , 2050 and 2080 . In climate change scenarios , models predict significant niche breadth changes in geographic space , principally in temperate sand fly species , while elevation shifts occur principally in tropical , and greatest , in vector species . Niche centroid shifts for individual species were predominately to the northwest , and secondarily to the northeast . The highest proportion of human population at-risk for contact with a vector species was with Lutzomyia diabolica and Lutzomyia shannoni . Despite the fact that Lutzomyia olmeca olmeca is the only confirmed vector species in Mexico , the present study demonstrates a significant niche identity between Leishmania spp . and Lutzomyia cruciata . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2013 | Current and Future Niche of North and Central American Sand Flies (Diptera: Psychodidae) in Climate Change Scenarios |
Animal heterotrimeric G proteins are activated by guanine nucleotide exchange factors ( GEF ) , typically seven transmembrane receptors that trigger GDP release and subsequent GTP binding . In contrast , the Arabidopsis thaliana G protein ( AtGPA1 ) rapidly activates itself without a GEF and is instead regulated by a seven transmembrane Regulator of G protein Signaling ( 7TM-RGS ) protein that promotes GTP hydrolysis to reset the inactive ( GDP-bound ) state . It is not known if this unusual activation is a major and constraining part of the evolutionary history of G signaling in eukaryotes . In particular , it is not known if this is an ancestral form or if this mechanism is maintained , and therefore constrained , within the plant kingdom . To determine if this mode of signal regulation is conserved throughout the plant kingdom , we analyzed available plant genomes for G protein signaling components , and we purified individually the plant components encoded in an informative set of plant genomes in order to determine their activation properties in vitro . While the subunits of the heterotrimeric G protein complex are encoded in vascular plant genomes , the 7TM-RGS genes were lost in all investigated grasses . Despite the absence of a Gα-inactivating protein in grasses , all vascular plant Gα proteins examined rapidly released GDP without a receptor and slowly hydrolyzed GTP , indicating that these Gα are self-activating . We showed further that a single amino acid substitution found naturally in grass Gα proteins reduced the Gα-RGS interaction , and this amino acid substitution occurred before the loss of the RGS gene in the grass lineage . Like grasses , non-vascular plants also appear to lack RGS proteins . However , unlike grasses , one representative non-vascular plant Gα showed rapid GTP hydrolysis , likely compensating for the loss of the RGS gene . Our findings , the loss of a regulatory gene and the retention of the “self-activating” trait , indicate the existence of divergent Gα regulatory mechanisms in the plant kingdom . In the grasses , purifying selection on the regulatory gene was lost after the physical decoupling of the RGS protein and its cognate Gα partner . More broadly these findings show extreme divergence in Gα activation and regulation that played a critical role in the evolution of G protein signaling pathways .
There are few well-understood examples of how signaling pathways evolved . In particular , it is not known how extant signaling molecules evolve characteristics including intrinsic activity , regulatory mechanisms and binding partners . Neutral selection theory proposes that genes released from constraints are gradually deleted from a genome . However , the processes whereby signaling genes are freed from constraints are not known and uninvestigated . Heterotrimeric G proteins are well characterized molecular switches that are activated in response to extracellular stimuli [1] , [2] . The G protein activation state is determined by the balance between rates of GDP-release ( nucleotide exchange ) and intrinsic GTP-hydrolysis by the Gα subunit of the heterotrimer [1] , [2] . For all metazoan and yeast Gα proteins , GDP-release is slower than GTP-hydrolysis , and thus the G protein is predominantly GDP-bound in its resting state . However , both nucleotide exchange and hydrolysis are conditionally controlled by regulatory proteins in cells [3] . In animals and yeast , G protein-coupled receptors ( GPCR ) accelerate GDP release to favor the active GTP-bound state . Regulator of G Signaling ( RGS ) proteins accelerate GTP hydrolysis to favor the inactive GDP-bound state . In contrast to this paradigm found in animals , Arabidopsis thaliana ( Arabidopsis ) Gα ( AtGPA1 ) spontaneously self-activates without the aid of a GPCR or non-receptor GEF [4] , [5] . Thus , in the absence of regulatory proteins , AtGPA1 would be predominantly GTP-bound [4] , [5] . Instead , AtGPA1 inactivation is regulated in vivo by the single Arabidopsis RGS protein , AtRGS1 [4] , [6] , [7] , which accelerates the intrinsically slow GTPase activity of AtGPA1 [4] , [7] . AtRGS1 is the first identified protein with a domain architecture consisting of an N-terminal 7TM domain fused to an RGS domain [7] , [8] . Plants and animals diverged from each other early in eukaryotic history . Based on recent evolutionary findings [9] , [10] , the plant kingdom is the most distinct group from animal lineages that are within Unikonts [10] . Our recent finding that G protein signaling in Arabidopsis differs greatly from that of animals raised the question of how these distinct signaling modules evolved in eukaryotes . Whether or not Bikonta other than Arabidopsis possess self-activating Gα proteins was unknown and , in cases , controversial . One group reported slow nucleotide exchange for the rice Gα [11] , while another group reported relatively fast nucleotide exchange [12] , albeit slower than the well-characterized Arabidopsis Gα protein [4] , [5] . The Gα protein from Glycine max ( soybean ) may also possess relatively rapid GDP release [13] although there is no direct biochemical evidence supporting this idea . Here we show that the plant kingdom employs G protein activation mechanisms distinct from those found in the animal kingdom . We analyzed plant genomes for G protein signaling components , and purified an informative subset of these components for biochemical analysis . We found that the trait of self-activating Gα was conserved throughout the plant kingdom . However , mechanisms that regulate G protein signal initiation differed throughout the plant kingdom , with some species lacking RGS proteins . We also provide evidence for the evolutionary route from one signal regulation mechanism to another . Specifically , we found in monocots that a single amino acid mutation in Gα disrupted the RGS-Gα interface and may have resulted in subsequent loss of the RGS genes from the genome . Collectively , these characteristics distinguish plant G protein signal regulation from the well-known paradigm from the animal kingdom . More broadly , this study illustrates the mechanism for how a strict functional pair ( i . e . a signaling component and its regulator ) , commonly found in eukaryotes , was disrupted and resulted in rewiring of a cellular signaling network .
To identify signaling modules in the plant kingdom , homologous sequences of Gα , Gβ and Gγ genes were collected from genomic or expression sequence tags ( EST ) databases as described in Materials and Methods ( Table 1 and Table S1 , Figures S1 , S2 , S3 ) . For reference , an evolutionary tree is provided in Figure 1A and includes the species described from here on . Typically vascular plants had one or two Gα genes , but G . max , a partially diploidized tetraploid , had four Gα genes . Physcomitrella patens ( moss ) , a non-vascular land plant , lacked Gα homologs , although another non-vascular plant , Marchantia polymorpha ( liverwort ) , possessed one Gα gene . One or two Gβ genes were encoded in all land plants analyzed , with the exception of soybean , which had four Gβ homologs . Likewise , Gγ genes resembling Arabidopsis Gγ genes [14] , [15] were conserved in all land plants , with a few gene duplications . Notably , moss contained genes encoding the Gβγ dimer , but lacked a canonical Gα protein ( Figure 1A ) . The moss genome encoded a gene ( XP_001772174 . 1 ) homologous to Arabidopsis extra large GTP-binding protein ( XLGA ) , although it should be noted that the moss gene lack a sequence for phosphate-binding loop ( P-loop ) and a glutamate residue in switch II region , each critical for G protein function . Chlamydomonas reinhardtii and Volvox carteri , ( unicellular and multicellular green algae , respectively ) had no homologous genes for Gα , Gβ , Gγ and RGS , but a partial sequence of a Gα homologue was found in the EST database of Coleochaete scutata ( a green alga , JG445935 ) , a descendant of the most probable immediate ancestral group to land plants . These results suggest that non-vascular plant and chlorophyta lost some elements of the heterotrimeric G protein complex in their lineages . Next , we searched for G protein regulatory elements . Previous analysis showed that plants lack canonical G protein coupled receptors [16] , [17] , and our analysis of new plant genomes/ESTs supported this finding . We discovered that genes encoding RGS proteins were not present in the most studied monocots , the cereals , even though RGS genes were present in all other vascular plants ( eudicots , gymnosperms and a spikemoss ) . Although all grasses lacked a standard 7TM RGS protein , the grass , Setaria italica ( foxtail millet ) and the non-grass monocot Phoenix dactylifera ( date palm ) each possessed a gene that appears to encode an RGS protein . Unlike eudicots , however , the S . italica RGS lacked the transmembrane domains ( Figure 1A and Table 1 and Table S1 ) . Two eudicot RGS genes ( Ricinus comunis and G . max RGS2 ) were predicted to have five transmembranes instead of seven transmembranes predicted for the founding member and prototype of the multi-transmembrane domain RGS family , AtRGS1 . No RGS-homologous genes were found in non-vascular plants ( liverwort and moss ) . Together , these results suggest that RGS proteins arose in an ancestor of vascular plants , but RGS genes were subsequently lost in many monocots . We then phylogenetically analyzed the evolution of G protein signaling components . Generally , phylogenies of genes encoding plant Gα , Gβ and 7TM-RGS matched those generated with other genes used for phylogeny construction [18] ( Figure 1B–1D ) . Near the end of angiosperm evolution , monocot and eudicot Gβ had approximately the same branch length from the common ancestor ( Figure 1D ) . However , Gα evolution was accelerated in the monocot branch: the branch length of Poaceae ( grass family ) Gα from the common ancestor with P . dactylifera was almost twice as long as that of date palm Gα ( Figure 1C ) . We hypothesize that this accelerated evolution of monocot Gα subunits compensated for the loss of RGS genes and/or was the result of the loss ( discussed below ) . We included representatives from a eudicot ( A . thaliana AtGPA1 ) , a grass ( Oryza sativa OsRGA1 ) , a gymnosperm ( Pinus taeda PtGα1 ) , and a nonvascular plant ( M . polymorpha MpGα1 ) . First , we characterized the nucleotide exchange rates of these proteins using the non-hydrolysable GTP analog , GTPγS ( Figure 2B and Table 2 ) . Consistent with previous results [4] , [5] , we found that AtGPA1 had fast nucleotide exchange ( Kobs = 5 . 80 min−1 ) . In contrast to an early report that suggested OsRGA1 had slow nucleotide exchange [12] , we found that OsRGA1 exchange nearly matched that of AtGPA1 ( OsRGA1 , Kobs = 0 . 92 min−1 ) . Our value is similar to that published in other studies [4] , [5] , [12] . Likewise , Gα from pine ( PtGα1 , Kobs = 6 . 85 min−1 ) and liverwort ( MpGα1 , Kobs = 1 . 84 min−1 ) also had fast nucleotide exchange . These nucleotide exchange rates were corroborated by measuring the rate of the activation-dependent change in intrinsic Gα fluorescence [19] , [20] ( Figure 2C and 2D ) . Together these data suggest that the trait of fast GDP release is conserved in land plants , and likely arose in a common ancestor of this super group ( Table 2 ) . For a Gα protein to be called “self-activating , ” it must release GDP/bind GTP faster than it hydrolyzes GTP . In other words , the Gα should accumulate in its active form without a regulatory protein . Thus , we measured the rate of Gα-GTP accumulation in the presence of hydrolysable GTP ( Figure 3A–3C ) . In this reaction , activated Gα would only be observed if the rate of nucleotide exchange was faster than the rate of GTP hydrolysis ( i . e . when the Gα protein is “self-activating” ) [5] . All tested plant Gα subunits accumulated in the active state with GTP ( Figure 3A–C ) . Grass Gα ( OsRGA1 ) and eudicot Gα ( AtGPA1 ) displayed sustained activation in the presence of GTP . However , the Gα from liverwort ( MpGα1 ) quickly returned to the inactive form , even in the presence of a 10-fold molar excess of GTP ( Figure 3C ) , suggesting that the liverwort Gα had a fast rate of GTP hydrolysis . To test this hypothesis , we directly measured the intrinsic rates of inactivation of the selected Gα subunits by quantifying release of 32PO4 from [γ-32P]GTP in single turnover GTPase experiments ( Figure 3D and Table 2 ) . AtGPA1 ( Kcat = 0 . 047 min−1 [4] , [5] ) and OsRGA1 ( Kcat = 0 . 052 min−1 ) had slow rates of GTP hydrolysis . In contrast , liverwort MpGα1 ( Kcat = 0 . 87 min−1 ) had a 16-times faster GTP hydrolysis rate than AtGPA1 and OsRGA1 , indicating that MpGα1 efficiently inactivates itself without an RGS protein , yet hydrolysis remains the rate-limiting step . Together , these results suggest that land plant Gα subunits are all “self-activating” due to rapid nucleotide exchange relative to GTP hydrolysis and that the controlled step for activating G signaling is at GTP hydrolysis . The ideal element for this control is a 7TM-RGS protein , represented by the prototype AtRGS1 . However , the absence of 7TM-RGS proteins in grasses indicates an alternative regulatory mechanism must exist in this class . Under neutral selection , genes freed from evolutionary constraint are rapidly deleted from the genome . This implies that 7TM-RGS was released from the strict functional constraint with Gα early in grass family history . To determine how this release may have occurred , we modeled the putative RGS - grass Gα protein interaction interface ( Figure 4C and 4D ) and found that a threonine residue in switch I ( Thr194 of AtGPA1 ) that is critical for interaction with RGS proteins [21] was changed to asparagine in most monocot Gα subunits ( Asn195 of OsRGA1 , Figure 4B ) . This threonine residue is conserved in the RGS-sensitive human Gαi and Gαq family ( Figure 4B ) and is located at the center of the interface with RGS protein [21] . The threonine residue is substituted to lysine in Gα12 and Gα13 , and this class of Gα subunits has dedicated RGS Homology ( RH ) proteins of Rho-family GEFs . Gα12 and Gα13 are not substrates for RGS proteins , which are dedicated GAPs of Gαi and Gαq . A mutation of this lysine of human Gα13 abolishes interaction with the RGS domain of a Gα13 effector , leukemia-associated Rho GEF ( LARG ) [22] . Two monocots were atypical in that they retained RGS-encoding genes . Examination of Gα and RGS sequences from these monocots provided insight into how other monocots may have lost RGS genes . First , the RGS protein in the monocot P . dactylifera has the typical ( i . e . the Arabidopsis prototype ) 7TM-RGS topology and its gene transcription is supported by EST data ( Figure S4 ) . The P . dactylifera Gα has the threonine critical for RGS interaction ( Figure 4A ) . Notably , the P . dactylifera gene is longer than the Arabidopsis AtRGS1 gene by 19 kb , primarily due to the dramatic expansion of intron between the 7TM and RGS domains ( Figure S4A ) suggesting that this part of the gene was subjected to DNA insertion , possibly through transposon activity . The monocot S . italica also encodes a single soluble RGS gene . However , the S . italica Gα has Asn instead of Thr in the switch I region ( Figure 4A ) . These analyses indicate that grass Gα subunits lost the ability to couple with RGS , thus releasing the genetic linkage between the Gα and the RGS protein , although it is also possible that deletion of RGS genes in grasses preceded the Gα mutation . To trace the evolutionary process leading to the deletion of the 7TM-RGS gene in grasses , we surveyed S . italica genomic sequences surrounding the coding region of the single RGS gene ( SiPROV019851m ) , and we found a hypothetical gene upstream of the RGS gene ( SiPROV032159m ) with sequence homologous ( E value = 1e−28 ) to transmembrane helices 2 through 4 of AtRGS1 ( Figure S5 and Table S2 ) . We also found nine ESTs that were homologous to the RGS domain . However , we found no sequence homologous to the region upstream of the 7TM domain ( Figure S6 ) . Moreover , the ratio of change in synonomous vs . nonsynonous residues ( dn/ds ) in comparing the S . italica homologies to their orthologs in date palm and Arabidopsis were higher for the 7TM region than for the RGS region . This suggests that the S . italica RGS domain continued with a function that was under strong purifying selection while the S . italica 7TM domain , although under purifying selection for most of the last 100 million years , has been under neutral or diversifying selection for the last few thousands or millions of years ( Table S3 ) . Closer examination of the assembled sequence provided a clue to the partial gene loss . In the Setaria RGS region , we found two transposons inserted between the conserved and transcribed RGS domain and the apparent 7TM domain ( Figure S6 ) . Lack of EST support , suggest that the 7TM domain became a pseudo gene . One insertion is a partial sequence of a LINE transposon , likely resulting from deletion after insertion because the polyA and target site duplication ( TSD ) are missing . The second insertion was of a previously unknown long terminal repeat ( LTR ) retrotransposon that we named Alubu . Unequal homologous recombination subsequently converted this insertion into a solo LTR with intact TSD . There are 2 additional Alubu solo LTRs ( but no complete elements ) in the current Setaria sequence assembly ( phytozome 7 http://www . phytozome . net/ ) . To identify other possible remnants of the 7TM-RGS gene in other grasses that lack RGS genes , we performed a tBLASTx search using the genomic sequences of the S . italica RGS-homologous region ( segment 13 , bases 1356001–1363646 ) against other monocot genomes . No homologous sequence of S . italica RGS was found in the other grasses analyzed , although the possibility of highly divergent RGS genes in plant genomes is not excluded . These results indicate that a vascular plant ancestor had the 7TM-RGS gene . Furthermore , these analyses suggest that grasses gradually lost the RGS gene once it was uncoupled from the Gα protein by mutation of the RGS-Gα interaction interface . More broadly , these analyses point to the mechanism whereby a single amino acid substitution can lead to rewiring of a new signaling network In this case , the mutation led to neutral selection and loss of a regulatory element from the signaling pathway . Our bioinformatics analyses suggested that the single amino acid substitution in the Gα protein-RGS interface was sufficient to release the RGS protein from evolutionary linkage to the Gα protein . To test this hypothesis experimentally , we substituted the threonine with an asparagine in the extant Arabidopsis Gα protein ( AtGPA1-T194N ) to recapitulate the monocot RGS interaction interface . We also made the comparable reverse substitution in OsRGA1 , a representative monocot Gα protein . These mutations did not affect intrinsic nucleotide exchange and GTP hydrolysis rates ( Figure 5A–5D ) . Next , we quantified interaction between these Gα proteins and the RGS protein from Arabidopsis ( Figure 6A–6E and Table 3 ) . As shown by surface plasmon resonance ( SPR ) analysis , AtGPA1 had high affinity for AtRGS1 ( KD = 17 . 4 nM ) , and OsRGA1 had a relatively lower affinity for AtRGS1 ( KD = 56 . 7 nM ) . We next tested two mutated Gα subunits , AtGPA1-T194N and OsRGA1-N195T . Although these mutations did not affect intrinsic nucleotide exchange and GTP hydrolysis rates ( Figure 5A–5D ) , the T194N mutation reduced AtGPA1 affinity for AtRGS1 by 7-fold ( KD = 115 nM ) . Reciprocally , the N191T mutation increased OsRGA1 affinity for AtRGS1 by 12-fold ( KD = 4 . 83 nM ) . As a second measure of RGS-Gα interaction , we quantified GTPase acceleration by AtRGS1 in a steady-state GTP hydrolysis experiment ( Figure 7A–7C and Table 4 ) . Consistent with the affinities from SPR analysis , the T194N mutation of AtGPA1 reduced the GTPase acceleration by AtRGS1 , and the N195T mutation of OsRGA1 increased GTPase acceleration by AtRGS1 . This change in RGS1 sensitivity conferred by single amino acid substitution was confirmed using enzyme titration assays ( Figure 7D–7H ) . Notably , intrinsic GTP hydrolysis by liverwort MpGα1 was fast ( 1 . 1±0 . 1 min−1 ) and was not further stimulated by AtRGS1 . These results suggest evolution of distinct regulatory systems of plant G proteins in the eudicots , grasses and liverworts . Collectively , our phylogenetic and biochemical analyses suggest that the grass Gα lost the ability to interact with the regulatory molecule early in the evolutional lineage by substituting one critical residue ( Figure 4A ) . The substitution of threonine to asparagine likely occurred in grasses before the loss of the RGS protein , a stage represented in S . italica , which contains the asparagine substitution , yet still encodes a remnant trace of the 7TM-RGS gene . This suggests that the physical uncoupling of Gα with RGS by single amino acid mutation broke the signaling pathway linkage permitting the subsequent deletion of RGS genes in grasses ( Figure 1B ) .
GDP release and GTP hydrolysis by Gα proteins are balanced to establish the steady-state level of the activated Gα subunit of the heterotrimeric G protein complex . In animals , G protein coupled receptors alter this balance to favor the GTP-bound state and relay signals from the outside of the cell to the inside of the cell . Likewise , RGS proteins accelerate GTP hydrolysis to favor the GDP-bound state and terminate intracellular signaling . Our recent discovery that these reactions are differently balanced in animals and Arabidopsis prompted us to examine divergence throughout the lineage and evolution of the G proteins and 7TM-RGS proteins within the plant kingdom . To complement our phylogenetic analyses of plant signaling components , we purified an informative set of plant Gα proteins that spanned the plant kingdom ( Figure 8 ) , and also investigated an amino acid substitution that was deduced to have occurred in the grass ancestral Gα protein . All tested Gα subunits were able to release GDP quickly without any other regulatory protein such as a GPCR or other guanine nucleotide exchange factor ( Table 2 ) ( i . e . they were “self-activating” ) . This finding is consistent with the fact that no unequivocal homologous GPCR gene has been characterized in the plant kingdom [16] , [17] . These results provide powerful evidence that plant G proteins use different regulatory mechanisms than vertebrates to activate and terminate G protein signaling . We previously proposed that AtRGS1 functions as a sugar receptor GAP of AtGPA1 , operating by a sugar-dependent GAP activity [6] , [7] . Here we found that liverworts , representing non-vascular plants , lack RGS genes altogether . In compensation , the liverwort Gα hydrolyzes GTP to GDP quickly without the aid of an RGS protein ( Figure 7C and Table 3 ) . The rates of liverwort GDP-release and GTP-hydrolysis were each fast and similar in value ( Table 2 ) , suggesting that liverwort Gα activity is equally balanced between the two reactions . Thus , liverwort Gα protein is likely regulated by other proteins yet to be identified . In contrast to liverwort , OsRGA1 ( representing monocots ) , shared similar intrinsic activation/inactivation properties with AtGPA1 ( representing eudicots ) . This self-activating property of OsRGA1 was surprising given that all studied grass genomes lost the standard 7TM RGS gene . As with the nonvascular plants , this finding points toward alternative regulatory mechanisms in grasses that were not identified based on homology to known G protein regulators from animals . These results indicate that plant G proteins use at least three different regulatory mechanisms , not only to activate , but also to terminate G protein signaling . In addition to GEFs , mammalian G proteins are also regulated by GDP dissociation inhibitor proteins ( GDI ) , which inhibit GDP release from the G protein and stabilize the GDP-bound state [23] . Since all plant Gα subunits spontaneously release GDP ( Table 2 ) , and some lack RGS proteins , plant G proteins are likely regulated by molecules having GDI activity . While several proteins and chemicals have GDI activity [23] , [24] , to date no GDI has been found in the plant kingdom other than the Gβγ dimer , AGB1/AGG1 , although this was shown to be insufficient to maintain Gα in the inactive state [25] . Our analyses also identified 7TM RGS gene loss in progress in the S . italica genome . It is not possible to determine whether the insertion of transposon-like elements found in the S . italica RGS gene actually caused loss of the 7TM domain function , or whether this functional loss predated the insertion events . Transposable elements are the most abundant DNA sequences coded in plant genomes , and confer rapid rearrangement of plant genome structure [26] , [27] . It is interesting , however , that the S . italica RGS domain continues to be expressed and under purifying selection . The S . italica RGS protein could be specifically coupled with the S . italica Gα with the Asn substitution . However , whether it is still involved in G protein signaling , without the need for the 7TM domain , is not known . The grass genomes that have been extensively sequenced are dominated by species that have been cultivated for centuries , including rice , sorghum , maize , barley , wheat and S . italica ( aka , foxtail millet ) . However , the observed amino acid change in the Gα protein and the loss of a normal 7TM RGS is not an outcome of domestication per se , as we see the same Gα protein and 7TM genome structure in S . viridis ( the wild ancestor of S . italica ) ( unpub . obs . ) . We have shown that deletion of AtRGS1 from Arabidopsis results in increased cell growth and proliferation [7] , [8] , [28] , [29] . Also , Gα mutants in Arabidopsis and rice show defects in their development [28] , [30] . Our analyses raise the intriguing possibility that G protein signaling regulates growth and development with different regulatory mechanisms in grasses and eudicots . Regulators other than 7TM-RGS await discovery , or the grass family could have divergent RH proteins not identified by BLAST search . In mammals , the Gs-class of α subunits lacks a known RGS protein . For the Gq-class of Gα subunits , phospholipase Cβ functions as a GAP of Gαq [31] . Therefore , it follows that G protein activity in grasses or the other plants may be regulated by divergent effecters or the other binding proteins yet to be identified in plants .
The sequences of G protein signaling components were found using BLASTp ( E value<0 . 1 ) against protein database and the translated BLAST ( tBLASTn , E value<0 . 1 ) against genomic DNA sequences registered in Phytozome v7 . 0 ( released on Apr/8/2011; www . phytozome . net ) by using A . thaliana genes as queries . Full-length or partial DNA sequences of Gα , Gβ , Gγ , and 7TM-RGS for Triticum aestivum , Hordeum vulgare , P . taeda , Picea glauca , and M . polymorpha were identified with tBLASTn in the nucleotide collection ( nr/nt ) database or the expressed sequence tags ( EST ) database at National Center for Biotechnology Information ( NCBI ) or the species-specific EST database ( E value<10 ) . The partial DNA sequences were combined to determine the full cDNA sequences . G protein components of P . dactylifera were found using assembled-gene sequences downloaded from Weill Cornell Medical College in Qatar ( http://qatar-weill . cornell . edu/research/datepalmGenome/download . html ) . 7TM-RGS gene of P . dactylifera was assembled manually . Gα genes of P . taeda and P . glauca were cloned from the cDNA libraries and their sequences were determined , because information from the databases was insufficient to define the full length sequence . To screen all the possible RGS-like genes , P . dactylifera RGS and S . italica RGS and RGS domain sequences from H . sapiens RGS4 , G protein-coupled receptor kinase , , LARG and sorting nexin 13 were also used as query sequences ( E value<10 ) . Phylogenetic trees were constructed with MEGA5 . 0 [32] . Full length Gα , Gβ and 7TM-RGS protein sequences were aligned with ClustalW using the following parameters; Gap opening penalty and gap extension penalty for initial pairwise alignment , 10 and 0 . 1; Gap opening penalty and gap extension penalty for multiple alignment , 10 and 0 . 2; Gonnet protein weight matrix; Residue-specific penalties , ON; Hydrophilic penalties , ON; Gap separation distance , 4; End gap separation , OFF; Use negative matrix , OFF . The maximum likelihood ( ML ) trees were made using the Complete-Deletion option of gaps and the JTT ( Jones-Taylor-Thornton ) substitution model [33] with gamma distributed rate variation , which was estimated as the best-fitting substitution model using MEGA5 . 0 . The consensus phylogenetic trees were shown with the bootstrap values from 1000 repetitions . Homo sapiens Gαi1 , Gαq , Gβ1 and Gβ5 were included as out groups . cDNAs of P . taeda RGS and Gα were amplified from the cDNA library and cloned into pENTR-D/TOPO vector . cDNAs corresponding to O . sativa or M . polymorpha Gα were synthesized with optimization of codon usage for E . coli . AtGPA1-T194N and OsRGA1-N191T mutants were created by site directed mutagenesis . The Gα cDNAs were subcloned into pDEST17 ( N-terminal 6×His ) . Recombinant His-tagged Gα proteins were expressed in ArcticExpress RP ( Agilent Technologies ) or Rosetta ( DE3 ) ( Novagen , used only for PtGα1 ) with 0 . 5 mM IPTG at 12°C , solubilized in buffer A ( 50 mM Tris-HCl ( pH 7 . 5 ) , 100 mM NaCl , 5 mM MgCl2 , 5 mM 2-Mercaptoethanol , 10 µM GDP , 1 mM PMSF and 1 µg/ml leupeptin ) with 0 . 25 mg/ml lysozyme and 0 . 2% NP-40 , captured from the soluble fraction with TALON Metal Affinity Resin ( Clonetech ) , washed with buffer A containing 500 mM NaCl and 0 . 1% sodium cholate , and eluted with buffer A including 500 mM imidazole . 5 mM imidazole was added in crude extracts to reduce nonspecific binding . The purified proteins were dialyzed with buffer A and stored in 20% glycerol at −80°C . Recombinant His-AtRGS1 ( 284–459 aa ) protein was prepared with the same method , except that MgCl2 and GDP were removed from buffer A . The rate of GTPγS binding was determined indirectly using intrinsic tryptophan fluorescence of Gα [20] and directly with [35S]GTPγS . The rate of GTP hydrolysis was determined with [γ32P]GTP . For GTPγS binding , GDP-loaded Gα ( 1 µM ) in TEDM buffer ( 50 mM Tris-HCl ( pH 7 . 0 ) , 1 mM EDTA , 1 mM DTT and 5 mM MgCl2 ) was mixed with an equal volume of TEDM buffer containing 5 µM [35S]GTPγS ( about 5000 cpm/pmol ) to start the binding reaction . At a given time points , 100 µl aliquots were quenched in 1 ml of ice-cold wash buffer ( 20 mM Tris-HCl ( pH 7 . 5 ) , 100 mM NaCl and 25 mM MgCl2 ) containing 50 µM GTP and immediately vacuum-filtered onto nitrocellulose . Filters were quickly washed three times with 3 ml of ice-cold wash buffer . The total amount of 35S bound to the filter was quantified by scintillation counting . For single-turnover GTP hydrolysis reactions , Gα subunit ( 800 nM ) was preloaded with radioactive [γ-32P]GTP in TEDL ( 50 mM Tris-HCl ( pH 7 . 5 ) , 10 mM EDTA , 1 mM DTT , and 0 . 05% lubrol-PX ) for 30 min on ice . The hydrolysis reaction was then started by adding 450 µl of TMDL+GTPγS ( 50 mM Tris-HCl ( pH 7 . 5 ) , 40 mM MgCl2 , 1 mM DTT , 0 . 05% lubrol-PX , and 400 µM GTPγS ) into 1 . 2 ml of preloaded Gα . At a given time point , duplicate 100 µl aliquots were taken into 1 ml of charcoal ( 25% ( w/v ) in 50 mM phosphoric acid ( pH 2 . 0 ) ) to remove non-hydrolyzed [γ-32P]GTP and proteins . The charcoal tubes were centrifuged , and amount of 32PO4 hydrolyzed was measured by scintillation counting of the centrifuged supernatants . GTP or GTPγS binding with Trp fluorescence and steady state GTP hydrolysis were performed as described previously [5] , [25] . Briefly , 400 nM Gα protein was incubated in a cuvette with 1 ml of TEMNG buffer ( 25 mM Tris-HCl ( pH 8 . 0 ) , 1 mM EDTA , 5 mM MgCl2 , 100 mM NaCl , and 5% glycerol ) . 800 nM GTP or 5 µM GTPγS was added to the cuvette and the change in the intrinsic fluorescence of Gα protein ( excitation at 284 nm , emission at 340 nm ) was monitored . Affinity between 2 different proteins was measured by Surface Plasmon resonance technology using BIAcore 2000 ( GE Healthcare ) . His-tagged AtRGS1 ( 284–459aa ) was immobilized on sensor chip CM5 with ammine coupling . Temperature , flow rate or running buffer were 25°C , 10 µl/min , or 10 mM Hepes , 150 mM NaCl , 3 mM EDTA , 0 . 005% Tween-20 , 100 µM GDP and 10 mM MgCl2 , respectively . Seven different concentrations ( 6 . 25 , 12 . 5 , 25 , 50 , 100 , 200 and 400 nM ) of His-AtGPA1 , AtGPA1-T194N , RGA1 , RGA1-N195T or Gβγ ( AGB1/AGG1 ) prepared in running buffer with 20 mM NaF and 100 µM AlCl3 were flowed onto the sensor chip for 3 min . Dissociation was monitored for 5 min , and the sensor chip was washed with the same running buffer for 10 min at a flow rate of 30 µl/ml . The association ( ka ) and dissociation ( kd ) rate constants were obtained by fitting the original sensorgrams with a 1∶1 Langmuir binding model . | Extracellular signals activate intracellular changes that lead to cell behaviors . This spatial coupling is mediated by cell-surface receptor activation of the heterotrimeric G protein complex located on the cytoplasmic side of the plasma membrane . Unlike the case for metazoans , plant G proteins are constitutively active . Plants use multiple mechanisms to keep the G protein complex in its resting state , and activation occurs by inhibition of this property . One mechanism involves a cell surface receptor that accelerates the return to the resting state through direct interaction with the G protein at a specific protein interface . This unique protein , AtRGS1 , has both an animal like receptor domain and a domain ( RGS box ) responsible for accelerating deactivation . One group of plants ( cereals ) lost this protein through , first , a mutation in the protein interface that reduces the affinity for the RGS box to the G protein , followed by gene loss . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"signal",
"transduction",
"molecular",
"cell",
"biology",
"biology",
"evolutionary",
"biology",
"molecular",
"biology"
] | 2012 | G Protein Activation without a GEF in the Plant Kingdom |
Many algorithms that compare protein structures can reveal similarities that suggest related biological functions , even at great evolutionary distances . Proteins with related function often exhibit differences in binding specificity , but few algorithms identify structural variations that effect specificity . To address this problem , we describe the Volumetric Analysis of Surface Properties ( VASP ) , a novel volumetric analysis tool for the comparison of binding sites in aligned protein structures . VASP uses solid volumes to represent protein shape and the shape of surface cavities , clefts and tunnels that are defined with other methods . Our approach , inspired by techniques from constructive solid geometry , enables the isolation of volumetrically conserved and variable regions within three dimensionally superposed volumes . We applied VASP to compute a comparative volumetric analysis of the ligand binding sites formed by members of the steroidogenic acute regulatory protein ( StAR ) -related lipid transfer ( START ) domains and the serine proteases . Within both families , VASP isolated individual amino acids that create structural differences between ligand binding cavities that are known to influence differences in binding specificity . Also , VASP isolated cavity subregions that differ between ligand binding cavities which are essential for differences in binding specificity . As such , VASP should prove a valuable tool in the study of protein-ligand binding specificity .
The comparative analysis of protein structures is widely used to infer protein function . Geometric alignment of entire structures or of individual domains can reveal that two proteins are related even if this is not evident from sequence . Numerous techniques have been developed for this purpose , most based on either the superimposition of the polypeptide backbone [1]–[5] , the comparison of geometric graphs [6] , [7] or the alignment of a matrix of distances between individual amino acids [8] . A second type of approach involves the direct comparison of functional sites , such as the geometric disposition of catalytic residues [9]–[13] or the comparison of the shapes of cavities on the protein surface [14]–[18] . Surface representations of proteins [19]–[24] are , in particular , widely used as they reveal shape recognition features that underlie binding specificity . Most approaches reported to date have focused on remote homology detection with the goal of identifying similarities between two or more proteins that can give hints as to biological function . However , a large class of phenomena depend on the ability of closely related proteins to bind similar but non-identical ligands . In such cases the function of a protein as normally defined is well-known but its binding preferences may not be . The problem we are specifically addressing concerns the case where two or more proteins have been structurally aligned and it is of interest to identify conserved and varying regions in their binding cavities . Conserved regions , for example , might bind a molecular fragment that is common to substrates acted on by the entire protein family , while the source of differences in intrafamily specificity would likely reside in regions where cavities vary . Our approach is based on a volumetric representation of binding cavities ( Figure 1 ) that is generated with a new program , VASP ( Volumetric Analysis of Surface Properties ) . VASP uses Constructive Solid Geometry ( CSG ) to compare regions in space defined by a polyhedral boundary [25] , [26] . Developed originally for the computer aided design of machine parts [26] , and adapted later for computer graphics [25] , CSG enables volumetric unions , intersections , and differences of two aligned regions to be computed as if they are solid objects . These CSG operations are a novel tool in the analysis of protein structures because they yield an approximation to the shape of solid regions that is varying or conserved , among protein structures and protein cavities , that is not possible with existing structure comparison methods . The solid representations used in VASP differ fundamentally from point-based and surface-based representations , which are used in existing methods to define and compare cavities . Point-based representations compare the geometric coordinates of atoms related by one-to-one correspondences . These correspondences cannot be fully constructed between all atoms of sidechains with different lengths , forcing the simplification of sidechain geometry into pseudo-atom or backbone-only representations . In contrast , solid representations compare regions defined by the molecular surface , whose shape reflects the position of any atom without simplification . Solid and surface-based representations both measure differences in molecular shape and curvature . However , surface representations cannot disassemble surface cavities to isolate conserved ( intersecting , Figure 1i ) or varying ( difference , Figure 1g , 1h ) regions , as VASP does with CSG , because surface representations do not represent the interior or exterior of a boundary surface . To our knowledge , VASP is the first application of CSG to protein structure comparison , although small molecules have been previously compared in a related manner with lattice points [27] and voxels [28] , which are both precursors to Marching Cubes [29] , the origin of our technique . These earlier techniques use rectilinear representations that cannot approximate the curvature of molecular surfaces , as VASP does . Other volumetric methods have also been developed to capture topological differences in electrostatic isocontours [30] and to represent regions where substrates overlap for the design of inhibitors that evade drug resistance [31] . The input to VASP includes the definition of binding cavities obtained from manual observation or cavity detection algorithms [16] , [22] , [32]–[36] , and structural alignments of entire proteins [1]–[15] , [17] , [18] . VASP then uses CSG comparisons of aligned cavity volumes to enable several unique capabilities . Unlike existing methods , VASP can identify individual amino acids and cavity subregions that create structural differences in ligand binding cavities that influence binding specificity . Such functionalities suggest novel applications in protein engineering and design and in the detailed characterization of the determinants of ligand binding specificity . We demonstrate VASP's capabilities with applications to the START domains and to the peptide binding cleft of serine proteases .
As input , Marching Cubes requires the desired output resolution , which specifies how finely the output region will be approximated , the desired CSG operation , union , intersection or difference , and two closed regions A and B ( Figure 2a ) , defined by their surface boundaries SA and SB , representing , in this work , aligned cavities . The output of Marching Cubes is a region represented by a boundary surface that is approximated with a triangular mesh ( Figure 2j ) . Using intersection as an example , the overall procedure ( Figure 2 ) is to approximate the shape of the overlapping region ( Figure 2a ) shared by A and B . First , we construct an axis aligned cubic lattice ( Figure 2b ) so that , along any dimension , every triangle of A and B is within the bounds of the lattice . We interpret the lattice as a grid of “lattice points , ” incrementally spaced along the primary axes according to the desired output resolution , or as a set of “lattice segments” connecting pairs of co-axial lattice points , or as a collection of identically sized “lattice cubes” sharing lattice segments . The lattice is a scaffold for generating the triangles of the output surface . Second , each lattice point p is determined to be either inside or outside the overlapping region by first testing if p is inside or outside A and B , individually ( Figure 2d ) . We determine if p is inside A by generating a randomly oriented ray originating at p . A is not infinitely large , so the ray must eventually extend outside SA , perhaps intersecting the triangles of SA several times . Beginning from the outside , we count these intersections backwards along the ray , crossing into and out of A each time the ray passes through SA . Therefore , for an even number of intersections ( Figure 2c1 ) , p is outside A . For an odd number of intersections , p is inside A . We apply the same even/odd method to test if p is inside B . If p is inside A and p is inside B , then p must be inside the overlapping region , as illustrated in Figure 2c2 . Otherwise , p must be outside the overlapping region . The third step begins by selecting lattice segments that connect a lattice point inside the overlapping region to a lattice point outside the overlapping region , as shown in Figure 2e . Since the overlapping region of two closed regions must be closed , all selected segments necessarily exit the overlapping region at a “crossing point” p0 ( Figure 2g ) where the selected segment intersects SA or SB or both . If only one of SA and SB intersect the selected segment , as shown in Figure 2f1 , or if SA and SB intersect at the same point , then p0 is that point of intersection . If SA and SB intersect the selected segment at different points , we call these points pA and pB . If pA is inside B , then pA is on the border of A but still inside B , so pA must be at the border of the overlapping region , and thus p0 = pA . Conversely , if pB is inside A , as shown in Figure 2f2 , then , for the same reasons , p0 = pB . Finally , we analyze each lattice cube . For each cube , there are 28 = 256 possibilities for the interior/exterior state of its 8 lattice points . Each state corresponds to a unique way for one or more parts of the output surface to pass through the lattice cube , leaving some combination of the lattice points inside or outside the overlapping region . The crossing points indicate precisely where the border of the overlapping region intersects with the lattice segments of the cube . All that remains is to connect the crossing points with triangles to approximate the border of the overlapping region inside the cube , as shown with four examples in Figure 2h . Since there exists 256 different triangular configurations , a lookup table , described elsewhere [29] , provides a triangular configuration for every possibility . Notably , the triangles have a directional orientation , defined to face away from the interior of the surface . To denote the orientation of a triangle , a fact we use later , the corners are enumerated in counterclockwise order , when viewed from an exterior perspective . These “output triangles” are depicted as black dotted lines in Figure 2i , since the figure is two dimensional . The output triangles approximate the border of the overlapping region , but are not necessarily identical to the triangles of either SA or SB . Proper selection of the output resolution can reduce inaccuracies in the output surface . The final output region ( Figure 2j ) is within the surface composed by the output triangles . As input , we begin with a closed region A represented by a boundary surface SA composed of oriented triangles . From the input , we compute the centroid c of all triangle corners ( Figure 3a ) . Looping through each triangle t in SA , we keep a running total , V , initially zero , of the volume within SA , while performing the subroutine below . After all triangles have been considered , the final value of V is the volume within SA . First , we compute the centroid of the triangle , tc , and the normal vector of the triangle , tn . tn is perpendicular to the plane of t , but for any plane , there are two perpendicular directions . Using the fact that t is oriented , we select tn to point away from the inside of SA ( Figure 3b ) . Second , we determine if t faces away from c or towards c , by measuring the dot product d between tn and the vector ( tc-c ) ( Figure 3c ) . Next , we generate the tetrahedron T , with corners based on the three corners of t , and the global centroid c . We measure the volume of T , v ( T ) , using Tartaglia's rule , described below . If d is positive , we add v ( T ) to V ( Figure 3d ) , if d is negative , we subtract v ( T ) from V ( Figure 3e ) . If d is zero , v ( T ) is also zero , in which case we do nothing and proceed to the next triangle . Tartaglia's Rule [40] is a three dimensional generalization of Heron's Formula for the area of a triangle [41] . Here , the volume V of a tetrahedron with corners a , b , c , and d , can be evaluated with the expressionwhere the distance between two corners x and y is dxy . We use SCREEN [35] to identify cavities as input for VASP . SCREEN produces lists of amino acids nearby the cavity , which we convert into a volumetric representation using the procedure illustrated in Figure 4: First , GRASP2 [3] is used to compute triangular meshes approximating the molecular surface based on a 1 . 4 Å probe ( Figure 4a ) , and an “envelope” surface based on a 5 . 0 Å probe ( Figure 4b ) . Second , all patches of triangles on the molecular surface with corners further than 2 Å from any location on the envelope surface are identified as the base of each surface cavity ( Figure 4c ) . Third , the patch closest to the amino acids produced by SCREEN is manually selected for the analysis that follows . Fourth , for every triangle in the selected patch , the closest atom in the structure is found and the amino acid it belongs to is added to a non-redundant list . This list contains all amino acids lining the selected patch ( Figure 4d ) . Fifth , the qhull program [42] , is used to compute the convex hull of the Van der Waals spheres of the amino acids lining the selected patch ( Figure 4e ) . From the region within the convex hull , the region within the molecular surface is removed using the CSG difference operation ( Figure 4f ) , as is the region outside the envelope surface ( Figure 4g ) . The resulting region ( Figure 4h ) defines the cavity . Occasionally , small disconnected regions are created in this process . All but the largest , based on surface area , are removed . In addition to SCREEN , other methods can be used to identify cavities as input for VASP . Cavities described by lists of amino acids , generated with algorithms for cavity detection [33] , [35] or local structural comparison [6] , [9] , [11]–[13] , [15] , [17] , [18] , can be converted into volumetric representations with the procedure described above . Cavities described with surfaces [20]–[23] , [34] , [35] , such as the exterior triangles of an alpha shape within a CAST pocket [34] , can be converted into volumetric representations by using the surface as if it was selected in Step 3 , above . CSG can also be used to define a subsite of a cavity . First , we follow the procedure described in Figure 4 to represent the entire cavity . Second , we position spheres in the subsite of interest based on the coordinates of bound ligands and select a radius for each sphere that is large enough to overlap the entire subsite ( Figure 4i ) . Third , we compute the CSG union of all the spheres . Fourth , we calculate the intersection between the sphere union and the cavity ( Figure 4j ) . The resulting region defines the shape of the subsite , without including the wider cavity . GRASP2 surfaces [3] , using Van der Waals radii taken from [43] , are exceptionally precise approximations of the molecular surface , averaging 384461 triangles per surface , and triangular area averaging . 026 Å2 on our data set . Some GRASP2 surfaces contain topological discontinuities where single contiguous surfaces are represented with disconnected patches . Input surfaces exhibiting topological discontinuities were first fixed using Polymender [44] . Cavities obtained from a given family of proteins were clustered by “volumetric distance” V ( x , y ) , where x and y are cavities , x∩y is the volumetric intersection of x and y , and V ( K ) represents the volume of a given region K , in Å3 . The shape of the region x∩y was determined with the CSG intersection , and V ( K ) was evaluated with the Surveyor's Formula . V ( x , y ) is the proportion of intersecting volume relative to the maximum theoretical degree of intersection , the volume of the smaller region , and thus a measure of volumetric similarity between x and y . We computed V ( x , y ) for all pairs of cavities in each set . Using the “neighbor” tool from Phylip [45] , we summarized the overall organization of volumetric conservations and variations using UPGMA clustering ( Unweighted Pair Group Method with Arithmetic mean , [46] ) of V ( x , y ) , over all pairs of cavity regions . We also clustered proteins in our data set using other metrics of similarity . Multiple sequence alignments were computed with ClustalW 2 . 0 . 7 [47] and the most parsimonious phylogeny was constructed with the “protpars” tool from Phylip [45] . Phylogenetic trees generated in this manner are unrooted , so a logical root was selected manually for visual comparison . Backbone structure similarity was computed with Ska [5] , and the RMSD of corresponding Cα atoms was clustered by UPGMA using the “neighbor” tool from Phylip . We begin with aligned proteins X and Y , with cavities x and y . First , we generate the molecular surface Sa of each amino acid a in X , individually . Second , we compute the CSG intersection between a and y , and measure the volume of the intersection using the Surveyor's Formula . Amino acids with a nonzero volume of intersection cause x to have a different shape than y . Regions conserved among aligned cavities are determined by repeated application of CSG intersection . Regions occupied by at least one cavity , among several , are determined with the CSG union . Regions in a cavity x that are not in a cavity y are determined with the CSG difference . For example , the region conserved in all trypsin cavities that overlaps no elastase cavity , illustrated in Figure 9d , is evaluated as the difference between the intersection of all trypsin cavities and the union of all elastase cavities . The Protein DataBank ( PDB - 06 . 15 . 2008 ) [48] contains the structures of 28 START domains and 582 serine proteases , from the chymotrypsin , trypsin , and elastase subfamilies . From each set , we removed functionally undocumented and mutant structures and then structures with greater than 90% sequence identity , leaving a non-redundant subset of 11 START domains and 14 serine proteases . Filtering in this order maximized the number of diverse representative structures , identifying START domains and serine proteases averaging 12% and 47% pairwise sequence identity , respectively . Hydrogen atoms , resolved in only four structures in our dataset , were removed for consistency . The START domains are lipid transporters whose available structures belong to distinct subgroups that have well documented ligand binding specificities [49] . Three proteins in our set exhibit a specific affinity for cholesterols: MLN64 ( pdb: 1em2 ) [50] , StarD5 ( pdb: 2r55 ) [49] , and StarD4 ( pdb: 1jss ) [51] . Five others exhibit binding with a wide range of lipids , including fatty acids , cytokinins , and flavonoids [52] and are referred to here as having “broad specificity” . These proteins include allergen-like proteins from birch ( pdb: 1bv1 ) , cherry ( pdb: 1e09 ) , celery ( pdb: 2bk0 ) , yellow lupine ( pdb: 1xdf ) , and mung bean ( pdb: 2flh ) . The remaining functionally characterized proteins in our set include the human phosphatidylcholine transfer protein ( pdb: 1ln1 ) , which only binds phosphatidylcholines [53] , human ceremide transporter ( CERT ) ( pdb: 2e3m ) , a highly specific transporter of ceremides of specific lengths [54] , and the yeast oxysterol binding protein Osh4 ( pdb: 1zht ) , which prefers oxysterols to cholesterols [55] . Using Ska [5] , the START domains were aligned to the major birch allergen ( pdb: 1bv1 ) , which was selected randomly . Cavities were defined in the START domains as described above , without subsite definition . The serine proteases were aligned via Ska to bovine gamma-chymotrypsin ( pdb: 8gch ) , because 8gch exhibits a tryptophan bound in the S1 specificity pocket of the larger peptide binding cleft . The S1 pocket was defined with the subsite technique described above . 5 Å spheres were positioned at all tryptophan atoms and at five waters at the bottom of the 8gch S1 pocket . With all S1 pockets aligned onto the S1 pocket of 8gch , the spheres defined the S1 subsite cavity in all serine proteases . Manually placed waters can also be used to define known subsites , but bound waters and substrate provided an objectively defined subsite for demonstration purposes . Structural alignments of all proteins in our datasets to an individual structure did not create bias in our results . As described in Text S2 and Figures S1 , S2 , S3 , rerunning our results on a realignment to any other dataset member produced no major differences in our results . VASP was developed in ANSI C/C++ and compiled on gcc 3 . 4 . 6 , for 32 and 64 bit ×86 computing platforms . Visualization was implemented using the OpenGL C/C++ library on Windows XP platforms running Intel Xeon , AMD Athlon 64 , and Nvidia Geforce 6800 and 7600 chipsets . Experimentation was performed on quad-core Opteron systems with at least 2 gigabytes of random access memory per core . VASP , a single threaded process , used one core and approximately 1 gigabyte of RAM . All results were computed at . 5 Å resolution , which produced accurate results with practical runtimes: CSG operations converting a known functional site into a volumetric representation involved the entire protein structure , and an average of 1 . 04 million voxels , 384 , 461 triangles , and 12 . 8 minutes ( 1355 voxels/sec ) . CSG operations computing the intersection of cavities , rather than whole structures , involved an average of 177 , 490 voxels , 59 , 677 triangles , and 5 . 9 minutes of computation ( 494 voxels/sec ) . Finally , CSG operations for individual amino acids involved an average of 2 , 958 voxels , 2 , 915 triangles , and 2 . 77 seconds ( 1068 voxels/sec ) . START domain cavities generally had much larger volume than serine protease cavities , and CSG runtimes reflected these differences . Additional runtime details are provided in Table S2 . To further clarify the runtime performance of VASP , in the Supporting Materials , we have provided additional performance details describing the runtime of typical CSG operations ( Text S3a , Figure S4 ) and the runtime/accuracy tradeoff at lower resolutions ( Text S3b , Figure S5 , S6 ) . These observations suggest that . 75 Å resolution can also yield reasonable accuracy , though the clustering of START domains was slightly less accurate at this resolution . In the future , adaptive approaches , using oct-trees instead of uniform voxels , and more efficient strategies for assessing the interior/exterior state of a given point , such as those described elsewhere [44] , could potentially reduce runtimes and memory usage while maintaining accuracy .
Figure 5 reports a clustering of START domains based on volumetric distance . It is evident that the tree separates the 11 proteins into distinct groups that are well correlated with their binding preferences . This separation indicates that VASP is successful in capturing cavity shape similarities and differences among the different proteins that relate to binding preferences . The single outlier in the tree is yellow lupine PR-10 ( pdb 1xdf ) which is not grouped with other broad specificity START domains . However , 1xdf has a kinked C-terminal helix that fills the ligand binding site and indeed the protein cannot bind ligands in this conformation [56] . Thus , volume-based classification correctly discriminates between 1xdf and the other broad-specificity START domains . It should be noted that global sequence and structure alignment also separated START domains into the correct clusters ( Figure S7 ) , but in these cases , 1xdf was included as part of the broad specificity cluster . Thus , global comparisons failed to detect a local change of cavity shape in the binding cavity . We used VASP to identify the regions of the protein responsible for the unusual binding properties of 1xdf . Figure 6 illustrates the degree of volumetric intersection between individual amino acids in 1xdf and the cavities of the other broad-specificity START domains , 1bv1 , 1e09 , 2bk0 , and 2flh . For most amino acids , the volume of intersection averaged 8 Å3 ( standard deviation 16 Å3 ) over all cavities . That so many amino acids have at least a small degree of overlap is due to the fact that all of these proteins have a very large internal cavity that has some degree of contact with almost every residue . In contrast to this baseline variation , residues 137–144 exhibited unusually high intersection volumes with all cavities considered , averaging 60 Å3 , with several surpassing 100 Å3 . These residues are located at the center of the kinked C-terminal helix that fills the binding site of 1xdf and prevents ligand binding ( inset , Figure 6 ) . Our ability to identify these residues illustrates how VASP can be used to identify locations in a structure that are responsible for specificity . In serine proteases , affinity for specific sequences of amino acids is associated with individual specificity pockets , S4 , S3 , . . S1 , S1' , S2' . . S4' , that recognize substrate residues P4 , P3 , . . P1 , P1' , P2' , . . P4' [57] . In trypsins , S1 exhibits a narrow affinity for amino acids with positively charged side chains [58]; in chymotrypsins , S1 exhibits greatest affinity for large hydrophobic sidechains [59] , and in elastases , S1 has greatest affinity for small hydrophobic sidechains [60] . Figure 7 illustrates the clustering of serine protease S1 pockets based on volumetric distance . Elastase S1 pockets were clustered tightly together and separately from the other serine proteases . With the exception of fire ant chymotrypsin ( pdb: 1eq9 ) , trypsins are also clustered tightly together , and separately from other serine proteases . Bovine chymotrypsin ( pdb: 8gch ) is separated distinctly from the trypsins and from elastases , but also from fire ant chymotrypsin ( pdb: 1eq9 ) . Global sequence and structure alignment separated the serine proteases similarly or less well ( Figure S8 ) . Figure 8 illustrates the degree of volumetric intersection between the individual amino acids of the serine proteases and the S1 cavity of bovine chymotrypsin ( pdb 8gch ) . Intersection volumes were almost always zero or near zero , with a few distinct exceptions: In elastases ( Figure 8a ) , Val216 and Thr226 occupy an average of 43 Å3 and 31 Å3 , respectively , within the 8gch cavity region . These amino acids are known to truncate the S1 pocket ( inset , Figure 8a ) to generate specificity for small hydrophobic amino acids [61] . In trypsins ( Figure 8b ) , Asp189 occupies an average of 25 Å3 within the 8gch cavity and is primarily responsible for the specificity of trypsin for basic residues [62] . Figure 8b illustrates how Asp189 occupies the bottom of the chymotrypsin cavity , which orients the negatively charged carboxylate group of Asp189 to face substrate resides and to sterically hinder the binding of aromatic amino acids . VASP also identifies Glu192 , a residue conserved among trypsins that occupies an average of 12 Å3 in the 8gch cavity that is not occupied by the Met192 conserved among chymotrypsins . Finally , in fire ant chymotrypsin ( pdb: 1eq9 ) ( Figure S9 ) , VASP identifies Asp226 , which exhibits a 32 Å3 overlap with the bovine chymotrypsin ( 8gch ) cavity . Residue 226 is typically glycine in mammalian chymotrypsins , and , as reported elsewhere [63] , Asp226 must rotate out of the way to accommodate the aromatic residues preferred by chymotrypsin . Figure 9 illustrates several regions within the serine protease S1 cavities that are volumetrically conserved or varying . The first region , where all S1 subsites in our dataset overlap ( Figure 9a ) occupies a volume of 107 Å3 and is located at the entrance of the S1 subsite . This global intersection includes a protruding region that extends into the center of the oxyanion hole , a tiny cleft critical for stabilizing hydrolysis reaction intermediates [64] . Only the central portion of the oxyanion hole was conserved among all serine proteases because of slight variations in structural alignments . It is clear that in any serine protease , if any region of the global intersection is obstructed , either P1 would be hindered in entering the S1 cavity or the oxyanion hole would be unable to stabilize reaction intermediates . By determining the global intersection of all S1 cavities , VASP can thus identify functionally significant subregions . The second region we studied , a 198 Å3 volume where all trypsin cavities overlap ( Figure 9b ) exhibits a distinct 70 Å3 protrusion that does not overlap with the region occupied by any elastase cavity ( Figure 9c ) . This conserved cavity protrusion accommodates the longer sidechains bound by trypsin S1 pockets that are occluded by elastase S1 pockets . Figure 9d illustrates one example where the peptide Gly-Ala-Arg , bound to Fusarium oxysporum ( pdb: 1fn8 ) , clearly extends its Arginine sidechain into the conserved cavity protrusion . By computing the volumetric difference between the intersection of all trypsins and the union of all elastases , VASP can identify conserved variations between subfamilies of serine proteases that influence specificity for different ligands .
We have presented a new volumetric method for the comparison of protein cavities that is embodied in the VASP program . To our knowledge , VASP is the first program capable of comparing cavities via CSG and it therefore enables a new approach to the characterization of protein binding sites . We demonstrate in an application to START domains that VASP is capable of reproducing known ligand binding specificities and of identifying differences in cavity shapes among proteins that , based on global sequence or structure similarity , might have been expected to be similar . Such differences can result from variations in backbone or sidechain conformation , which are two factors contributing to subtle changes in the shape of binding cavities that would otherwise be hard to detect . We demonstrate a number of applications of VASP that are not possible with existing methods . One involves the identification of amino acids that contribute to differences in cavity shape . We identified several such amino acids among the START domains and serine proteases and , in each case , reproduced known determinants of ligand binding . A second application is the identification of conserved and varying regions in protein cavities . Among the S1 subsites of the serine proteases , VASP identified conserved regions that are critical for ligand binding , and varying regions that selectively accommodate certain ligands . Overall , we find that VASP creates new opportunities to comparatively analyze and isolate the structural influence of individual elements within protein cavities . As a first step in the comparison of protein and cavity shape via CSG , VASP exhibits considerable potential for broader applications . When applying VASP more broadly , input structure alignments could include local structure alignments , which would enable proteins with different folds but similar functional sites to enter the analysis . Likewise , as VASP is not a cavity detection algorithm , methods for converting the wide range of cavities detected by existing methods [16] , [22] , [32]–[35] into a volumetric representation could allow a broader space of input to be analyzed . VASP has useful applications in contexts where existing protein structure comparison techniques have not been applied . For example , efforts to engineer proteins with altered binding specificities face the practical challenge of being able to test only a few mutants from a combinatorial space of possibilities . By identifying amino acids that influence differences in cavity shape , VASP can suggest a set of mutations to consider . Another possible application is for the annotation of ligand binding specificity on function annotation servers: Given a query protein , function annotation servers can find neighbor proteins with global structure similar to the query . Using VASP , neighbors with bound ligands can be analyzed locally , at their binding sites , to assess volumetric similarity with a known or predicted binding site on the query . Patterns of local volumetric similarity and variation between the query and neighbor might correlate with patterns of ligand binding preferences . Together with other sources of information , volumetric comparison of structurally aligned proteins may thus offer an important tool in protein engineering and function annotation . | Proteins carry out vital and specific functions by physically binding other molecules . Understanding specificity , the preferential binding of certain molecules to one another , is essential for numerous medical and industrial applications . Given the structure of a protein with unknown function , algorithms are available that suggest hypothetical functions based on structural similarities to better-studied proteins , even at vast evolutionary distances . In contrast , few algorithms identify structural differences that relate to differences in specificity among closely-related proteins . To address this problem , we present a Volumetric Analysis of Surface Properties ( VASP ) . VASP differs from existing methods because it compares solid representations of protein structures and cavities based on principles from computer graphics and computer aided design . In our results , solid representations enabled VASP to isolate elements of protein structure that create differences in binding sites and thereby lead to differences in binding preferences . These observations point to applications for the annotation and engineering of protein specificity . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"computational",
"biology/macromolecular",
"structure",
"analysis",
"biochemistry/bioinformatics"
] | 2010 | VASP: A Volumetric Analysis of Surface Properties Yields Insights into Protein-Ligand Binding Specificity |
The 22q11 . 2 deletion syndrome ( 22q11 . 2DS; velo-cardio-facial syndrome; DiGeorge syndrome ) is a congenital anomaly disorder in which haploinsufficiency of TBX1 , encoding a T-box transcription factor , is the major candidate for cardiac outflow tract ( OFT ) malformations . Inactivation of Tbx1 in the anterior heart field ( AHF ) mesoderm in the mouse results in premature expression of pro-differentiation genes and a persistent truncus arteriosus ( PTA ) in which septation does not form between the aorta and pulmonary trunk . Canonical Wnt/β-catenin has major roles in cardiac OFT development that may act upstream of Tbx1 . Consistent with an antagonistic relationship , we found the opposite gene expression changes occurred in the AHF in β-catenin loss of function embryos compared to Tbx1 loss of function embryos , providing an opportunity to test for genetic rescue . When both alleles of Tbx1 and one allele of β-catenin were inactivated in the Mef2c-AHF-Cre domain , 61% of them ( n = 34 ) showed partial or complete rescue of the PTA defect . Upregulated genes that were oppositely changed in expression in individual mutant embryos were normalized in significantly rescued embryos . Further , β-catenin was increased in expression when Tbx1 was inactivated , suggesting that there may be a negative feedback loop between canonical Wnt and Tbx1 in the AHF to allow the formation of the OFT . We suggest that alteration of this balance may contribute to variable expressivity in 22q11 . 2DS .
The 22q11 . 2 deletion syndrome ( 22q11 . 2DS ) , also known as velo-cardio-facial syndrome ( MIM# 192430 ) or DiGeorge syndrome ( MIM# 188400 ) is a congenital malformation disorder that is caused by a hemizygous 1 . 5–3 million base pair ( Mb ) deletion of chromosome 22q11 . 2 . It occurs with a frequency of 1:1 , 000 fetuses [1] and 1:4 , 000 live births [2] . Approximately 60–70% of affected 22q11 . 2DS individuals have congenital heart disease ( CHD ) due to malformations of the aortic arch and/or cardiac outflow tract [3] . There are over 46 known coding genes in the 3 Mb region , including TBX1 ( T-box 1; MIM# 602054 ) , encoding a T-box containing transcription factor [4] . TBX1 has been considered the strongest candidate gene for CHD , based upon studies of mouse models [5–7] and discovery of mutations in some non-deleted patients [8 , 9] . The basis of variable phenotypic expression is under intense investigation . Understanding responsible genetic factors upstream and downstream of TBX1 is necessary to test for relevancy as modifiers in human 22q11 . 2DS patients . We are taking mouse genetic approaches to identify genes and networks that may act as modifiers . Tbx1 heterozygous mice have mild aortic arch anomalies or ventricular septal defects , at reduced penetrance , while all homozygous null mutant mice die at birth and have a persistent truncus arteriosus ( PTA ) , which is the most serious heart defect that occurs in 22q11 . 2DS patients [5–7] . In mammals , Tbx1 is expressed strongly in the embryonic pharyngeal apparatus , but not the heart tube itself suggesting that its critical functions are in this tissue [4] . In the early vertebrate embryo , the heart forms as a bilateral cardiac crescent of mesodermal cells , termed the first heart field that fuses to form the primitive heart tube [10 , 11] . Additional mesodermal cells derived from the pharyngeal apparatus , referred to as the second heart field ( SHF ) migrates and helps to expand the heart tube in both directions [12] [13] [13–16] . These cells remain in a progenitor state , allowing them to migrate and build the length of the heart tube , where they differentiate into smooth and cardiac muscle and endothelial cells [17 , 18] . The SHF itself , can be further subdivided to the anterior heart field ( AHF or anterior SHF ) forming the cardiac OFT and right ventricle as well as the posterior SHF forming the inflow tract , respectively , based upon gene expression and cell lineage studies [19–21] . Of interest , Tbx1 is strongly expressed in the pharyngeal mesoderm , including the AHF , but it is not noticeably expressed in the posterior SHF or heart tube [22–24] . One of the key functions of AHF cells is to maintain a progenitor cell state and to prevent premature differentiation . [25] Gene expression profiling of the AHF , within pharyngeal arches two to six , in Tbx1-/- embryos versus wild type littermates [24] and embryonic stem cell lineage studies [22] , suggest that Tbx1 serves to restrict premature differentiation of the pharyngeal mesoderm , so as to allow the OFT to elongate properly [25] . However , the tissue specificity and key molecular mechanisms are not well defined . The basis for premature differentiation in the AHF in Tbx1 mutant embryos is unknown . Major signaling pathways likely have a role in this process . The canonical Wnt signaling pathway is mediated by β-catenin , which has critical functions in most aspects of embryonic development . There are multiphasic functions of Wnt/β-catenin in the pharyngeal mesoderm required for heart development [26] . Several years ago , it was shown that canonical Wnt/β-catenin has a major role in the AHF in forming the cardiac OFT [27] . Further , one study showed that increased or decreased Wnt/β-catenin in the pharyngeal mesenchyme ( DermoCre ) resulted in a decrease or increase in Tbx1 expression , implicating antagonistic functions upstream of Tbx1 [28] . However , genetic interaction studies were not explored nor were gene expression profiling performed to understand possible molecular connections . Such studies would provide possible modifier genes to investigate in human 22q11 . 2DS to understand its variable expressivity . In this report we performed genetic rescue experiments between Tbx1 and β-catenin in the AHF , using mouse models .
Wnt/β-catenin and Tbx1 are expressed in the opposite domains of the SHF , with Tbx1 higher in the AHF and Wnt/β-catenin higher in the posterior SHF , as denoted by Wnt2 and Mef2c-AHF-Cre [18] lineage compared to canonical Wnt signaling ( Fig 1A–1E ) . We were interested in further exploring the function of β-catenin when completely diminished ( Mef2c-AHF-Cre/+;β-cateninflox/flox , referred to as β-cat LOF [29] ) or constitutively active ( Mef2c-AHF-Cre/+;β-cateninEx3/+ [30] , referred to as β-cat GOF ) in the AHF . To identify downstream genes affected by these changes , gene expression profiling was performed on the distal pharyngeal apparatus containing the AHF micro-dissected from β-cat LOF and β-cat GOF embryos at E9 . 5 ( Fig 1F–1H ) . Note that the dissection of the AHF did not include the heart tube . In order to highlight the genes with the greatest fold change , we created a dot plot of log2 fold changes ( Fig 1I ) . Loss of both β-catenin alleles in the Mef2c-AHF-Cre domain resulted in strongly reduced expression of muscle structural genes in the AHF , while constitutive activation of β-catenin in this domain , had the opposite effect and caused a strong increase in expression of the same genes in the AHF ( Fig 1I ) . This increase was strikingly similar to that in the AHF of global Tbx1 null mutant embryos that were previously reported [22 , 24 , 31] . We then examined cardiac phenotypes upon inactivation of Tbx1 in the Mef2c-AHF-Cre lineage to determine whether Tbx1 had a specific role in the AHF . To determine a specific role of Tbx1 in the Mef2c-AHF-Cre domain , we generated two different genotypes , Mef2c-AHF-Cre/+;Tbx1f/- and Mef2c-AHF-Cre/+;Tbx1f/f . Embryos at E14 . 5 with both genotypes had a persistent truncus arteriosus ( PTA ) with complete penetrance ( n = 50; Fig 2A ) . Most but not all had an accompanying ventricular septal defect ( VSD; n = 30; Fig 2B and 2C ) , in contrast to Tbx1-/- mutant embryos , which all has a PTA with a VSD . The PTA was observed as early as E12 . 5 ( S1 Fig ) . Due to the similarity in phenotype ( Fig 2A ) , the two genotypes were combined and further referred to as Tbx1 LOF . In Tbx1 LOF embryos at E9 . 5 , the pharyngeal apparatus and individual arches within appeared grossly normal ( Fig 2F–2G ) compared to control littermates ( Fig 2D and 2E ) . This is distinctly different as compared to the global Tbx1-/- null mutant embryos or mesoderm specific Tbx1 conditional loss of function embryos at this stage [32] that have a severely hypoplastic distal pharyngeal apparatus . This rules out extreme morphology defects , such as absence of neural crest cell populations , as being responsible for the presence of a PTA in Tbx1 LOF embryos . We also performed lineage tracing ( Fig 2D–2G ) and observed that the Mef2c-AHF-Cre lineage in the AHF was only slightly reduced in Tbx1 LOF embryos versus control littermates at E9 . 5 ( Fig 2H ) . By in situ hybridization analysis , Tbx1 expression was greatly reduced in Tbx1 LOF embryos ( S1 Fig ) and this was confirmed by qRT-PCR ( Fig 2I ) . Cell proliferation and apoptosis in the Tbx1 LOF versus control embryos did not show any significant difference in the Mef2c-AHF-Cre lineage in the AHF region between groups at E9 . 5 ( S2 and S3 Figs ) . This is different than what was previously found for Tbx1-/- [33] or Nkx2-5Cre [22] conditional mutant embryos , which have changes in proliferation and apoptosis . We suggest the improved appearance of the distal pharyngeal apparatus in Tbx1 LOF embryos is due to differences in the Mef2c-AHF-Cre recombination domain . In relation to β-catenin , we noted a slight decrease in Tbx1 expression in β-cat GOF embryos ( S4 Fig ) . This was consistent , although not as dramatic , as was found previously using a broader mesenchymal Cre driver ( DermoCre ) [28] . We found β-catenin mRNA is significantly increased in expression in the AHF of Tbx1 LOF embryos by qRT-PCR ( Fig 2I ) . As for β-catenin gain of function in the AHF , we were interested in determining the function of loss and gain of Tbx1 in the AHF . We previously generated a tissue specific constitutively expressing Tbx1 gene [34] . Homozygous mice were crossed with Mef2c-AHF-Cre mice to overexpress Tbx1 in the same domain as other alleles , and the embryos are referred to as Tbx1 GOF . Gene expression profiling of Tbx1 LOF and GOF embryos was performed of AHF tissue at E9 . 5 , to test whether loss or gain of Tbx1 would have opposing effects on muscle structural protein differentiation genes and to compare with findings of β-catenin loss and gain mutant embryos ( Fig 3A–3C ) . The dot plots of global gene expression changes in the AHF between β-cat GOF versus Tbx1 LOF embryos showed increase in gene expression in the same direction ( Fig 3A ) . The genes with the largest increase were the muscle structural protein genes . Similarly , β-cat LOF versus Tbx1 GOF showed the same strong decrease of expression of muscle differentiation genes . The genes with the largest decrease were the muscle structural protein genes . These results provide functional genetic insight as to the previously implicated antagonistic relationship between Tbx1 and β-catenin in the pharyngeal mesenchyme [28] , that they perhaps are needed to balance cell differentiation . Some of the genes were tested by qRT-PCR for Tbx1 LOF and β-cat LOF embryos ( Fig 3C ) . We found top genes decreased in expression in Tbx1 LOF embryos were not generally decreased in β-cat LOF embryos as top genes that were increased in expression . Tbx1 was slightly increased in expression in β-cat LOF embryos ( Fig 3C ) . Based upon the opposing gene expression changes between Tbx1 and β-catenin in the AHF , that also included β-catenin mRNA itself in Tbx1 LOF embryos ( Fig 2I ) , we tested whether we could rescue heart defects in the Tbx1 conditional loss of function mutant embryos by inactivating one allele of β-catenin in the Mef2c-AHF-Cre domain ( Mef2c-AHF-Cre/+;Tbx1f/- and Mef2c-AHF-Cre/+;Tbx1f/f ) . Details of the background and crosses are provided in the Methods section and details of the control genotypes are provided in S1 Table . Inactivation of one allele of β-catenin in the Mef2c-AHF-Cre domain did not result in any cardiovascular defects ( S1 Table and [27] ) . Significant rescue ( p < 0 . 001 , Fisher’s exact test ) was obtained in both sets of double mutant rescue genotype embryos ( Fig 4 ) . Upon combining all double mutant embryos together ( n = 56 ) , a total of 61% ( n = 34/56 ) showed some rescue of the PTA phenotype ( Fig 4 ) . Specifically , complete distal OFT and partial proximal OFT septation and/or complete septation between the ventricles were present in these hearts ( Fig 5A–5D ) . Ten percent showed complete rescue . Additional and more posterior sections can be found in S5 Fig Lineage tracing of the double mutant rescue genotyped embryos showed no significant difference in the number of cells in the AHF , of Mef2c-AHF-Cre lineage compared to the control or Tbx1 LOF embryos at E9 . 5 ( Fig 5E ) . Finally , qRT-PCR was performed and Tbx1 and β-catenin mRNA expression were reduced in the AHF of these embryos compared to the control ( Fig 5F ) . Since we identified the greatest increase of expression in Tbx1 LOF and decrease in β-cat LOF embryos pertaining to muscle differentiation genes , we tested if there is global normalization of expression in the embryos of the double mutant , rescue genotype . For this test , gene expression profiling was performed on these embryos , in the same way for the individual mutant embryos and we found this to be the case . Expression of genes with greatest increase in Tbx1 LOF embryos ( >1 . 3 fold ) , primarily the differentiation genes and greatest decrease in β-cat LOF embryos were largely normalized in rescued embryos ( Fig 6A ) . However , we did not observe this same strong finding for genes increased in β-cat LOF embryos . As mentioned , most of the genes with the strongest increase of expression in Tbx1 LOF and decrease in β-cat LOF embryos that were normalized ( p<0 . 01 ) in rescued embryos , were genes that encode smooth or cardiac muscle genes ( Figs 3C and 6B ) . This also included major transcription factors such as Pitx2 , Tbx5 , Gata4 and Gata6 , that are required for cardiac muscle differentiation [35–37] . The canonical Wnt gene , Wnt2 , showed a similar pattern ( Figs 3C and 6B ) . A full heatmap of the experiment is shown in S6 Fig Some additional genes of note , increased in expression in Tbx1 LOF embryos by gene profiling include Myocd , Bmp2 , Bmp10 , Erbb4 and Sfrp5 , which were oppositely affected in β-cat LOF embryos , and normalized in rescued embryos ( Fig 6B and S6 Fig ) . This data supports the idea that pro-differentiation by canonical Wnt/β-catenin might be modulated by Tbx1 in the Mef2c-AHF-Cre lineage . Not all genes with increase in expression in Tbx1 LOF embryos and decrease in β-cat LOF showed normalization in rescued embryos ( Hand1 , Zfpm2 , Smarcd3 and Tbx20; Fig 6B and S7 Fig ) . Further , genes reduced in expression in both types of LOF mutants ( S7 Fig ) might not be relevant for the observed rescued phenotype since they were not normalized in rescue genotyped embryos . This suggests that other pathways are required for OFT development , and explains , in part , why complete rescue did not occur . Nonetheless it provides insights as to the nature of the relationship of the two genes , Tbx1 and β-catenin as well as their independent functions .
Loss of β-catenin using various pharyngeal mesoderm engineered Cre drivers , including Mesp1-Cre [38] , Nkx2-5-Cre [39] , Isl1-Cre [40 , 41] and Mef2c-AHF-Cre [27] results in embryonic lethality due to the presence of cardiac outflow tract defects . The mechanisms mediating these abnormalities , in particularly within the pharyngeal mesoderm of the AHF , have not been well defined . This is especially important because there are many divergent and distinctive functions of β-catenin during cardiac development [42–45] [45] [38 , 46 , 47] . Our interest was to follow up on a previous study in which Tbx1 expression was affected oppositely by loss or constitutively active β-catenin in the pharyngeal mesenchyme using a mesenchymal Cre , termed DermoCre [28] . Based upon this finding , we investigated the two genes in the AHF tissue at stage E9 . 5 , when the heart tube is elongating . We found that β-catenin promotes muscle differentiation in the AHF . We also found that Tbx1 and Wnt/β-catenin act antagonistically to provide a balance of expression of pro-differentiation genes in the AHF that may be required for cardiac outflow tract development . This sheds new light onto the importance of the two genes in heart development as outlined in the model shown in Fig 7 . In the model in Fig 7 , we illustrate the Tbx1 expression domain in the SHF as a triangle , with the strongest expression anteriorly , in the AHF tissue and weakest in the posterior SHF . On the other hand , Wnt/β-catenin expression and function is strongest in the posterior SHF and weakest in the AHF , at E9 . 5 during mouse embryogenesis . In the panel on the left , we created a simple negative feedback loop , which is consistent with our findings in this study and previous findings using DermoCre [28] . The center panel of the model illustrates the situation when Tbx1 is inactivated or β-catenin is constitutively active in the AHF . Here , in these embryos differentiation occurs prematurely in the AHF , prior to reaching the heart tube ( Fig 7 , middle panel ) . This results in impaired cardiac outflow tract development . In our study , we found that loss of Tbx1 in the Mef2c-AHF-Cre domain , along with loss of one allele of β-catenin provided significant rescue of heart defects ( Fig 7 , right panel ) . This supports the importance of their interaction . However , rescue is not complete . An explanation for this is that β-catenin is only partially diminished when one allele is inactivated , such that complete normalization is not possible . Another explanation is that both Tbx1 and Wnt/β-catenin act in many complex pathways in the AHF at this time point ( E9 . 5 ) , for which only the overlapping functions were normalized in rescued embryos [48–51] . Further , genes changing in expression at E9 . 5 may only be partially reflected in PTA defects observed at E12 . 5 . We also note that the defects in the cardiac outflow tract between Tbx1 loss and β-catenin gain in the Mef2c-AHF-Cre domain are different [27] , supporting this idea . In particular , neonatal lethality occurs in Tbx1 LOF embryos due to the presence of a PTA , while gain of β-catenin results in mid-gestational lethality with a short , hypercellular outflow tract . One of the main functions of Tbx1 in the AHF is to maintain a progenitor state and restrict premature differentiation prior to reaching the elongating heart tube [22 , 24] . Supporting this idea , constitutive overexpression of Tbx1 in the Mef2c-AHF-Cre domain results in a decrease in expression of muscle differentiation genes [22] . Based upon our gene expression profiling data , we suggest that Tbx1 may directly or indirectly , repress expression of key transcription factors that regulate this process i . e . , it maintains the AHF cell fate . We suggest that there is a small decrease in AHF cell numbers but more importantly , a change in cell fate . Lack of observable morphology defects in the distal pharyngeal apparatus and lack of significant change in proliferation or apoptosis of the AHF progeny at E9 . 5 support this . Interestingly , in Tbx1 LOF mutant embryos , we found an increase in expression of genes required for cardiomyocyte specification , such as Gata4 , Tbx5 and Smarcd3 ( Baf60c ) [52–55] [56–61] [35 , 62–65] . Tbx5 and Gata4 proteins are co-expressed and physically interact to regulate expression of downstream muscle structural protein genes [52] [53] . The combination of Tbx5 , Gata4 and Smarcd3 are sufficient to differentiate mouse embryonic mesoderm to beating cardiomyocytes [54] . Another intermediate protein is Serum Response Factor ( SRF ) , which directly promotes expression of genes encoding muscle structural proteins that were found increased in Tbx1 null mutant embryos [22] . Inactivation of Tbx1 resulted in expansion of expression of SRF protein but not mRNA [22] . Similarly , we did not find Srf expression levels altered in Tbx1 LOF embryos . Of interest , the above transcription factors may interact with SRF protein to induce differentiation [66] , supporting a continued role of SRF in Tbx1 biology [22] [36 , 67] . Of interest , the Wnt2 , Tbx5 , Tbx20 , Gata4 and Gata6 genes are expressed and have function in the posterior SHF for formation of the inflow tract . It is not yet known if any are directly or indirectly regulated by Tbx1 . An additional role of Tbx1 may be to restrict posterior SHF fate in the AHF so as to maintain the appropriate sub-populations within the SHF for proper heart development . We previously found expression of these posterior SHF genes were greatly expanded in the AHF tissue in Tbx1 global null mutant embryos [24] and were increased in the same tissue in the conditional mutant embryos by qRT-PCR . It was previously found that Wnt2 and Gata6 act in the same genetic pathway in the posterior SHF during heart development and when inactivated cause atrial septal defects among other anomalies [68 , 69] . We observed an increase in expression , but did not identify atrial septal defects . Since we did not observe a severe morphological defect in Tbx1 LOF embryos at E9 . 5 , we suggest that some of these molecular changes will then affect later development . One of the challenges in human genetics is to identify risk factors of complex traits , such as congenital heart disease [70 , 71] . The 22q11 . 2DS , although rare in the general population , offers a relatively homogenous cohort to investigate the basis of variable phenotypic heterogeneity among affected individuals . Rare deleterious DNA variants altered in expression in Tbx1 LOF embryos and acting antagonistically to canonical Wnt/β-catenin , might act as genetic modifiers of CHD in 22q11 . 2DS . Examination of whole exome sequence of 22q11 . 2DS subjects [72] is underway with a larger cohort , to identify such variants connected to Tbx1 and Wnt/β-catenin gene networks or pathways needed to provide proper balance of critical cell fate choices . The work here provides a basis in the near future to translate efforts to studies of human subjects . In this study , we showed that inactivation of Tbx1 in the AHF using Mef2c-AHF-Cre allele , results in a PTA that is also observed in the most seriously affected 22q11 . 2DS ( velo-cardio-facial/DiGeorge syndrome ) patients . The PTA defect in Tbx1 conditional loss of function mutant embryos , was partially , but significantly rescued by decreasing one allele of the β-catenin gene in this domain , and this also resulted in a normalization of gene expression changes specifically for muscle differentiation but not necessarily for other classes of genes . Thus , we conclude that Tbx1 in the Mef2c-AHF-Cre domain acts antagonistically with Wnt/β-catenin in the SHF to modulate differentiation prior to entering the heart tube .
The following mouse mutant alleles used in this study have been previously described: Tbx1+/- [7] , Tbx1f/+ ( flox = f ) [73] , Tbx1-GFP [34] , β-cateninf/+ and β-cateninE3/+ [29 , 30] , Mef2c-AHF-Cre/+ [18] , ROSA26-GFPf/+ ( RCE:loxP ) [74] and Wnt/β-catenin signaling reporter mice ( Tg ( TCF/Lef1-HIST1H2BB/EGFP ) 61Hadj/J; TCF/Lef:H2B-GFP [75] . To generate Mef2c-AHF-Cre/+;Tbx1f/- mutant embryos ( Tbx1 LOF ) , Mef2c-AHF-Cre/+ transgenic male mice were crossed to Tbx1+/- mice to obtain male Mef2c-AHF-Cre/+;Tbx1+/- mice that were then crossed with Tbx1f/f females . Alternatively , to generate Mef2c-AHF-Cre/+;Tbx1f/f mutant embryos , Mef2c-AHF-Cre/+ transgenic male mice were crossed to Tbx1f/f mice to obtain male Mef2c-AHF-Cre/+;Tbx1f/+ mice , and these were then crossed with Tbx1f/f females . Wild type and Me2fc-AHF-Cre/+;Tbx1f/+ littermates were used as controls for the experiments ( First Tbx1 LOF and rescue crosses , S1 Table ) . Tbx1 gain of function embryos ( Tbx1 GOF ) were generated by crossing male Mef2c-AHF-Cre/+ mice with Tbx1-GFPf/f females . To generate Mef2c-AHF-Cre/+;β-cateninf/f mutant embryos ( β-cat LOF ) , male Mef2c-AHF-Cre/+ transgenic mice were crossed to β-cateninf/f females to obtain male Mef2c-AHF-Cre/+;β-cateninf/+ mice that were then crossed with β-cateninf/f females . β-catenin gain of function ( β-cat GOF ) embryos i . e , Mef2c-AHF-Cre/+;β-cateninE3/+ or male Mef2c-AHF-Cre/+ transgenic mice were crossed to β-cateninE3/E3 females . Double mutant embryos were generated by addition of one copy of the β-cateninflox allele to Tbx1 LOF embryos resulting in what we denote as rescue genotyped embryos . In this case , the females used for the experimental crosses were of the Tbx1f/f;β-catenin f/f genotype , which have been maintained as an inbred line deriving from a mixed C57Bl/6; Swiss Webster background . The reporter ROSA26-GFP f/+ allele was added to the Tbx1f/f and Tbx1f/f;β-catenin f/f lines when visualizing Mef2c-AHF-Cre lineage . To evaluate Wnt/β-catenin signaling in wild type embryos , TCF/Lef:H2B-GFP/+ reporter mice were used . The Mef2c-AHF-Cre/+;Tbx1f/+ and the Mef2c-AHF-Cre/+;Tbx1+/- mice are congenic in Swiss Webster . The Tbx1f/f;β-cateninf/f mice are in an inbred line , as above . The Mef2c-AHF-Cre/+;Tbx1+/- x Tbx1f/f; β-cateninf/f crosses were performed 2 years before the Mef2c-AHF-Cre/+;Tbx1f/+ x Tbx1f/f; β-cateninf/f crosses . The Tbx1f/f and ROSA26-GFPf/+ lines are congenic in Swiss Webster . The β-cateninE3/+ and β-cateninf/f mice were in a mixed C57Bl/6; Swiss Webster background . To exclude the possibility that a strain background might affect the possible rescue by β-catenin LOF allele , half of both Tbx1 LOF and the rescue genotyped embryos were generated by using Tbx1f/+;β-cateninf/+ females ( second Tbx1 LOF and rescue crosses; S1 Table ) . Here , both Tbx1 LOF and the rescue genotyped embryos were littermates . The PCR strategies for mouse genotyping have been described in the original reports and are available upon request . All experiments including mice were carried out according to regulatory standards defined by the NIH and the Institute for Animal Studies , Albert Einstein College of Medicine ( https://www . einstein . yu . edu/administration/animal-studies/ ) , IACUC protocol # 2013–0405 . Institutional Animal Care and Use Committee ( IACUC ) approved this research . The IACUC number is 20160507 . Whole-mount RNA in situ hybridization with non-radioactive probes was performed as previously described [76 , 77] , using PCR-based probes , Tbx1 [78] , Wnt2 forward primer: 5’ TGGCTCTGGCTCCCTCTGCT 3’ and reverse primer: 5’ CAGGGAGCCTGCCTCTCGGT 3’ and Wnt4 forward primer: 5’ CCGCGAGCAATTGGCTGTACC 3’ and reverse primer: 5’ TGGAACCTGCAGCCACAGCG 3’ . Following whole-mount protocol , the embryos were fixed overnight in 4% paraformaldehyde ( PFA ) and then dehydrated through a graded ethanol series , embedded in paraffin and sectioned at 10 μm . Minimum of 5 embryos from 3 independent litters were analyzed per embryonic stage . After fixation as described above , frozen sections were obtained at a thickness of 10 μm and then permeabilized in 0 . 5% Triton X-100 for 5 min . Blocking was performed with 5% serum ( goat or donkey ) in PBS/0 . 1% Triton X-100 ( PBT ) for 1 hour . Primary antibody was diluted in blocking solution ( 1:500 ) and incubated for 1 hour . Proliferation of cells was assessed by immunofluorescence using the primary antibody anti-phospho Histone H3 ( Ser10 ) , a mitosis marker ( 06–570 Millipore ) . Sections were washed in PBT and incubated with a secondary antibody for 1 hour . Secondary antibody was Alexa Fluor 568 goat a-rabbit IgG ( A11011 Invitrogen ) at 1:500 . Slides were mounted in hard-set mounting medium with DAPI ( Vector Labs H-1500 ) . Images were captured using a Zeiss Axio Observer microscope . To perform statistical analysis of cell proliferation , we first counted the Mef2c-AHF-Cre , GFP positive cells in the pharyngeal apparatus located behind the heart in embryo sections and then calculated the average cell counts per tissue section for each embryo . Then we counted all proliferating cells in each section and calculated the ratio of proliferating cells within the Mef2c-AHF-Cre lineage . Then , we estimated the mean and standard error of the average cell counts for controls , Tbx1 LOF and rescued embryos and compared them using the t-test . Apoptosis was assessed on 10 μm thick frozen sections by using TMR Red In situ Cell Death kit ( 2156792 Roche ) following the manufacturer’s instructions . Natural GFP from the reporter or an antibody for GFP ( Abcam 6290 ) was used to distinguish the AHF cells in both assays described above . Representations of the complete AHF region from at least 4 embryos per genotype from at least 3 independent litters were used in each assay . Wnt/β-catenin signaling reporter mice , TCF/Lef:H2B-GFP [75] were used to observe Wnt/β-catenin signaling by direct fluorescence of green fluorescent protein ( GFP ) in wild type embryos at embryonic day E9 . 5 ( 19–21 somite pairs ) . Mouse embryos were fixed and cryosectioned at 10 μm . Slides were mounted in hard-set mounting medium with DAPI to visualize DNA ( Vector Labs H-1500 ) . Images were then captured using a Zeiss Axio Observer microscope . Nuclear Wnt/β-catenin signaling was counted as the GFP positive signal that co-localized with the DNA . A minimum of 5 embryos from 3 independent litters was analyzed . Mouse embryos were isolated in phosphate-buffered saline ( PBS ) and fixed in 10% neutral buffered formalin ( Sigma Corp . ) overnight . Following fixation , the embryos were dehydrated through a graded ethanol series , embedded in paraffin and sectioned at 5 μm . All histological sections were stained with hematoxylin and eosin using standard protocols . Staining was performed in the Einstein Histopathology Core Facility ( http://www . einstein . yu . edu/histopathology/page . aspx ) . For Tbx1 LOF mutants , a total of 70 hearts at E14 . 5 were obtained from more than 50 independent crosses and analyzed morphologically using light microscopy . For the rescue crosses , 56 hearts at E14 . 5 were obtained and the Fisher’s exact test was performed to compare the proportion of rescued phenotypes observed between rescued genotype hearts and the Tbx1 LOF mutants . Images were generated from GFP expressing embryos by direct fluorescence immediately following dissection . For tissue sections , embryos were fixed for 2 hours with embryos stage ≤ E10 . 5 ( 30–32 somite pairs ) . Fixation was carried out in 4% PFA in PBS at 4°C . After fixation , tissue was washed in PBS and then cryoprotected in 30% sucrose in PBS overnight at 4°C . Embryos were embedded in OCT and cryosectioned at 10 μm . Images were then captured using a Zeiss Axio Observer microscope . Embryos at E9 . 5 ( 19–21 somites pairs ) were used for global gene expression studies . To obtain enough RNA for microarray hybridization experiments , microdissected AHFs ( defined here as: pharyngeal arches 2–6 ) from 27 of each of the following genotypes: Tbx1 LOF and its control ( Tbx1f/+ ) , Tbx1 GOF and its control ( Tbx1-GFP/+ ) , β-cat LOF and its control ( β-cateninf/+ ) , β-cat GOF and its control ( β-catE3/+ ) , rescue and its control ( Tbx1f/+;β-cateninf/+ ) , were pooled in groups of three or six according to the genotype . For this experiment we used controls that did not have Cre . Between 4–6 microarrays were performed per genotype in 2–3 batches . The tissue was homogenized in Buffer RLT ( QIAGEN ) . Total RNA was isolated with the RNeasy Micro Kit according to the manufacturer’s protocol . Quality and quantity of total RNA were determined using an Agilent 2100 Bioanalyzer ( Agilent ) and an ND-1000 Spectrophotometer ( NanoDrop ) , respectively . Biotinylated single-stranded cDNA targets were amplified from 100 nanograms ( ng ) starting total RNA using the Ovation RNA Amplification System V2 and FLOvation cDNA Biotin Module V2 ( NuGEN ) . A total of 3 . 75 mg of cDNA was hybridized to the GeneChip Test3 array ( Affymetrix ) to test the quality of the labeled target . Nucleic acid samples that passed quality control were then hybridized to the Affymetrix Mouse GeneST 1 . 0 chip . Hybridization , washing , staining and scanning were performed in the Genomics Core at Einstein ( https://www . einstein . yu . edu/research/shared-facilities/cores/46/genomics/ ) according to the Affymetrix manual . Data analysis was performed in the R statistical package . GeneChip data were pre-processed by the ‘oligo’ package [79] , which implements Robust Multichip Average ( RMA ) algorithm with background correction , quantile normalization and gene level summarization [80] . Afterwards , for convenience of comparison , only probe-set assigned to genes were kept for subsequent analysis . Multiple probe-sets for the same genes were collapsed by “average” to obtain a single measurement per gene [81] . As some arrays were assayed in different batches , we performed UPGMA ( unweighted pair group method with arithmetic mean ) clustering of samples by transcriptomic profile similarities based on the Spearman correlation coefficients . This analysis indicated clear batch effects , especially for β-cat LOF and Tbx1 LOF data ( data not shown ) . Hence , we applied ComBat , an efficient batch effect removal approach , to remove batch effects [82] . This analysis detected some individual arrays of poor quality that were then excluded . In the end , to keep a balance between controls and mutants , we analyzed 4 arrays per genotype . The ‘Limma’ package was used for determining differential expression [83] . To address the issue that adjustment of batch effect by any linear model based approach ( including ComBat ) can introduce a systematic correlation structure in the data , which may lead to exaggerated confidence in differential expression analysis [84] , we accounted for this correlation in Limma by adding ‘blocking for batch’ in the model . In the end , genes with p-values < 0 . 05 were further explored . The microarray data has been deposited to the GEO database ( accession number: GSE78125 ) . Embryos at E9 . 5 ( 19–21 somites pairs ) were used for quantitative gene expression studies of microdissected AHFs from each of the following genotypes: Tbx1 LOF and its control ( Tbx1f/+ ) , Tbx1 GOF , β-cat LOF , β-cat GOF and rescue were pooled in groups according to genotype . Tbx1f/+ was used as control . To obtain enough total RNA and minimize the variability of gene expression in individual embryos , each biological replicate of RNA contained microdissected AHFs from six embryos of the same genotype at E9 . 5 collected from at least 3 independent litters . Three biological replicates were performed per genotype . The tissue was immediately frozen , samples were homogenized and total RNA was isolated with the RNeasy Micro Kit ( Qiagen ) . Quality and quantity of total RNA was determined using an Agilent 2100 Bioanalyzer ( Agilent ) and a ND-1000 Spectrophotometer ( NanoDrop ) , respectively . Single-stranded cDNA targets were amplified from 100 nanograms ( ng ) starting total RNA using the Ovation RNA Amplification System V2 and FL- Ovation cDNA Biotin Module V2 ( NuGEN ) . The mRNA levels were measured using TaqMan Gene Expression assays ( Applied Biosystems ) for each gene and were carried out in triplicate using 18S ( RNA , 18S ribosomal 1 ) , Actb ( Actin , beta ) and B2m ( Beta-2-microglobulin ) genes as normalization controls . TaqMan probes and primer sets were obtained from the Applied Biosystems Gene Expression Assay database ( http://allgenes . com ) . Samples were processed in standard 96-well plates ( 20 ul final volume per reaction and each reaction in triplicate containing 25 ng of cDNA ) on an ABI 7900HT Q-PCR apparatus . The SDS 2 . 2 software platform ( Applied Biosystems ) was used for the computer interface with the ABI 7900HT PCR System to generate normalized data , compare samples , and calculate the relative quantity . Statistical significance of the difference in gene expression was estimated using ANOVA and the two-tailed t-test independently when type of comparison allowed it . http://www . omim . org http://genome . ucsc . edu/ http://www . R-project . org | To understand the genetic relationship between Tbx1 and canonical Wnt/β-catenin , we performed gene expression profiling and genetic rescue experiments . We found that Tbx1 and β-catenin may provide a negative feedback loop to restrict premature differentiation in the anterior heart field . This is relevant to understanding the basis of variable expressivity of 22q11 . 2DS , caused by haploinsufficiency of TBX1 . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"cardiovascular",
"anatomy",
"cardiac",
"ventricles",
"ventricular",
"septal",
"defects",
"animal",
"models",
"developmental",
"biology",
"model",
"organisms",
"experimental",
"organism",
"systems",
"embryos",
"morphogenesis",
"car... | 2017 | Reduced dosage of β-catenin provides significant rescue of cardiac outflow tract anomalies in a Tbx1 conditional null mouse model of 22q11.2 deletion syndrome |
Mutations in Pten-induced kinase 1 ( PINK1 ) are linked to early-onset familial Parkinson's disease ( FPD ) . PINK1 has previously been implicated in mitochondrial fission/fusion dynamics , quality control , and electron transport chain function . However , it is not clear how these processes are interconnected and whether they are sufficient to explain all aspects of PINK1 pathogenesis . Here we show that PINK1 also controls mitochondrial motility . In Drosophila , downregulation of dMiro or other components of the mitochondrial transport machinery rescued dPINK1 mutant phenotypes in the muscle and dopaminergic ( DA ) neurons , whereas dMiro overexpression alone caused DA neuron loss . dMiro protein level was increased in dPINK1 mutant but decreased in dPINK1 or dParkin overexpression conditions . In Drosophila larval motor neurons , overexpression of dPINK1 inhibited axonal mitochondria transport in both anterograde and retrograde directions , whereas dPINK1 knockdown promoted anterograde transport . In HeLa cells , overexpressed hPINK1 worked together with hParkin , another FPD gene , to regulate the ubiquitination and degradation of hMiro1 and hMiro2 , apparently in a Ser-156 phosphorylation-independent manner . Also in HeLa cells , loss of hMiro promoted the perinuclear clustering of mitochondria and facilitated autophagy of damaged mitochondria , effects previously associated with activation of the PINK1/Parkin pathway . These newly identified functions of PINK1/Parkin and Miro in mitochondrial transport and mitophagy contribute to our understanding of the complex interplays in mitochondrial quality control that are critically involved in PD pathogenesis , and they may explain the peripheral neuropathy symptoms seen in some PD patients carrying particular PINK1 or Parkin mutations . Moreover , the different effects of loss of PINK1 function on Miro protein level in Drosophila and mouse cells may offer one explanation of the distinct phenotypic manifestations of PINK1 mutants in these two species .
PD is a neurodegenerative disorder characterized by the dysfunction and loss of dopaminergic ( DA ) neurons in the substantia nigra , although neurons in other brain regions are affected as well . Mutations in PINK1 and Parkin are linked to familial forms of early-onset PD [1] , [2] . PINK1 encodes a Ser/Thr kinase with a mitochondrial targeting sequence , whereas Parkin encodes an E3 ubiquitin ligase . Studies in Drosophila first revealed that PINK1 and Parkin act in a common pathway to impact mitochondrial function and DA neuron maintenance [3]–[6] , in part through the regulation of mitochondrial fission/fusion dynamics [7]–[11] . At least in primary cultured mammalian hippocampal neurons and DA neurons , PINK1 and Parkin have been shown to exert similar effects on mitochondrial dynamics as seen in Drosophila DA neurons [7] , [12] . PINK1 and Parkin are also implicated in mitochondrial quality control [13] . Decreased mitochondrial membrane potential stabilizes the normally labile PINK1 , which recruits Parkin to damaged mitochondria , leading to ubiquitination of mitochondrial proteins and marking damaged mitochondria for removal by autophagy [14] . Both mitochondrial fission/fusion dynamics and autophagy are considered important aspects of the mitochondrial quality control mechanism that mediates PINK1/Parkin function in DA neuron maintenance [11] , [15]–[17] . In some PINK1- or Parkin-linked PD patients , symptoms of peripheral neuropathy were also reported [18]–[20] . It is not clear whether this is caused by defects in the aforementioned functions or some other unknown function of PINK1/Parkin . Peripheral neuropathy is a clinical term used to describe various forms of damages to nerves of the peripheral nervous system by distinct mechanisms [21] . Many types of peripheral neuropathy are dependent on the length of neuronal axon , with neurons carrying long axons frequently affected . It is hypothesized that this is caused by defects in the axonal transport of key proteins and/or organelles such as mitochondria , which are critical for maintaining the axonal and synaptic physiology of those extremely polarized neurons [22] . This notion has gained significant support from recent studies of the inherited forms of peripheral neuropathies [22] , [23] . Defective mitochondrial transport has also been considered a pathogenic event in other neurodegenerative diseases [22] , [24] , [25] , including rodent models of PD [26] , [27] . In primary cultured rat hippocampal neurons , overexpression of PINK1 has been shown to inhibit the lateral movement of photoactivated , mito-Dendra2-labelled mitochondria [12] , raising the possibility that defects in the axonal transport of mitochondria may actively participate in PINK1-related PD pathogenesis . A major aspect of axonal transport is mediated by motor proteins that travel on axonal microtubules , which are polarized and uniformly orientated , with their plus-ends pointing towards nerve terminals . The kinesin and dynein motors are involved in microtubule plus-end ( anterograde ) and minus-end directed ( retrograde ) transport , respectively [28] . Mitochondria are mainly produced in neuronal cell body and delivered to sites where metabolic demand is high , such as the synapses and nodes of Ranvier [29] . The functions of Mitochondrial Rho ( Miro ) , a mitochondrial outer membrane GTPase [30] , and the cytosolic protein Milton are critical for mitochondrial transport , as they serve to link mitochondria with kinesin motors and the microtubule cytoskeleton [31] , [32] . In Drosophila Miro or Milton mutants , mitochondria accumulate in neuronal soma and fail to move into the axons [31] , [32] . In cultured mammalian cells , overexpression of a constitutively active mutant of Miro was shown to induce cell death , suggesting that mitochondrial transport or some other aspect of Miro function is important for cell survival [30] . Whether this is relevant to in vivo conditions such as neurodegenerative disease settings is not known . The fruit fly Drosophila melanogaster has served as an excellent model for studying neurodegenerative diseases [10] . It was in Drosophila that the in vivo function of PINK1 was first revealed [3]–[6] . PINK1 mutant flies exhibit abnormal wing postures , reduced flight ability and thoracic ATP level , degeneration of indirect flight muscle and DA neurons , and male sterility , which are caused by the accumulation of dysfunctional mitochondria , thus suggesting a role of PINK1 in mitochondrial function and/or quality control [3]–[6] . Further genetic studies in Drosophila have also uncovered important functions of PINK1 in regulating mitochondrial morphology and electron transport chain activity [7]–[9] , [33] . The power of the Drosophila neurodegenerative disease models lies in the ability to facilitate unbiased genetic modifier screens to identify new players involved in the disease process . Using this approach , we show in this study that PINK1 genetically interacts with the mitochondrial transport machinery . Reduction of function in Miro , Milton , or kinesin heavy chain effectively rescued the PINK1 mutant phenotypes . On the other hand , overexpression ( OE ) of Miro led to the formation of enlarged mitochondria and resulted in DA neuron loss , thus phenocopying PINK1 mutants . By monitoring mitochondrial movement in live Drosophila larval motor neurons , which possesses long axons and could serve as a model system for studying peripheral neuropathy , we provide evidence that PINK1 directly regulates mitochondrial transport . The function of PINK1 in mitochondrial transport may contribute to PD pathogenesis in DA neurons and underlie the peripheral neuropathy symptoms associated with certain PINK1 mutations in some PD patients . Our biochemical analysis demonstrated that overexpressed PINK1 in cooperation with Parkin could regulate Miro protein ubiquitination and stability , which might contribute to the regulatory effect of PINK1 on mitochondrial motility . While our paper was under review , it was suggested that PINK1 phosphorylates Miro at a conserved S156 residue , and that this phosphorylation event is required to activate proteasomal degradation of Miro in a Parkin-dependent manner [34] . However , our in vitro kinase assay using active recombinant PINK1 failed to show direct phosphorylation of Miro by PINK1 . Moreover , a mutant form of hMiro1 with the S156 site mutated to Ala was equally susceptible to PINK1/Parkin-mediated degradation in HeLa cells . Thus , the exact molecular mechanism by which the PINK1/Parkin pathway regulates Miro protein level will require further investigation .
By taking advantage of the easily identifiable phenotype of abnormal wing posture induced by dPINK1 inactivation , we performed a genetic screen for modifiers of PINK1 . The scheme was similar as described before [35] . In this screen , we identified components of the mitochondrial transport machinery as genetic modifier of PINK1 . Knockdown of Miro , Milton or Kinesin heavy chain ( Khc ) each rescued the muscle phenotypes in PINK1B9 null mutant , including abnormal wing posture , decreased fly ability and ATP depletion ( Figure 1A–1C ) . Conversely , overexpression ( OE ) of Miro and Khc enhanced such phenotypes ( Figure 1A–1C ) . These results demonstrated strong genetic interaction between PINK1 and the mitochondrial transport machinery as a whole , supporting that mitochondrial transport is the underlying mechanism mediating their genetic interaction . In this study , we will focus our analysis on Miro , as a previous study in cultured cells suggested that mammalian Miro might physically interact with PINK1 [36] . To test the relevance of the functional interaction between PINK1 and the mitochondrial transport machinery to PD pathogenesis , we examined their interaction in DA neurons , the disease-relevant cell type . As in the muscle , Miro-RNAi effectively rescued PINK1 mutant phenotypes in DA neurons , both in terms of mitochondrial aggregation ( Figure 2A–2D ) and DA neuron loss ( Figure 2F ) . Moreover , Miro-OE alone , driven by the TH-Gal4 driver , caused aberrant mitochondrial aggregation ( 2E , 2G ) and DA neuron loss ( Figure 2F ) , thus phenocopying PINK1 loss-of-function effects ( Figure 2B , 2F , 2G ) , although the Miro-OE effect was noticeably stronger than PINK1 mutant . It is worth noting that TH-Gal4-driven Miro-OE in PINK1 mutant background resulted in dramatically reduced viability ( data not shown ) , although the surviving adults did not show further DA neuron loss than that induced by Miro-OE alone ( Figure 2F ) . Together , these results demonstrate that PINK1 and Miro also exhibit strong genetic interaction in DA neurons , with decreased dMiro level/activity ameliorating the detrimental effects caused by the loss of dPINK1 , whereas increased dMiro level or activity phenocopying dPINK1 mutants . The strong genetic interaction between PINK1 and Miro raised the interesting possibility that PINK1 might directly regulate mitochondrial transport , the impairment of which might contribute to PINK1-related parkinsonism . This was further supported by the DA neuron loss induced by Miro-OE alone , which presumably acted by altering mitochondrial transport . To test this idea , we examined the effect of PINK1 on mitochondrial movement in Drosophila larval motor neurons , a system amenable to live imaging of mitochondrial transport . Mitochondrially-targeted GFP ( mitoGFP ) expressed specifically in motor neurons was used to track mitochondrial movement via live imaging in anesthetized third instar larvae ( Figure 3A ) , using well-established procedures [37] , [38] . To highlight the mitochondria undergoing active transport , a 61 . 5 µm-long segment of motor neuron was photobleached and the movement of fluorescently labeled mitochondria moving into the bleached area from both directions was recorded at 1 frame/2 s for 300 s ( Figure 3B , Videos S1 , S2 , S3 , S4 , S5 , S6 , S7 ) . From these videos , mitochondrial flux ( the normalized number of mitochondria that passes certain point over time ) , mitochondrial net velocity ( the normalized mitochondrial net displacement over time ) , and mitochondrial morphology ( e . g . mitochondrial length ) in different genetic backgrounds were analyzed . In general , mitochondrial net velocity is controlled primarily by the intrinsic properties and quantities of motor proteins associated with the mitochondria [38] , while mitochondrial flux can also be significantly affected by mitochondrial morphology , as changes in mitochondrial morphological features such as length can increase or decrease the number of motile mitochondria . We found that PINK1-OE decreased mitochondrial flux as well as net velocity in both anterograde and retrograde directions , similar to the effect of Miro-RNAi , although the PINK1-OE effect appeared to be slightly weaker ( Figure 3B–3D; Videos S1 , S2 , S4 ) . In contrast , PINK1-RNAi and Miro-OE both increased the net velocity of anterograde mitochondrial transport , with retrograde transport largely unaffected ( Figure 3B , 3C; Videos S3 , S5 ) . PINK1-RNAi also increased anterograde mitochondrial flux ( Figure 3C ) , while mitochondrial flux in Miro-OE background was reduced in both anterograde and retrograde directions ( Figure 3C ) . The reduction of mitochondrial flux by Miro-OE could be partially explained by the formation of very long mitochondria in Miro-OE motor neurons ( Figure 3E , 3F ) , as previously observed [38] . In addition to mitochondrial motility , PINK1 also affected mitochondrial length in motor neurons . PINK1-RNAi increased mitochondrial length in the axons of larval motor neurons as in Miro-OE case , although the effect of Miro-OE was much stronger . Conversely , PINK1-OE and Miro-RNAi both decreased mitochondrial length ( Figure 3E , 3F ) . To address whether PINK1-induced mitochondrial motility change was due to its effect on mitochondrial length , we examined mitochondrial transport in genetic backgrounds where mitochondrial fusion/fission machinery was directly manipulated to alter mitochondrial length . Increasing mitochondrial fission by overexpression of the fission protein Fis1 or knockdown of the fusion protein Marf led to decreased mitochondrial length ( Figure 3E , 3F ) , similar to the effects of Miro-RNAi or PINK1-OE . However , in contrast to the decreased mitochondrial flux and net velocity as observed in the Miro-RNAi or PINK1-OE backgrounds , Fis1-OE and Marf-RNAi both increased mitochondrial flux and net velocity in anterograde and retrograde directions ( Figure 3B–3D; Videos S6 , S7 ) , suggesting that mitochondrial length and transport kinetics are not always directly correlated , which is consistent with a previous report [38] . Collectively , these results support the notion that PINK1 regulates mitochondrial transport and that its effect on mitochondrial motility is direct , rather than a secondary effect of mitochondrial length change . In addition to mitochondrial motility , we also examined mitochondrial distribution at motor neuron nerve terminals at the larval neuromuscular junction ( NMJ ) , which can be used as an indirect measure of mitochondrial motility . Consistent with a previous report [31] , Miro-OE led to the accumulation of mitochondria in the most distal boutons , which is likely the consequence of net anterograde transport ( Figure 4A , 4B ) . PINK1 knockdown showed similar effect ( Figure 4A , 4B ) . Thus , the dysfunctional mitochondria in PINK1 mutant might gain longer retention time in the distal segment of motor neuron axons where synapses are formed . This finding may have clinical implications for PINK1 pathogenesis . In contrast , Miro-RNAi and PINK1-OE both led to decreased accumulation of mitochondria in the most distal boutons ( Figure 4A , 4B ) . We conclude that PINK1 regulates mitochondrial distribution in motor neuron nerve terminals , likely through its effect on mitochondrial transport . Our results so far showed that PINK1 and Miro exert opposite effects on mitochondrial morphology , motility and distribution . We next explored the biochemical mechanisms underlying their negative genetic relationship . We first used HeLa cells to test whether Miro protein level might be regulated by PINK1 and possibly Parkin , which tends to work together with PINK1 in a common pathway [3]–[5] , [39] . There are two Miro homologues in human cells , hMiro1 and hMiro2 , that are ∼60% identical [30] . Overexpression of either hPINK1 or hParkin did not lead to obvious change of exogenous hMiro1 protein level under normal conditions , but a modest reduction of hMiro1 level was observed when hPINK1 and hParkin were co-expressed ( Figure 5A , lane 5 ) . A decline in mitochondrial membrane-potential induced by the mitochondrial uncoupler carbonyl cyanide m-chlorophenylhydrazone ( CCCP ) was reported to activate the PINK1/Parkin pathway [13] , [14] , [39] . Under CCCP treatment condition , hPINK1 or hParkin each significantly stimulated hMiro1 ubiquitination ( Figure 5A ) . Since HeLa cells express very little endogenous Parkin [13] , the effect of hPINK1 alone on hMiro1 ubiquitination ( Figure 5A , lane 8 ) suggested that other E3 ligase ( s ) might be recruited by hPINK1 to ubiquitinate hMiro1 . However , this ubiquitination event did not appear to lead to destabilization of hMiro1 ( Figure 5A , IB: Myc ) . In contrast , coexpression of hPINK1 and hParkin dramatically reduced hMiro1 level in the presence of CCCP ( Figure 5A , lanes 9 ) . Importantly , pathogenic mutations in hPINK1 or hParkin abolished this effect ( Figure 5B , lanes 4 and 5 compared with lane 3 , and lanes 9–11 compared with lane 8 ) , indicating that functional hPINK1 and hParkin are both required in the destabilization of hMiro1 . Previously , many outer mitochondrial membrane ( OMM ) proteins were shown to be degraded by the ubiquitin proteasome system ( UPS ) pathway in a PINK1/Parkin-dependent reaction at an early step of mitophagy , while other OMM proteins might be eliminated by subsequent autophagosome-dependent events [40] , [41] . Thus , direct or indirect substrates of PINK1/Parkin could be distinguished by their degradation kinetics [40] . In our experiments , hMiro1 was more rapidly degraded than another OMM protein VDAC1 ( Figure 5B , VDAC1 ) , a reported Parkin substrate involved in mitophagy [14] , supporting that hMiro1 is a direct substrate of Parkin . Similar to hMiro1 , hMiro2 could also be ubiquitinated by PINK1 and Parkin co-expression or after CCCP treatment . However , the degradation of hMiro2 was at a much slower rate compared to hMiro1 ( Figure S1 ) , consistent with a previous result [40] . Furthermore , like exogenous hMiro1 , endogenous hMiro1 was also rapidly degraded by PINK1/Parkin overexpression in HeLa cells and its level was dramatically reduced within 15 min of CCCP treatment . The degradation of endogenous hMiro2 was again at a much slower rate than that of hMiro1 ( Figure 5C , 5D ) . We next examined the effect of the PINK1/Parkin pathway on Miro protein level in an in vivo setting . Similar to the results in HeLa cells , Drosophila dMiro protein level was decreased in the brain extracts of PINK1 or Parkin overexpression adult flies ( Figure 5E ) . Conversely , dMiro level was increased in PINK1B9 mutant brain extracts ( Figure 5E ) . These results are consistent with dPINK1 negatively regulating dMiro protein level in vivo . It is worth noting that different from the effects seen in HeLa cells , overexpression of PINK1 or Parkin alone was sufficient to reduce dMiro level in adult Drosophila brain , and the co-expression of PINK1 and Parkin did not lead to much further reduction of dMiro level than PINK1-OE alone , suggesting that the endogenous levels or activities of PINK1 and Parkin are already sufficient to support each other's action in the Drosophila brain . Removal of damaged or dysfunctional mitochondria through mitophagy could be one mechanism by which the PINK1/Parkin pathway maintains mitochondrial health , at least under some conditions , and the accumulation of those abnormal mitochondria in PINK1 mutants could be the underlying cause of disease pathogenesis . Consistent with this notion , it was previously shown that enhancing autophagy could efficiently rescue dPINK1 mutant phenotypes [35] . To better understand the rescuing effect of Miro-RNAi in PINK1 mutant background , we examined the effect of Miro knockdown on mitophagy . We used CCCP treatment to induce mitochondrial damage in HeLa cells stably transfected with venus-Parkin , and subsequently monitored the removal of damaged mitochondria over time by examining the protein levels of mitochondrial markers on the inner/outer membrane or in the matrix and inter-membrane space . Simultaneous knockdown of hMiro1 and hMiro2 significantly accelerated the mitochondrial removal process , with all the mitochondrial markers disappearing faster in hMiro knockdown cells than in the control siRNA-treated cells ( Figure 6A ) . This suggested that there was more active mitophagy after hMiro knockdown . To confirm this result , we monitored the mitochondrial network by immunofluorescence staining . Compared to the control siRNA-treated cells , knockdown of either hMiro1 or hMiro2 led to the accumulation of mitochondria in the perinuclear region , and knockdown of both hMiro1 and hMiro2 further enhanced this effect ( Figure 6B ) . Fluorescence from the immunostaining of Tom20 ( an OMM marker ) but not HSP60 ( a matrix marker ) in hMiro1 and hMiro2 double knockdown cells was noticeably weaker than that in control siRNA treated cells at 3 h after CCCP treatment ( Figure 6C ) , supporting the notion that hMiro knockdown facilitated an early event in Parkin-mediated mitophagy . We further investigated the molecular mechanisms by which the PINK1/Parkin pathway regulates Miro protein level or stability . While our paper was under review , a report showed that PINK1 phosphorylates Miro at a conserved S156 residue , and that this phosphorylation activates proteasomal degradation of Miro in a Parkin-dependent manner [34] . However , repeated in vitro kinase assays using an active GST-dPINK1 recombinant protein capable of efficient autophosphorylation [42] failed to show phosphorylation of GST-dMiroΔTM , a GST fusion protein of full-length dMiro with the transmembrane domain deleted ( Figure 7A ) . Drosophila or mammalian PINK1 protein affinity purified from HEK293 cells by immunoprecipitation also failed to phosphorylate GST-dMiroΔTM in our assays ( data not shown ) . To further probe the significance of S156 phosphorylation in facilitating the proteasomal degradation of Miro promoted by the PINK1/Parkin pathway , we introduced S156A mutations into hMiro1 or hMiro2 and examined the stability of the mutant proteins in HeLa cells co-transfected with PINK1 and Parkin , under normal or CCCP treatment conditions . As shown in Figure 7B , the wild type and S156A mutant forms of hMiro1 were equally susceptible to PINK1/Parkin- mediated degradation under both conditions . The wild type and S156A mutant forms of hMiro2 behaved similarly as well ( data not shown ) . Thus , the PINK1/Parkin pathway may regulate Miro protein level independent of S156 site phosphorylation under these experimental conditions in HeLa cells . We also examined the effect of loss of PINK1 function on Miro protein level in mammalian cells . For this purpose , we used both HeLa cells with PINK1 knockdown and MEF cells derived from PINK1 ( −/− ) knockout mice . Surprisingly , unlike the situation in Drosophila , endogenous Miro1 or Miro2 protein levels were significantly reduced in PINK1 RNAi cells under normal or CCCP treatment conditions ( Figure 7C , PINK1 RNAi in HeLa cells or Venus-Parkin stably transfected HeLa cells ) and in PINK1 ( −/− ) MEF cells ( Figure 7D ) . The introduction of Venus-Parkin resulted in CCCP/PINK1-dependent degradation of Miro1 ( Figure 7C , Control RNAi in Venus-Parkin transfected HeLa cells ) . Thus , loss of PINK1 function in mammalian cells can lead to reduced expression of Miro1 and Miro2 proteins , presumably through mechanisms distinct from that operating under PINK1/Parkin co-overexpression condition .
Mitochondrial dysfunction has long been implicated in the pathogenesis of PD . However , the exact mechanisms by which mitochondrial dysfunction arises in the disease process and how cells , particularly neurons , handle dysfunctional mitochondrial are not well understood . The identification of a mitochondrial quality control system involving two FPD genes , PINK1 and Parkin , has provided a much-needed point of entry to elucidate the role of mitochondria in the pathogenesis of PD . Here we showed that PINK1 directly regulates mitochondrial transport and that it affects the stability and/or activity of Miro , a mitochondrial Rho GTPase with a well-establish function in mitochondrial transport . Our conclusion is supported by the following evidence: 1 ) dMiro protein level is negatively regulated by PINK1 and Parkin in vivo in Drosophila; 2 ) Overexpressed PINK1 and Parkin act together to promote the ubiquitination and degradation of hMiro1 in HeLa cells; 3 ) Reduction of the activities of Miro or other components of the mitochondrial transport machinery effectively rescued dPINK1 mutant phenotypes . 4 ) Overexpression of dMiro in DA neurons phenocopied dPINK1 loss-of-function effects; 5 ) Manipulation of dPINK1 activity produced clear mitochondrial motility phenotypes opposite to that observed for dMiro manipulation in Drosophila larval motor neurons . Together , these results support that the mitochondrial transport defects caused by PINK1 inactivation represent one of the key pathogenic events that contribute to PD pathogenesis in the Drosophila model . Neurons are highly polarized cells that rely heavily on axonal transport to distribute to axons and synapses critical proteins and organelles synthesized in the cell body , thereby maintaining neuronal function and health . Defects in axonal transport are often linked to diseases affecting peripheral neurons that tend to extend very long axons [21] , [22] , [29] . Although the symptoms of PD patients mainly arise from the loss of DA neurons , some PD patients carrying particular PINK1 and Parkin mutations developed peripheral neuropathy with unknown cause [18]–[20] . Our results showing the PINK1/Parkin pathway playing a critical role in regulating mitochondrial transport offers one potential explanation of the peripheral neuropathy symptoms observed in these PINK1 or Parkin-linked PD cases . It would be interesting to examine whether Miro protein level or activity is affected by these particular mutations in human cells . Moreover , we propose that defects in PINK1/Parkin-regulated mitochondrial transport may offer one explanation of the selective vulnerability of DA neurons observed in PD patients and animal models . DA neurons that make elaborate and long projections may be particularly vulnerable to impairment of the mitochondrial transport system . Our results offer new insights into the mode of action of the PINK1/Parkin pathway in mitochondria quality control . We showed that , in Drosophila models , PINK1 OE led to decreased mitochondrial flux and net velocity , as observed in Miro knockdown background . In addition , we found that Miro knockdown could facilitate an early step of mitophagy in mammalian cells . These observations , together with the finding that the normally labile PINK1 protein is stabilized on damaged mitochondria [39] , suggest a scenario whereby the accumulation of PINK1 on damaged mitochondria and the subsequent turnover of Miro could exert neuroprotection by ( 1 ) preventing damaged mitochondria from being anterogradely transported along the axons , thus increasing their chance of getting eliminated in the soma; and ( 2 ) promoting elimination of damaged mitochondria through mitophagy . This potentially explains the normal protective function of PINK1 . When PINK1 function is impaired , however , on one hand mitochondria become dysfunctional as evidenced by morphology changes and impaired electron transport chain function [3]–[5] , [33] , [43]–[45] , on the other hand , the anterograde mitochondrial transport is enhanced as shown in this study in Drosophila models . As a result , the dysfunctional mitochondria would have increased retention in the axons and synapses , resulting in increased reactive oxygen species ( ROS ) production , oxidative damage , and subsequent synaptic and axonal degeneration and eventual neuronal loss , at least in the Drosophila models . Many details of this model await further experimental validation . For example , it has been suggested that the reported effect of PINK1/Parkin on mitochondrial autophagy may not operate in the same manner in primary neurons as compared to cultured non-neuronal cells [46] . It also remains to be determined whether the effects of Miro on mitochondrial transport and mitophagy reflect a functional antagonism between these two processes , or two distinct functions of Miro in neuronal maintenance . In this respect , it is worth noting that the effect of Miro overexpression on cell survival in Drosophila is cell type-dependent: it causes DA neuron loss but has no obvious effect on muscle integrity ( data not shown ) . It is possible that different tissues may have different sensitivities to impairments of Miro function . For example , muscle cells may be less susceptible to mitochondrial transport defects than neurons . One interesting difference we observed between Drosophila and mouse systems was that although activation of the PINK1/Parkin pathway led to reduced Miro protein level in both systems , the loss of PINK1 in Drosophila resulted in increased steady-state Miro protein level , whereas its loss in mammalian cells as in PINK1 ( −/− ) MEF cells or PINK1 RNAi HeLa cells had the opposite effect . The mechanism of Miro downregulation in PINK1 loss-of-function mammalian cells is currently unknown , but it is presumably different from that used by activation of the PINK1/Parkin pathway . Since the upregulation of dMiro in dPINK1 mutant background is likely causal to DA neuron degeneration , as indicated by the rescue of DA neuron loss in dPINK1 mutant by dMiro-RNAi and the induction of DA neuron loss by dMiro-OE alone , it is tempting to speculate that the downregulation of Miro levels in PINK1 ( −/− ) mouse , as opposed to the dMiro upregulation in Drosophila PINK1 mutant , might contribute to the lack of DA neuron degeneration phenotype in the mouse PINK1 models [47]–[50] . Testing this hypothesis will require in vivo studies boosting Miro expression levels in wild type and PINK1 ( −/− ) mouse . Our results also provide new insights into the process by which the PINK1/Parkin pathway promotes mitophagy . Previous studies suggested that upon recruitment to damaged mitochondria , Parkin activates the ubiquitin proteasome system to effect wide-spread degradation of OMM proteins in an autophagy-independent manner , and it was further proposed that this remodeling of OMM is important for a subsequent step of mitophagy [40] . The previously identified Parkin substrates , Mfn1 and Mfn2 , although important for the effect of Parkin on mitochondrial fission/fusion dynamics , are not necessary for Parkin-induced mitophagy [40] , [51] . Here we show that removal of Miro by the PINK1/Parkin pathway , in a presumably autophagy-independent but ubiquitination-dependent manner , facilitated mitophagy . Interestingly , knockdown of mammalian Miro itself promotes the formation of ring-like or round-shaped mitochondrial morphology , which is often observed in depolarized , mitophagy-ready mitochondria ( Figure 6B; [13] ) . It is possible that the removal of Miro from OMM exposes certain recognition signals for the autophagy machinery , or that Miro/Milton/Kinesin-mediated mitochondrial transport may normally antagonize the mitophagy process . Supporting the latter scenario , an interaction between the BECLIN 1-interacting protein AMBRA1 and the dynein motor complex has been implicated in mammalian autophagy [52] . It also remains to be understood at the mechanistic level how PINK1 cooperates with Parkin to promote the ubiquitination and degradation of Miro . One attractive hypothesis is that PINK1 may directly phosphorylate Miro to promote its subsequent ubiquitination and degradation by Parkin , as suggested by a recent study [34] . However , our biochemical data have so far failed to support this hypothesis . It is possible that the divergent results are due to the different cell lines used or other experimental conditions . Alternatively , PINK1 may directly act on Parkin to promote Parkin's mitochondrial recruitment or activity in activating ubiquitin proteasome system-mediated ubiquitination and degradation of Miro . Further studies are needed to elucidate the molecular mechanisms of PINK1/Parkin action . Finally , it is worth mentioning that studies in Drosophila models have identified a number of genetic modifiers of PINK1/Parkin [35] , [53] , [54] . While some of these genetic modifier genes may be directly related to the seemingly diverse biological activities of the PINK1/Parkin pathway , possibly mediated by distinct PINK1/Parkin substrates , others may reflect cellular compensatory responses to cope with the mitochondrial dysfunction caused by PINK1/Parkin inactivation . The fact that manipulations of each of these different cellular processes exert clear functional rescue of PINK1/Parkin mutant phenotypes suggests that there exists a signaling network linking the diverse activities of PINK1/Parkin in mitochondria biology with the nuclear-encoded cellular responses to mitochondrial dysfunction , and that many key players in this network represent novel and rational therapeutic targets .
Flies were raised according to standard procedures at indicated temperatures . Sources of fly strains and other reagents are as follows: dPINK1B9: Dr . J . Chung [3]; UAS-dMiro and anti-dMiro antibody: Dr . K . Zinsmaier [31]; TH-GAL4 , UAS-PINK1 , UAS-PINK1 RNAi and rabbit anti-Drosophila TH antibody: described before [5]; UAS-Miro RNAi106683: Vienna Drosophila RNAi Center; UAS-Miro RNAi27695 , UAS-Milton RNAi28385 and UAS-Khc RNAi25898: Harvard Transgenic RNAi Project ( TRiP ) and Bloomington Drosophila Stock Center; all other fly lines: Bloomington Drosophila Stock Center; FLAG-hParkin mutants and HA-ubiquitin: Drs . N . Matsuda , K . Tanaka and S . Hatakeyama; Myc-hMiro1 and Myc-hMiro2 plasmids: Dr . P . Aspenström [30]; hPINK1 cDNAs were cloned into pcDNA3-FLAG vector . Antibodies used in this study are as follows: anti-RHOT1/Miro1 ( 4H4 , Abnova ) , anti-RHOT2/Miro2 ( Protein technology Group ) , anti-PINK1 ( Novus ) , anti-Parkin ( PRK8 , Santa Cruz Biotechnology ) , anti-Tom20 ( FL-145 , Santa Cruz Biotechnology ) , anti-VDAC1 ( Abcam ) , anti-OXPHOS Complex IV subunit I/COX I ( Invitrogen ) , anti-Tim23 ( BD ) , anti-NDUFA9 ( Invitrogen ) , anti-Cytochrome c ( BD ) , anti-HtrA2/Omi ( as described in [55] ) , anti-Hsp60 ( BD ) , anti-PDHA1 ( Abcam ) , anti-HA ( 3F10 , Roche ) , anti-Myc ( 4A6 , Millipore; #2272 , Cell Signaling Technology ) , anti-β-actin ( AC-15 , Sigma-Aldrich ) , anti-α-tubulin ( DM1A , Millipore ) , Peroxidase anti-Guinea Pig IgG antibody ( Jackson ImmunoResearch ) , Texas Red-conjugated anti-HRP ( Jackson ImmunoResearch ) , Alexa Fluor 488 nm-conjugated goat anti-chicken IgG ( Invitrogen ) and Alexa Fluor 594 nm-conjugated goat anti-rabbit IgG ( Invitrogen ) . These assays were carried out essentially as described before [35] . The thoracic ATP level was measured using a luciferase based bioluminescence assay ( ATP Bioluminescence Assay Kit HS II , Roche applied science ) as described [35] . Whole-mount brain immunohistochemistry for TH and mitoGFP was performed as described previously [35] . For DA neuron mitochondrial morphology analysis , mitoGFP was expressed in Drosophila DA neurons using the TH-Gal4 driver . Brains from 3-day-old adult flies of the indicated genotypes were immunostained with the anti-TH antibody to label DA neuron and anti-GFP antibody to label mitochondria . For measurement of mitochondrial size distribution , the size of each mitochondrial aggregate was represented by the length of its longest axis . The percentage of DA neurons in the PPL1 cluster that have one or more mitochondria exceeding the indicated size was shown . Six flies of each genotype were used for the analysis . The motor neuron-specific OK6-Gal4 driver was used to express UAS-mitoGFP in the larval segment neurons , and 3rd instar male larvae raised at 29°C were used for live imaging . Larvae were briefly washed with water and anesthetized for 4 min in 1 . 5 ml eppendorf tubes containing 4 µl suprane ( Baxter International Inc . ) , before being placed ventral side up into a small chamber . The chamber was created on glass slide with double-sided tape and cover glass . Additional suprane ( 4 µl ) was introduced into the chamber before it was sealed with Valap ( 1∶1∶1 amount of vaseline , lanolin , parafin wax ) . Mitochondria were viewed with an upright Leica DM6000 B microscope equipped with a laser scanner and a 63× oil-immersion objective . Larvae were positioned to have their ventral ganglion ( VG ) appearing on the right of the acquired image and segmental nerves aligned horizontally across the image . Before recording mitochondrial movement , a centered region of 1024×200 pixel ( 61 . 5 µm×12 . 0 µm , 4× digital zoom ) close to the VG was photobleached for 30 s with 488 nm excitation argon laser set at 80% output power . The viewing field was then zoomed out ( 2 . 5× digital zoom ) and mitochondrial movement was immediately recorded by time-lapse video ( 2 s/frame , 300 s total ) in a region of 1024×150 pixel ( 98 . 4 µm×14 . 4 µm ) with the laser power reduced to 10% of the maximum output . The pinhole was set at 200 µm for all the experiments . Time-lapse images were acquired within 30 min of anesthetization . Mitochondrial flux was calculated by normalizing the number of mitochondria that passes certain point within the time frame examined . Mitochondrial net velocity was calculated by normalizing mitochondrial net displacement with time . At least 5 larvae of each genotype were analyzed . HeLa cells were maintained at 37°C with 5% CO2 atmosphere in DMEM ( Wako ) supplemented with 10% FCS ( GIBCO ) and non-essential amino acids ( Invitrogen ) . Plasmids and siRNA duplexes ( Invitrogen ) were transfected using Lipofectamine 2000 ( Invitrogen ) and Lipofectamine RNAiMAX ( Invitrogen ) , respectively , according to manufacturer's instructions . To depolarize the mitochondria , HeLa cells were treated with 10 µM CCCP ( Sigma-Aldrich ) at 36 hr ( for plasmids ) or 72 hr ( for siRNA ) post-transfection . The S156A mutations were introduced into human Miro1 and Miro2 by PCR-based mutagenesis using the following primers: For hMiro1 , Forward primer: 5′- GCA gag ctcttttatt acgcac -3′; Reverse primer: 5′- tatg ttcttcaggt ttttcgc -3′ . For hMiro2 , Forward primer: 5′- GCA gagct gttctactac gc -3′; Reverse primer: 5′- ga tgttcctcag gttcttggc -3′ . Full-length sequences of hMiro1 S156A and hMiro2 S156A were confirmed by sequencing . For brain extract preparation , 6 fly heads were quickly dissected and homogenized in 60 µl SDS-PAGE sample buffer . 5 µl brain extracts from each genotype were loaded onto SDS-PAGE gel . Guinea pig anti-dMiro antibody ( 1∶20000 , from Dr . K . Zinsmaier ) and Peroxidase Anti-Guinea Pig IgG antibody ( 1∶10000 , Jackson ImmunoResearch Labs ) were used for Western Blot . For HeLa cell-based biochemical analysis , cells were lysed in 1% Triton X-100 -based lysis buffer ( 10 mM Tris-HCl [pH 7 . 6] , 120 mM NaCl , 5 mM EDTA , 1% Triton X-100 and protease inhibitor [Nacalai Tesuque] ) . Immunoprecipitation was performed using Immunoprecipitation Kit-Dynabeads Protein G ( Invitrogen ) according to manufacturer's instructions . In vitro kinase assay was performed essentially as described [42] , using a 2× GST-dPINK1 fusion protein with GST fused at both the N- and C-terminus of dPINK1 as the kinase source and a GST-dMiroΔTM fusion protein as the substrate . GST-dMiroΔTM covers amino acids 1–634 of the full-length dMiro protein . The GST-dMiro-ΔTM plasmid was constructed by amplifying a Myc-tagged dMiro fragment without the transmembrane domain from a pUAST-Myc-dMiro plasmid ( obtained from Dr . K . Zinsmaier ) using 5′-CGCCCG-GGTGAGCAGAAACTCATCTCTGAAGAAG-3′ and 5′-ATGCGGCCGCTACTTGGG-GTCCTCCGTC-ATC-3′ as primers . The amplified fragment was inserted into the SmaI and NotI cloning sites of the pGEX-6P-1 vector . Recombinant GST fusion proteins were purified from bacteria according to standard protocols . Cells were fixed with 4% paraformaldehyde in PBS and permeabilized with 0 . 2% Triton X-100 ( for mitophagy in Figure 6C ) or 0 . 5% Triton X-100 ( for mitochondrial morphology in Figure 6B ) in PBS . Cells stained with the appropriated antibodies and counterstained with DAPI were imaged using a laser-scanning microscope ( LMS510 META; Carl Zeiss , Inc . ) with a Plan-Apochromat 63×NA1 . 4 or 100×/1 . 4 Oil differential interference contrast objective lens . Image contrast and brightness were adjusted in Image Browser ( Carl Zeiss , Inc . ) Two-tailed Student's t tests were used for statistical analysis . p values of <0 . 05 , <0 . 01 , and <0 . 005 were indicated with one , two , and three asterisks ( * ) , respectively . | Parkinson's disease ( PD ) is the second most common neurodegenerative disease . It mainly affects movement in elderly people and was traditionally considered a sporadic disease with no known cause . Discoveries of genes associated with familial PD ( FPD ) have demonstrated that PD pathogenesis can be significantly influenced by an individual's genetic makeup . Understanding the functions of these FPD genes will allow better understanding of the sporadic PD cases . PINK1 and Parkin are genes associated with FPD that affect patients at an early age . Mutations in PINK1 and Parkin lead to the accumulation of damaged mitochondria , the powerhouse of the cell , as a result of impairments of the mitochondrial quality control system . However , the mechanism of PINK1/Parkin action remains poorly understood . Here we show that PINK1 and Parkin act together to regulate Miro , a key component of the mitochondrial transport machinery , and that altered activities of PINK1 cause aberrant mitochondrial transport . Regulation of mitochondrial transport may be a critical aspect of the mechanisms by which the PINK1/Parkin pathway governs mitochondrial quality control . Dysfunction of this process could contribute to the loss of DA neurons , the cardinal feature of PD , as well as the peripheral neuropathy symptom associated with particular PINK1 or Parkin mutations . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"animal",
"models",
"medicine",
"cellular",
"structures",
"model",
"organisms",
"molecular",
"cell",
"biology",
"cell",
"biology",
"neurological",
"disorders",
"neurology",
"biology",
"neuroscience"
] | 2012 | Parkinson's Disease–Associated Kinase PINK1 Regulates Miro Protein Level and Axonal Transport of Mitochondria |
Comprising nearly half of the human and mouse genomes , transposable elements ( TEs ) are found within most genes . Although the vast majority of TEs in introns are fixed in the species and presumably exert no significant effects on the enclosing gene , some markedly perturb transcription and result in disease or a mutated phenotype . Factors determining the likelihood that an intronic TE will affect transcription are not clear . In this study , we examined intronic TE distributions in both human and mouse and found several factors that likely contribute to whether a particular TE can influence gene transcription . Specifically , we observed that TEs near exons are greatly underrepresented compared to random distributions , but the size of these “underrepresentation zones” differs between TE classes . Compared to elsewhere in introns , TEs within these zones are shorter on average and show stronger orientation biases . Moreover , TEs in extremely close proximity ( <20 bp ) to exons show a strong bias to be near splice-donor sites . Interestingly , disease-causing intronic TE insertions show the opposite distributional trends , and by examining expressed sequence tag ( EST ) databases , we found that the proportion of TEs contributing to chimeric TE-gene transcripts is significantly higher within their underrepresentation zones . In addition , an analysis of predicted splice sites within human long terminal repeat ( LTR ) elements showed a significantly lower total number and weaker strength for intronic LTRs near exons . Based on these factors , we selectively examined a list of polymorphic mouse LTR elements in introns and showed clear evidence of transcriptional disruption by LTR element insertions in the Trpc6 and Kcnh6 genes . Taken together , these studies lend insight into the potential selective forces that have shaped intronic TE distributions and enable identification of TEs most likely to exert transcriptional effects on genes .
Transposable Elements ( TEs ) are major factors that have shaped the landscape of the mammalian genome through evolution . Most TEs in mammals are inactive remnants of ancient TE insertions , buried in the host genome for millions of years . In rodents and primates , TEs comprise 38–45% of the genome [1] , [2] , and about 90% of all human RefSeq genes contain TEs in their introns . These TEs can be divided into four major classes: long interspersed elements ( LINEs ) , short interspersed elements ( SINEs ) , long terminal repeat ( LTR ) retroelements ( including endogenous retroviruses ( ERVs ) ) , and DNA transposons [3] . The first three classes are retrotransposons , which utilize an RNA intermediate during their retrotransposition process and account for most TEs in mammalian genomes . On the other hand , DNA transposons move directly to new genomic loci without being reverse-transcribed . Although most mammalian TEs are neutral components of the genome with no significant biological effects [4] , [5] , some elements do impact the cell/organism by acting as insertional mutagens , inducing DNA rearrangements , assuming cellular functions and altering gene regulation [4] , [6] , [7] , [8] . Biologically significant TEs are usually discovered and studied on a case-by-case basis , although bioinformatics approaches have also been used to identify potentially functional TEs . Genomic comparisons between species have identified deeply conserved TEs that function as regulatory elements [9] , . TEs that serve as alternative exons , promoters or polyadenylation signals are also straightforward to detect by looking for chimeric transcripts between the TE and neighboring genes [11] , [12] , [13] , [14] . Global TE distribution patterns in mammalian genomes have been intensely studied in the past decade , and such analyses have provided insight into the selective forces that influence fixation probabilities of TE insertions . For example , some studies have evaluated the relationships between TE distributions and imprinted genes [15] , and gene expression patterns [16] , [17] , [18] . TE-free regions have also been used as markers to identify potentially critical regulatory regions [19] , [20] . Moreover , it is clear that LTR elements and LINEs are more prevalent in intergenic regions compared to gene introns , and most of those that do reside in gene introns are in the antisense orientation with respect to the enclosing genes [3] , [21] . This pattern reflects stronger selection against sense-oriented elements , likely due to the greater chance that such elements will disrupt gene transcript processing [22] . While cases have been reported of influential TEs far from genes , those elements near or within genes likely have a greater potential of affecting gene expression . However , our current knowledge of the distribution of TEs within gene introns is very limited , and it remains unclear why some intronic TEs perturb gene transcription while most do not . To fully understand their biological effects , it would be useful to determine which intronic TEs are most likely to affect gene expression , so they can be prioritized for functional analyses . With a growing appreciation for SINE and LINE insertional polymorphisms in human [23] , [24] , [25] , [26] , [27] , [28] , such predictions would be particularly helpful in identifying polymorphic TE insertions with the greatest probability of affecting gene transcription and , therefore , possibly contributing to phenotypic variability or disease susceptibility in humans . In this study , we conducted a set of bioinformatics analyses of TE distribution patterns within human and mouse genes and revealed TE underrepresentation zones and distributional biases in gene introns . TEs that do occur within the underrepresentation zones are more likely to be involved in aberrant gene splicing and known cases of intronic disease-causing TE insertions are primarily located within these zones , strongly suggesting that TEs in these locations are more likely to be harmful and be selected against . The results of our study reveal a distinct tendency for TEs to affect gene transcription when poised near exons , and point to their continued role in catalyzing genome evolution .
According to our genomic survey , 85–90% of mouse and human protein coding genes contain TE sequences in their introns . In a recent study of the relationship between Alu SINEs and alternative splicing , Lev-Maor et al . reported a drop of Alu density within 150 bp from intron boundaries [29] . Based on this observation and the fact that most intronic splice signals are located at the 5′- and 3′-end of introns [30] , we hypothesized that de novo intronic TE insertions near exons are more likely to be mutagenic , and consequently , that the frequency of TEs would be significantly lower than expected in general near intron ends . To analyze the distributions of various TE classes within introns , we first conducted computer simulations to determine theoretical TE distribution patterns ( see Materials and Methods ) . Then we determined the actual distribution pattern of intronic TEs according to their distance to the nearest exon . To alleviate our concern about the potential effect of “distance shifting”- a hypothesized result of later TE insertions or other rearrangements occurring between a specific TE and its nearest exon , we also analyzed the distribution of the youngest 20% of intronic TEs . However , we observed only minor differences compared to all intronic TEs in the genome ( data not shown ) . To clearly show the difference between simulated and actual TE distributions at each predefined position in introns , we calculated the ‘standardized frequency’ of observed TEs ( see Materials and Methods ) . Briefly , the level of TE representation at each predefined intronic interval is determined from the difference between the actual TE distribution in the genome ( observed ) and the computer simulation of random TE insertions ( expected ) . When this value is positive , it reflects an overrepresentation of a given TE class within the corresponding intronic region; however , when negative it indicates underrepresentation . As expected , we found that all four major TE classes are highly underrepresented near intron boundaries in both human ( Figure 1A in Text S1 ) and mouse ( data not shown ) . We next applied the same distribution analysis for only full-length or near full-length TE sequences ( see Table 1 for “full-length” definitions ) . Again , as shown in Figure 1B in Text S1 for human , full-length TEs were highly underrepresented when close to exons , but most TE classes except SINEs showed larger underrepresentation zones ( hereafter shortened to U-zone ) compared with the all-TE distributions . We also noticed that intronic regions more than 20 kb from exons showed a significant underrepresentation of SINEs compared to random simulations . Unlike patterns close to exons , intronic TE distributions greater than 20 kb from exons are less likely due to purifying selection so we searched for other explanations . SINE elements are more abundant in G/C-rich regions [1] , [21] and , since large introns resemble intergenic regions in terms of G/C content ( which is generally A/T rich ) [31] , we postulated that the drop of SINE frequency compared to random simulations in deep intronic regions was an effect of local G/C content . To determine if this was the case , we normalized our random simulations with the local G/C content as described in Materials and Methods . Indeed , after applying such normalization , the underrepresentation of SINEs in deep intronic regions greatly flattened out , while the sizes of the U-zones near exons were not affected . Hence all our subsequent analyses employed this normalization . Figure 1 shows the normalized plots for all human TEs ( Figure 1A ) and full length TEs ( Figure 1B ) , and these plots are very similar for mouse TEs ( Figure 2 in Text S1 ) . Interestingly , the sizes of the U-zones near intron boundaries are different between TE classes ( Table 1 ) . Original insertion site preferences , natural selection and genetic drift could all contribute to global TE distributions . While determining the initial integration site preference of TEs is difficult if not impossible ( especially for ancient families ) , a limited number of de novo TE integration studies showed that TEs in today's human genome are distributed very differently from their initial target site preferences [32] , [33] . Indeed , since 99% of TEs in the human genome and 93% in the mouse genome have been fixed for more than 25 million years [1] , it is reasonable that their current distributions will bear little resemblance to any original insertion site preferences but will primarily be the result of selection and genetic drift . Therefore , the TE U-zones identified here most likely result from purifying selection , rather than original avoidance of these regions during the integration process . The larger U-zones for full length TEs ( compare Figures 1A and B ) suggests that purifying selection acts at much greater distances on full-length elements than on their partly deleted counterparts . This effect is not observed for SINEs but these elements have a much shorter full-length size ( ∼300 bp for human Alus ) [1] , [8] , will generally carry fewer cryptic transcriptional regulatory signals and are less harmful to the enclosing genes than other TEs [34] . For the above reasons , full-length SINE elements may be better tolerated at a closer distance to exons . We next compared the average length of intronic TEs within and outside their full-length U-zones and found a significant difference for all TE classes in both species ( Figure 2 for human; Figure 3 in Text S1 for mouse ) . In fact , most elements within their respective U-zones are truncated , while a greater portion of TEs beyond such zones are full-size elements , resulting in a much bigger size variance ( see the difference between upper whiskers in Figure 2 for human and also Figure 3 in Text S1 for mouse ) . Therefore , the length of individual TEs is an important aspect dictating their genomic distributions , indicating that larger elements are more likely to be genotoxic when positioned near exons . These results also support previous work regarding L1 LINEs , indicating that , compared to shorter elements , full length L1s have more potentially disruptive splice and polyadenylation signals [35] , have greater effects on expression of enclosing genes [36] and have a greater fitness cost [37] . We next examined the distribution of intronic TEs in the sense orientation versus those in antisense with respect to the enclosing genes ( see Figure 3A for human and Figure 4A in Text S1 for mouse ) . Since DNA transposons only comprise about 3% of both the human and the mouse genomes and almost all of them are ancient elements without evidence of any transposition activity during the past 50 Myr ( million years ) [1] , [2] , we excluded them from the following analyses to avoid uncertainties introduced by their relatively small numbers . While previous studies have found an overall antisense orientation bias in genes ( particularly for LTR elements and LINEs ) [21] , [22] , we show here the existence of a much stronger bias in antisense for both LINEs and LTR elements near exons . The excess of antisense TEs compared with sense elements near intron boundaries is probably the result of purifying selection , like the genome-wide orientation bias of TEs in genes . This indicates in general that sense-oriented TEs near splice sites have a higher probability to influence normal gene transcription and are potentially more harmful to the host gene . Interestingly , for SINEs we observed the same strong antisense bias in the mouse ( Figure 4A in Text S1 ) , but in the human genome we observed a sense orientation bias instead of antisense for SINEs at a close distance of 20–200 bp from exons ( Figure 3A ) . These data are consistent with the Alu SINE study of Lev-Maor et al . [29] , in which the authors also observed more sense-oriented Alu elements near intron termini . Since Alus account for two-thirds of human SINE elements and many antisense Alus possess a strong cryptic SA signal [13] , selection against antisense-oriented elements may explain the unusual underrepresentation of antisense oriented SINEs near splice sites in humans . Furthermore , we also looked for evidence of any distributional bias of intronic TEs in terms of their proximity to either splice donor sites ( SDs ) or splice acceptor sites ( SAs ) . We found the total numbers of elements near SA sites are much lower than SD sites for all three retrotransposon classes examined ( see Figure 3B for human and Figure 4B in Text S1 for mouse ) . Since the core intronic splice signals at SD sites usually only consist of about 6 bp of terminal intron sequence compared with 20–50 bp at SA sites [30] , selection against physical disruption of critical splice motifs likely underlies this TE underrepresentation near SA sites . Theoretically , harmful antisense transcripts of protein-coding exons may be generated by read-through transcription of antisense TEs near SD sites . If such antisense transcripts have significant detrimental effects , then one might expect a larger proportion of TEs near SD sites to be in sense rather than in antisense due to purifying selection . However , as shown in Figure 4A ( human ) and Figure 5A in Text S1 ( mouse ) , such predicted bias of sense orientated TEs near SD sites was not found except for human SINEs , which is likely explained by the fact mentioned previously that antisense Alus possess cryptic SA signals . In fact , for other TE classes we observed more SD-associated elements oriented in antisense , probably indicating that antisense transcription is effectively silenced or not a general problem , and that sense oriented TE insertions are more detrimental . The same analysis of TEs near SA sites revealed similar orientation bias patterns as for TEs near SD sites . If the reduced frequency of TEs near intron boundaries reflects the force of selection against harmful insertions , one would predict that a higher fraction of mutagenic TEs in gene introns would be located within these TE underrepresentation zones . To evaluate this prediction , we compiled information on documented intronic mutagenic TE insertions and examined their integration sites in introns . Based on the TE activity and data availability , we focused on the following three TE families in our analyses: human Alu ( SINE ) , human L1 ( LINE ) and mouse LTR elements . First , as the most abundant TE family , Alus have successfully propagated in the human genome and reached a total number of over one million copies [1] . Even today , some of these elements are still active , generating new insertions and causing mutations linked to diseases [8] , [38] , [39] . Based on the information provided by the dbRIP database ( http://dbrip . brocku . ca/ ) [27] , we found six de novo Alu insertions associated with human diseases within introns , all of which belong to the AluY subfamily ( the youngest subfamily of Alu ) and cause splice defects of the enclosing gene ( Table 1 in Text S2 ) . Second , de novo disease-causing insertions of L1 , the active LINE family in humans , have also been reported [5] , [40] , [41] , [42] . These elements play important roles in human retrotransposon-mediated pathogenesis because not only do they encode reverse-transcriptase ( RT ) and other proteins required for their own retrotransposition , but also for mobilizing Alus [43] . In this study , our search of the dbRIP database identified a total of five intronic L1s associated with human diseases ( Table 2 in Text S2 ) , all of which cause transcriptional disruptions . Last , since no mutagenic LTR insertions and only a few insertionally polymorphic ERVs or LTRs have been reported in human [4] , [6] , [44] , we turned to the mouse genome , where ERVs/LTR elements cause ∼10% of germline mutations , many of which have been well studied [7] . In total we collected 40 cases of mutagenic LTR elements in mice: 15 from the Intracisternal A Particle ( IAP ) family , 18 from the Early Transposon/Mouse Type D retrovirus ( ETn/MusD ) family , and seven from other LTR elements or ERVs . Again , all these ERV-induced intronic mutations in mice are due to transcriptional disruptions on the enclosing gene ( Table 3 in Text S2 ) . For the three TE families listed above , we compared the intronic distribution of mutagenic elements with all full-length counterparts in the reference genomes and found highly consistent results ( Figure 5 and Table 2 ) . As shown in Figure 5A , all six mutagenic Alu insertions are within the U-zone of SINEs ( i . e . <100 bp from the nearest exon ) , and all are oriented antisense with respect to the enclosing gene . Moreover , five out of the six cases are near SA sites . In comparison , only 1 . 83% of all full-length AluYs in the reference human genome are located within the 100 bp U-zone - strikingly lower than the mutagenic elements and also more than two-fold lower than that expected by chance ( p<2 . 2e-16; one-sample proportion test ) . For all full-length AluYs within the U-zone we observed 47 . 7% elements in antisense , slightly lower than the random level ( 50% ) but much lower than mutagenic insertions . Since intronic TEs show their strongest splice site bias when they are in extreme close proximity to an exon ( Figure 3B ) , we examined full-length intronic AluYs located less than 20 bp from exons and observed only 10% of such elements near SA sites . Although we cannot directly compare this result to the case of mutagenic Alus due to their insufficient number within 20 bp to exons , the fact that five out of six mutagenic Alus are near SAs is noteworthy . Similarly , Figure 5B shows that all five mutagenic L1 elements are within the U-zone for full-length LINEs ( i . e . <2 kb from the nearest exon ) . Among them , four are sense-oriented and four are near SA sites . In contrast , only 23 . 0% of full-length intronic L1s in the reference genome are within the U-zone , which is significantly lower than both the mutagenic L1s and our random simulation ( p<0 . 0004 and p<2 . 2e-16 , respectively; two-/one-sample proportion test ) . Of those elements within the U-zone , only 27 . 7% are in sense , again significantly lower than both mutagenic insertions and the simulation ( p<0 . 035 and p<2 . 2e-16 , respectively; two-/one-sample proportion test ) . Although the number of full-length L1s in the reference genome within 20 bp to exons is very limited , among a total of seven cases only two were found near SA sites . We also examined the same parameters for mouse LTR elements ( Figure 5C and Table 2 ) . As we expected , a high fraction of these mutagenic insertions ( 72 . 5% ) are within the U-zone of full-length mouse LTR elements ( i . e . <2 kb from the nearest exon ) . More remarkably , all 15 mutagenic insertions from the IAP family were within the 2 kb U-zone . Since the orientation information of some mutagenic LTR elements within the U-zone was not indicated in their original reports , we checked the remaining 26 cases and found 20 ( 76 . 9% ) were oriented in sense . Among these mutagenic insertions in mice , five are located within 20 bp of exons , with three of them near SA sites ( 60% ) . However , the situation is completely different for all full-length LTR elements in the sequenced mouse genome ( strain C57BL/6J , or B6 ) . In contrast to mutagenic insertions , only 14 . 3% of full-length LTR elements in the reference genome were located within the 2 kb U-zone ( p<2 . 2e-16; two-sample proportion test ) , and of these elements only 30 . 1% are in the sense orientation ( p<2 . 65e-09; two-sample proportion test ) . At a distance less than 20 bp to exons , we found six full-length LTR elements in the B6 reference genome but only one of them is near the SA site ( 16 . 7% ) . In summary , the above analyses of mutagenic versus all full-length elements for the three retrotransposon families consistently showed an overrepresentation of mutagenic TEs within their respective U-zones but an underrepresentation of all full-length elements within the same regions . Moreover , apparent differences in orientation and splice-site biases were also observed between mutagenic TEs and all full-length elements in the reference genomes . These observations strongly suggest that intronic TE insertions within the U-zone have a much higher potential to be deleterious to the enclosing gene , particularly when oriented in antisense for human SINEs and in sense for LINEs and LTR elements . When intronic TE insertions are in extreme proximity ( e . g . <20 bp ) to an SA site , they are very likely to be harmful and may cause functional abnormality of the enclosing gene . We next extended our analyses to polymorphic AluY and L1 insertions not associated with any disease based on the dbRIP data . These elements are considered as relatively young since they are not fixed in humans . If , indeed , selection is still working upon these TEs , one might see an intermediate distribution pattern between that of mutagenic and all elements . However , for both polymorphic AluYs and L1s we observed no significant differences from all full-length elements in the reference human genome ( data not shown ) . While the limited total number of polymorphic insertions documented in dbRIP may partially account for this result , it is very likely that the distribution of these polymorphic TEs has already been shaped by selection . However , for the youngest insertionally polymorphic mouse LTR elements , we have previously shown that they do have a distinct prevalence in introns and orientation bias compared with older elements [45] . This suggests that some of these insertions are detrimental but have not been eliminated due to the artificial breeding environment of inbred strains [2] , [7] , [45] . Indeed , some known detrimental LTR insertions have even become fixed in one or a few mouse strains [46] , [47] . We therefore analyzed a list of polymorphic LTR insertions in four mouse strains from our previous study [45] , in which we had detected different distributions between polymorphic and common LTR elements . Here we used polymorphic IAP and ETn/MusD elements that are present in only one of the four analyzed mouse strains ( presumed to be the youngest elements ) and found that 34 . 8% of intronic insertions were within the 2 kb U-zone ( Figure 5C and Table 2 ) , a fraction very close to the simulated prediction of a random distribution but significantly higher than all full-length LTR elements in the mouse reference genome ( 14 . 3%; p<5 . 58e-13; two-sample proportion test ) and lower than the mutagenic insertions ( 72 . 5%; p<9 . 79e-05; two-sample proportion test ) . Moreover , we observed 23 . 2% of polymorphic LTRs in the U-zone as sense-oriented , which shows no statistical difference from that of all LTRs but is highly significantly lower than the mutagenic cases ( p<6 . 26e-07; two-sample proportion test ) . Since our list of polymorphic LTR insertions in mice does not contain any intronic insertions within 20 bp of an exon , we could not perform the analysis of splice site proximity bias . Nonetheless , the above observation of an intermediate distribution pattern of polymorphic insertions between mutagenic and all full-length TEs in the reference genome demonstrates that , indeed , purifying selection is the most likely underlying force shaping the observed intronic TE distribution patterns , and evidence suggests that such a process is ongoing . If TEs within their respective U-zones are more likely to be harmful by causing splicing abnormalities , one can make two predictions . One prediction is that TEs located in the U-zones would be associated with chimeric TE-gene transcripts more often than TEs located elsewhere in introns . To test this prediction , we downloaded and analyzed the human expressed sequence tag ( EST ) data from the UCSC Genome Browser ( http://genome . ucsc . edu ) , in which only spliced transcripts were included . We then screened for all spliced ESTs overlapping with intronic TEs ( i . e . chimeric ESTs ) . As shown in Figure 6A , we observed that 11 . 7% of human SINE elements within the 100 bp U-zone are associated with chimeric ESTs . In contrast , this ratio is only 1 . 6% for SINE elements outside the U-zone . Similarly , for human LINEs in their 2 kb U-zone , we found 4 . 6% of them associated with chimeric ESTs , while outside the U-zone the ratio significantly drops to 0 . 7% . Lastly , we identified 2 . 9% of human LTR elements as chimeric-EST-related in the 5 kb human LTR U-zone , but for elements outside the U-zone we observed only 0 . 9% . All the above results are highly statistically significant ( all p-values<2 . 2e-16; two-sample proportion test ) , which reinforces the notion that TEs within their U-zones are more likely to be involved in aberrant splicing . It should be pointed out , however , that the splicing events detected by this analysis are of unknown relevance and , indeed , because these TEs are fixed , are unlikely to have significant detrimental effects . A second prediction is that TEs which were not eliminated from the U-zone would have weaker splicing signals compared with other TEs . To examine this issue , we computationally analyzed potential splice sites within randomly selected solitary LTR sequences in human introns using NNSplice [48] ( see Materials and Methods ) . As shown in Figure 6B , as the distance between the intronic LTR and its nearest exon decreases , the average number and the strength of predicted splice sites in these LTR sequences also decrease . This observation indicates that LTRs carry fewer and weaker cryptic splice sites within the U-zone , especially when they are located in close proximity to exons . While the above EST analysis suggests the importance of U-zones in TE-gene interactions , it would be useful to predict which particular intronic TEs are most likely to influence gene transcription based on their size , distance to the nearest exon , orientation , and proximity to particular splice site . To conduct an initial evaluation of this concept , we examined a panel of polymorphic LTR element insertions in inbred mouse strains because they are currently highly active and , as discussed above , their genomic distribution suggests that some are likely detrimental but are maintained due to the artificial breeding environment . In order to take the advantage of the available EST/mRNA data in the B6 reference genome , we restricted our set of intronic polymorphic LTR elements to those present in the B6 mouse strain [45] . After excluding solitary LTRs and complex cases due to multi-gene families , we identified 44 full-length polymorphic LTR elements within the 2 kb U-zone ( data not shown ) . We then inspected each region using the UCSC Genome Browser ( mouse genome version: mm9 ) to look for chimeric ESTs/mRNAs involving the LTR element and the enclosing gene and found such transcripts for 19 of the 44 genes . For most of these 19 genes , the aberrant forms appear to be minor in abundance and it is difficult to estimate their overall impact on gene expression . However , among these 19 genes , transcription of three of them ( Cdk5rap1 , Adamts13 , and Wiz ) has been shown to be significantly affected by the embedded LTR element [46] , [49] , [50] . Judging from the frequency of annotated chimeric transcripts , two other genes among the group of 19 , Kcnh6 ( potassium voltage-gated channel , subfamily H ( eag-related ) , member 6 ) and Trpc6 ( transient receptor potential cation channel , subfamily C , member 6 ) , are of special interest . While no evidence of transcriptional disruption caused by LTR element insertions has been reported in the literature for these genes , UCSC Genome Browser snapshots of their deposited mRNAs suggest significant involvement in the transcription of each gene . For Trpc6 , two of seven mRNAs in the database terminate within a polymorphic IAP LTR element ( Figure 7A ) , and for Kcnh6 , one of three annotated mRNAs terminates within another IAP insertion ( Figure 7B ) . Trpc6 plays an important role in vascular and pulmonary smooth muscle cells and its deficiency impairs certain allergic immune responses and smooth muscle contraction [51] . Kcnh6 , also termed Erg2 ( eag related protein 2 ) , encodes a pore forming ( alpha ) subunit of potassium channels , and may serve a role in neural activation [52] . To examine the potential effect of the IAP polymorphisms on transcription of these two genes , we first confirmed the presence or absence of these insertions by genomic PCR in a panel of mouse strains including B6 , A/J , and 129SvEv . Indeed , an IAP is present in B6 and A/J but not in 129SvEv for the Trpc6 gene , and the IAP in the Kcnh6 gene is present only in B6 but not in A/J and 129SvEv ( data not shown ) . Since both genes are highly expressed in the brain , we conducted quantitative RT-PCR on brain cDNA from all three mouse strains by setting one primer pair upstream of the insertion site and another primer pair flanking the insertion site , as indicated in Figure 7 . In mouse strains carrying the IAP insertion , we found a significant decrease in the amount of normally spliced transcripts involving exons flanking the ERV insertion , compared with exons upstream of the insertion . In contrast , we saw less difference between the upstream and flanking primer sets in strain ( s ) without the IAP insertion ( Figure 8 ) . The blockage of normal Kcnh6 transcription is particularly striking , with very little normal splicing occurring for exons flanking the IAP in the B6 strain . These data suggest significant transcriptional interference of these two genes mediated by the embedded IAPs , and it would be interesting to determine if this interference results in phenotypic differences between mouse strains with and without these insertions . Over a million TEs have become fixed in human or mouse gene introns during evolution , and the vast majority of them presumably have no functional impact on the gene . Yet , new disease-causing TE insertions do occur in introns and exert detrimental effects mainly by disrupting normal gene transcript processing . The emergence of high throughput technologies has facilitated the discovery of an increasing number of TE germline polymorphisms and somatic insertions in human cancers , with the recent advances on studies of human L1 polymorphisms as the best example [23] , [24] , [25] , [26] . However , little attempt has been made thus far to identify which of these polymorphic or somatically-acquired TEs may contribute to allele-specific gene expression differences and potential phenotypic variation or disease . Methods are therefore needed to evaluate which TEs are most likely to affect gene transcription . Here we have identified intronic underrepresentation zones near exons , where fixed TEs occur less often than expected by chance . Strikingly , all documented human intronic Alu and L1 insertions and most mouse intronic LTR elements known to cause disease are located within these U-zones , strongly suggesting that TE elements in these locations are more likely to cause transcriptional disruptions and be eliminated by selection . Moreover , TEs within their U-zones are more likely to be involved in spliced chimeric transcripts than those located elsewhere in introns , suggesting that some may be slightly detrimental . Presumably in most of these cases the transcriptional effects must be insufficient to cause such insertions to be eliminated by purifying selection . However , it is possible that even apparently subtle effects on gene splicing could have functional consequences . On the other hand , previous studies have also demonstrated that TEs fixed in the host genome can participate in gene transcription , producing alternative transcript isoforms that might have functional importance [12] , [53] , [54] , [55] , [56] . Equally important as identifying potentially deleterious TE insertions , it is also of great value to identify fixed TEs that contribute to normal gene expression and cell functionality . The U-zones identified here , coupled with TE size , orientation bias , and location relative to SD or SA sites can all be combined to help predict those TEs with a higher likelihood of functional significance , while yielding new insights into the effects of TEs on gene regulation and evolution .
To establish a baseline of TE distributions in gene introns , we applied computational simulations of random TE insertions in both the hg18 human genome and the mm9 mouse genome . We used the RefSeq gene annotation data downloaded from the UCSC genome browser ( http://genome . ucsc . edu ) in our study . For each round of simulation , we generated 1 , 000 , 000 random genomic loci across the entire host genome to mimic randomized TE insertions . Next , we divided intronic regions into the following 13 bins with gradually increasing bin size according to their distance to the nearest exon: 0–20 bp , 20–50 bp , 50–100 bp , 100–200 bp , 200–500 bp , 500–1000 bp , 1–2 kb , 2–5 kb , 5–10 kb , 10–20 kb , 20–50 kb , 50–100 kb , >100 kb . The intention of using increasing bin size was to establish a higher resolution at the interesting regions near intron boundaries while also maintain a good overview of other intronic regions . We then calculated the fraction of simulated TE insertions located in each bin with respect to the total simulated insertions in introns . The same simulation process was applied three times and the average was taken for each genome as a control distribution for all further analyses . Finally , our calculation of the “standard error of the mean” based on three rounds of simulations showed a negligible sampling error for each bin ( data not shown ) , confirming the eligibility of using these results to represent the theoretical random TE distribution . To minimize the influence on TE distribution by local G/C content , we corrected our computational simulations of random TE distribution according to the overall G/C preference of each TE class . Specifically , we first performed a genome-wide evaluation of the G/C preference of each TE class by dividing the entire host genome into a set of consecutive 20 kb windows and calculating both the density of each TE class and the G/C density for each window . Then we grouped these 20 kb windows by G/C density level ( with a resolution of 1% ) and calculated their average TE density at each G/C level . Based on the assumption that TE density should be close to the overall genome-wide TE density anywhere in the genome when there is no G/C preference , we calculated the fold-difference of the actual TE density at each G/C level compared with the genomic background level for each TE class . In this way , we derived a list of ‘fold change’ values of TE density at each G/C level , which was then used as the normalization coefficient to correct the simulated distribution of random TE insertions . To determine how different the ‘observed’ TE frequency is from the ‘expected’ at each predefined distance bin , we used the concept of residual to measure the standardized TE frequency:where c is the residual of a given distance bin , obs is the total observed occurrence of a given TE class in that bin , and exp is the expected number of such TE insertions derived from our computational simulations . Common logarithm ( log10 ) was used here to equalize the value ranges of over- and under-represented data , and the addition of “1” in the formula is to fulfill the requirement that the subject of logarithm cannot be a negative number . Literally , the absolute value of residual c shows the degree of relative difference between the ‘observed’ and ‘expected’ , and when c is positive , it means the corresponding TE class is overrepresented in this region; when c is negative , it means such TE class is underrepresented . Here we used the web-based interface of the NNSplice program ( http://www . fruitfly . org/seq_tools/splice . html ) , which is a bioinformatics tool based on artificial neural networks and used for predicting the presence and the strength of potential splice sites in any given input DNA sequence . Due to the limitation of the maximum length of total input sequences that NNSplice can take , here we only chose sense-oriented solitary LTR sequences ( i . e . LTR sequences annotated with a size between 200–600 bp ) in human introns in this analysis . The intronic region was divided into a set of consecutive bins with increasing bin size according to the distance from the LTR to the nearest exon as the following: 0–200 bp , 200–500 bp , 500–1000 bp , 1–2 kb , 2–5 kb , >5 kb . For all bins except the first bin , a total number of 100 LTR sequences was sampled randomly for three times independently , and the averaged total numbers and strength of potential splice sites based on the three samples were taken as the final values for each bin . Since the first bin ( 0–200 bp ) contains only 101 cases in total , we took all those cases to calculate the average total number and strength of potential splice sites for this bin without sampling . Notably , when we calculated the average strength of potential splice sites for each bin , only the site with the highest score was considered for each LTR sequence . | Sequences derived from transposable elements ( TEs ) are major constituents of mammalian genomes and are found within introns of most genes . While nearly all TEs within introns appear harmless , some de novo intronic TE insertions do disrupt gene transcription and splicing and cause disease . It is unclear why some intronic TEs perturb gene transcription whereas most do not . Here , we examined intronic TE distributions in both human and mouse genes to gain insight into which TEs may be more likely to affect transcription . We found evidence that TEs near exons are likely subject to strong negative selection but the size of the region under selection or “underrepresentation zone” differs for different TE classes . Strikingly , all reported human disease-causing intronic TE insertions fall within these underrepresentation zones , and the proportion of TEs contributing to chimeric TE-gene transcripts is significantly higher when TEs are located in these zones . We also examined insertionally polymorphic mouse TEs located within underrepresentation zones and found evidence of transcriptional disruption in two genes . Given the growing appreciation for ongoing activity of TEs in human , our results should be of value in prioritizing insertionally polymorphic TEs for study of their potential contributions to gene expression differences and phenotypic variability . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"genomics",
"genome",
"evolution",
"evolutionary",
"biology",
"gene",
"regulation",
"molecular",
"genetics",
"biology",
"computational",
"biology",
"genomic",
"evolution"
] | 2011 | Distributions of Transposable Elements Reveal Hazardous Zones in Mammalian Introns |
Localization of specific mRNAs is an important mechanism through which cells achieve polarity and direct asymmetric growth . Based on a framework established in Saccharomyces cerevisiae , we describe a She3-dependent RNA transport system in Candida albicans , a fungal pathogen of humans that grows as both budding ( yeast ) and filamentous ( hyphal and pseudohyphal ) forms . We identify a set of 40 mRNAs that are selectively transported to the buds of yeast-form cells and to the tips of hyphae , and we show that many of the genes encoded by these mRNAs contribute to hyphal development , as does the transport system itself . Although the basic system of mRNA transport is conserved between S . cerevisiae and C . albicans , we find that the cargo mRNAs have diverged considerably , implying that specific mRNAs can easily move in and out of transport control over evolutionary timescales . The differences in mRNA cargos likely reflect the distinct selective pressures acting on the two species .
Cell polarity – asymmetry in shape , protein distribution , and/or sub-cellular function – is an essential feature of most eukaryotic cells and underlies such fundamental processes as cell division , cell differentiation , and cell-cell communication . One mechanism for achieving cellular asymmetry is through the localization of specific mRNAs to different parts of the cell . For instance , the spatial distribution of specific mRNAs in the oocytes of Drosophila melanogaster and Xenopus laevis underlies establishment of embryo polarity in these organisms [1] , [2] , [3] , [4] , [5] . In chick fibroblasts , transport of beta-actin mRNA promotes actin assembly at the leading edge of the cells [6] , [7] , [8] , and in mammalian neurons , transport of RNA to dendrites for localized protein synthesis is critical to synaptic activity [9] , [10] , [11] , [12] . In each of these examples , RNA localization occurs via active transport along cytoskeletal elements: microtubules in the Drosophila embryo , microfilaments in chick fibroblasts , and both structures in the Xenopus embryo and in mammalian neurons . Selective RNA transport is also a key feature of fungi . In the maize pathogen Ustilago maydis , the Rrm4 protein binds RNA and moves along microtubules . Loss of Rrm4 , or mutation of its RNA-binding domain , results in polarity defects and reduced virulence of the organism [13] , [14] . One of the best understood RNA localization mechanisms is the Saccharomyces cerevisiae She system , a riboprotein complex that uses actomyosin transport to move a set of mRNAs from the mother cell to the bud during mitosis [15] , [16] , [17] , [18] , [19] . Within the She complex , She2 is thought to be the primary RNA binding protein that links specific mRNAs to Myo4 , a type V myosin motor , via the adaptor protein She3 [20] , [21] , [22] , [23] . Thus , a small set of mRNAs , selected by binding to She2 , is transported from the mother cell to the bud . One such mRNA encodes Ash1 , a transcriptional repressor of HO , an endonuclease required for mating-type interconversion; Ash1 localization to daughter cells ensures that only mother cells express HO and thereby undergo this type of programmed DNA rearrangement [24] , [25] , [26] , [27] . In this study , we investigated the biological role of She-dependent RNA transport in Candida albicans , a commensal fungus and an opportunistic pathogen that can cause severe infection in immunocompromised humans . In the host , C . albicans exists in a variety of morphological forms , including budding yeast , pseudohyphae ( chains of elongated ellipsoidal cells ) , and hyphae ( chains of long , cylindrical cells with parallel cell walls ) [28] . The ability to rapidly switch among these forms in response to external cues is one of numerous factors contributing to virulence . The hyphal form in particular has been associated with numerous virulence attributes such as passage through endothelial and epithelial barriers and host tissue damage . C . albicans hyphae are formed by polarized growth at the apical cell ( the hyphal tip cell ) . Several morphological and molecular characteristics distinguish the hyphal tip cell from the sub-apical ( i . e . , non-tip ) cells of the filament . Newly formed apical cells inherit most of the cytoplasm and are cytologically active , while the mother or sub-apical cells are extensively vacuolated and undergo temporary cell cycle arrest [29] . Further , the Golgi complex is continuously redistributed to tip cells [30] , suggesting a means by which hyphae achieve localized secretion at their tips . As in other filamentous fungi , the tip of C . albicans hyphae contains the Spitzenkörper , a cluster of exocytic vesicles that drives polarized growth by concentrating secretion at the tip [31] . Finally , there is evidence that hyphal tip cells serve a specialized function during C . albicans invasion of host tissues . Electron micrographs have shown a zone of clearing around hyphae penetrating mammalian epithelia , suggesting a concentration of hydrolytic enzymes at the invading tip [32] . At least one such enzyme , phospholipase B , has been shown to be preferentially secreted from the hyphal tip cells [33] . In this study , we establish the existence of a She3-dependent mRNA transport system in C . albicans . In addition , we ( 1 ) identify a set of RNA transcripts specifically bound to She3; ( 2 ) determine the cellular localization of She3-bound transcripts; ( 3 ) characterize the phenotypes associated with loss of She3 , and ( 4 ) study the effects of deleting individual genes whose mRNAs are She3-bound . From the results of these experiments , we conclude that C . albicans has a She3-mediated system that transports selected transcripts into both daughter cells of budding yeast and into tip cells of the hyphae . We further show that approximately one third of these transcripts have roles in hyphal development . Finally , we show that She-based RNA transport , although not required for hyphal growth per se , is important for proper hyphal morphology and for specific aspects of hyphal function , specifically , invasive hyphal growth and tissue damage . Although the general features of the C . albicans She transport system appear conserved with those of S . cerevisiae , the mRNAs carried by She3 differ considerably between the two species , suggesting relatively rapid evolutionary turnover in the set of cargo mRNAs . This finding is analogous to comparisons of transcriptional circuits between C . albicans and S . cerevisiae; although the transcriptional regulators are often highly conserved , the genes they regulate can differ considerably .
A search of the genome sequence of C . albicans ( www . candidagenome . org ) revealed ORF19 . 5595 , predicted to encode a 377 amino acid protein , as a likely ortholog of S . cerevisiae SHE3 . An alignment of this protein with other putative fungal She3 orthologs indicates that the region of highest conservation is in the amino-terminal half of the protein , the putative myosin interaction domain [23] . No clear SHE2 ortholog was identified in C . albicans; either She3 may serve as the RNA-binding protein , or another , yet-unidentified protein may fulfill this function in C . albicans . The C . albicans genome contains a single gene encoding a class V myosin , MYO2 ( orf19 . 5015 [34] ) ; if the RNA transport mechanisms are similar in S . cerevisiae and C . albicans , Myo2 is most likely the motor linking She3 to actin filaments . Previous work has supported the idea that a She3-dependent mechanism of RNA transport may operate in C . albicans . C . albicans Ash1 protein is restricted to the tip cells of hyphae , as well as to daughter cells of budding yeast [35] . When the mRNA encoding C . albicans Ash1 was expressed in S . cerevisiae , it accumulates in daughter cells [36] , indicating that the C . albicans ASH1 transcript may contains localization signals that are recognized by the S . cerevisiae She complex . To directly test whether C . albicans possesses a She3-dependent RNA transport system , we deleted both copies of the C . albicans SHE3 gene ( C . albicans is diploid ) ( strains used in this study are listed in Table 1 ) . We observed that Ash1 now appears in both mother and daughter nuclei in yeast , and in nuclei of multiple cells of hyphae ( Figure 1 ) . We used fluorescent in situ hybridization ( FISH ) to detect localization of the endogenous ASH1 transcript in wild type and she3Δ/she3Δ cells . We observed that ASH1 mRNA accumulates in yeast daughter cells and in the tips cells of hyphae in a She3-dependent manner . The results indicate that C . albicans Ash1 localization ( to daughter cells in yeast and to tip cells in hyphae ) is mediated by She3 and likely occurs through specific localization of the ASH1 transcript – as occurs in S . cerevisiae . We used immunoprecipitation ( IP ) of She3-RNA complexes , followed by hybridization to whole genome microarrays , to identify the set of RNAs bound and potentially localized by C . albicans She3 ( Figure S1 ) . Cellular lysates were prepared from a C . albicans strain ( YSE25 ) containing a single copy of She3 fused to a tandem affinity purification tag ( She3-TAP ) [37] , which was grown in the yeast form ( YEPD medium 30°C ) or induced to form hyphae by addition of serum at 37°C for 30 minutes , one hour , or three hours . The tagged She3 protein was immunoprecipitated from these lysates , and the associated RNAs were eluted . Labeled cDNA generated from the She3-associated RNA was compared to reference cDNA by competitive hybridization to C . albicans microarrays representing the entire genome [38] . We used two different types of reference RNA: ( 1 ) total RNA from the She3-TAP strain , or ( 2 ) RNA isolated from a mock IP performed with an untagged strain . Use of the first type of reference risks false positives inherent to the IP methods ( i . e . , “sticky RNAs” ) , whereas use of the second is subject to potential complications arising from the use of two different strains . In a given experiment , we used either one reference or the other , and we combined results for data analysis , as explained below . This approach allowed us to eliminate false positives inherent to either method . Stringent filter criteria were used to identify the set of candidate She3-associated RNAs . Data were derived from twelve microarrays from yeast ( six each using either of the two reference samples ) and 24 from hyphae ( from each of three time points , four arrays each using the two reference populations ) . To pass the initial filter , array elements ( spots ) must have produced interpretable hybridization in greater than 50% of arrays from any single experiment ( i . e . , from one growth condition using one reference population ) and must have had a median percentile rank of at least 98 . A second filter required that transcripts be identified using both reference populations , and , for those identified from hyphal lysates , be identified in at least two time points . These criteria identified a set of 31 high-confidence transcripts bound by She3 in yeast and a largely overlapping set of 38 high-confidence transcripts bound by She3 in hyphae ( Table 2 ) . The genes represented by the set of She3-bound transcripts act in a variety of cellular processes , including mitosis and cytoskeletal dynamics , cell polarity , transcription , small molecule transport and regulation , virulence , and cell wall structure and function ( Table 2 ) . Ten genes encode proteins of unknown function . ASH1 was identified as She3-associated in both yeast and hyphae , validating the approach and providing independent evidence that Ash1 protein is localized via She3-dependent transport of ASH1 RNA . For the most part , the She3-bound mRNAs are the same in yeast and in hyphae; only two of these transcripts were identified as She3-bound solely in yeast , and nine were identified as bound only in hyphae . In general , these patterns do not reflect the relative abundance of the transcripts in yeast versus hyphae , as determined by previous studies ( www . candidagenome . org ) . Of the 24 RNAs identified as She-transported in S . cerevisiae , clear orthologs of only two – ASH1 and WSC2 – were also identified as She-associated in C . albicans ( Table 2 and [16] ) . Two possibilities could explain this difference: 1 ) the mRNAs transported by She3 differ considerably between the two species , or 2 ) the difference is an artifact of overly stringent filter criteria . To distinguish between these alternatives , we analyzed the C . albicans yeast form IP data to determine the percentile ranking of close homologs ( including orthologs ) of those mRNAs transported in S . cerevisiae . Excluding ASH1 and WSC2 , we identified clear C . albicans homologs ( or , at least , best BLAST hits ) for 13 of the S . cerevisiae She-transported mRNAs . We considered the percentile rank of all spots that met our basic threshold criteria ( i . e . , produced interpretable hybridization in greater than 50% of arrays from one type of experiment ) ; based on these criteria , array spots representing eleven genes were included in the analysis . The median percentile rank across all these array spots was 59 , suggesting that these transcripts are not significantly enriched in the C . albicans She3-TAP IPs . When the same analysis was applied to the set of transcripts that were She3-associated in C . albicans , the median percentile rank was 99 . Thus , the She machinery appears to bind distinct sets of transcripts in C . albicans and in S . cerevisiae . The RNA elements that specify She3-dependent transport in S . cerevisiae are incompletely understood; they appear to be a complex combination of RNA secondary structure and RNA primary sequence [19] , [39] , [40] , [41] , [42] . For these reasons , simple sequence inspection of the transported mRNAs could not reveal whether the C . albicans She3-dependent transport system uses signals similar to those in S . cerevisiae . Based on these results , we predicted that transcripts bound to C . albicans She3 would accumulate in yeast daughter cells ( buds ) and in the tip cells of hyphae and that this accumulation would be She3-dependent . We tested this prediction for 21 transcripts bound by She3 using FISH . Each probe was hybridized to wild type and she3Δ/she3Δ cells under conditions in which the transcripts had been identified from the She3 IP experiments . Fourteen mRNAs were clearly detectable by FISH in the wild type background . In yeast cells , hybridization was observed in the presumptive bud site and/or the bud . In hyphae , signal accumulated at the distal tip of the germ tube ( the nascent form of hyphae where a yeast cell sends out a long projection ) and in the tip cells of mature hyphae . Signal accumulation in the bud and/or hyphal tip cell was absent in the she3Δ/she3Δ cells . Table 3 summarizes the results of the FISH experiments , and representative examples of She3-dependent RNA localization are shown in Figure 2; additional images are provided in Figure S2 . In some cases , ( e . g . , CHT2 in yeast ) , fluorescence was clearly visible and diffuse in the mutant strain . In other examples ( e . g . , RBT4 in hyphae ) , the fluorescence in the mutant strain was not detectable above background . The lower signal in she3Δ/she3Δ cells , particularly in hyphae , suggests that site-specific accumulation is critical for visualizing the signal . It is unlikely that a lower signal in the deletion strain reflects reduced mRNA expression or stability; microarrays comparing the transcriptional profiles of SHE3/SHE3 and she3Δ/she3Δ strains showed no overall decrease in levels of She3-associated transcripts ( nine of the 14 probes with clear FISH results were analyzed; data not shown ) . In any case , the majority of probes ( 14/21 ) revealed that mRNAs identified as She3-bound were localized in a She3-dependent fashion , validating the IP and microarray analysis methods for identifying transported transcripts . In order to exclude the possibility that She3 transports all or most mRNAs in C . albicans , we performed FISH with three control probes , ACT1 ( orf19 . 5007 ) , ACC1 ( orf19 . 7466 ) and ADH1 ( orf19 . 3997 ) . These genes all had a median percentile rank of less than 75 in the She3 binding experiments . In each case , no specific localization was detected in yeast or in hyphae . Moreover , there was no detectable difference in distribution of the signal between wild type and she3Δ/she3Δ cells ( Figure 2B and 2D ) . This result supports the conclusion that She3 localizes only a specific set of transcripts and that the S . cerevisiae and C . albicans She3 systems transport different mRNAs . Based on the role of She3 in localizing transcripts to the hyphal tip , we next investigated the requirement for She3 in the formation and proper function of hyphae . When grown in liquid serum-containing medium , the she3Δ/she3Δ strain forms germ tubes that are initially indistinguishable from those of the matched wild-type strain ( Figure 3A ) , indicating that She3-regulated RNA transport is not required for the initiation of hyphal growth . However , subtle defects become apparent as the filaments grow . Normal hyphae have parallel sides with no constrictions at septal junctions , and their first septa are formed within the germ tube [28] . By two hours of serum exposure , approximately two-thirds of the she3Δ/she3Δ cells that had initiated germ tube formation failed to form normal hyphae; instead they displayed a range of defects , including constrictions at their septal junctions and uneven filament width ( Figure 3B ) . By the same criteria , only five percent of wild type hyphae were abnormal . These data indicate that She3-mediated RNA transport is not required for germ tube formation , the earliest stage in hyphal formation , but comes into play at later stages of hyphal growth . A more striking defect caused by deletion of the SHE3 gene is observed on filament-inducing solid media ( we use the term filament to include both hyphae and pseudohyphae ) . When grown on YEPD agar with 10% serum or on Spider agar , wild type C . albicans colonies develop a wrinkled central region ( a mixture of yeast , hyphae and pseudohyphae ) , as well as peripheral filaments ( predominantly hyphae ) that invade the agar . she3Δ/she3Δ colonies specifically lack these peripheral filaments; the central wrinkled region is expanded , but otherwise indistinguishable from wild type colonies ( Figure 4 ) . This pronounced phenotype is observed at both 30°C and 37°C , and the identical defect was observed in six independently derived she3Δ/she3Δ strains , representing three different strain backgrounds . In the SN152 background [43] , the she3Δ/she3Δ defect was complemented ( that is , peripheral hyphae were restored ) by re-introduction of the wild type SHE3 gene ( see Methods ) . While other loss-of-function mutations that preferentially affect central or peripheral filaments have been described [44] , none of these completely and selectively eliminates peripheral filaments without affecting the central portion of the colony . In order to visualize the defect caused by deletion of SHE3 in greater detail , we monitored the initial events in the formation of peripheral filaments . Wild type and she3Δ/she3Δ cells were seeded onto thin agar slabs or standard agar plates and colony growth under a cover slip was observed for 48 hours . Colonies from both strains initially grew as yeast and began to filament by 36 hours . Early filaments of wild type colonies , observed at the interface between the agar and the cover slip , were a mixture of hyphae and pseudohyphae . Invasive filaments , which appeared by 48 hours , were predominantly hyphae . In the she3Δ/she3Δ strain , in contrast , hyphae were never observed ( i . e . , all filaments were pseudohyphae ) , and the overall extent of filamentous growth and invasion of the agar was markedly decreased ( Figure 5 ) . These results , taken together , suggest that She3-regulated RNA transport is required for hyphal growth on solid media . It appears that the pseudohyphae of the she3Δ/she3Δ strain are inefficient at invasive growth , and that the absence of peripheral filaments in the she3 null colonies stems from an inability to make invasive hyphae . We next tested whether the defect in invasive growth on solid agar might correlate with a defect in damage to host cells . A she3Δ/she3Δ strain was tested for the ability to damage monolayers of human epithelial and endothelial cells [45] , [46] . While cells lacking SHE3 were able to damage endothelial cells as efficiently as wild type , their ability to damage epithelial cells was reduced by about 40% , a statistically significant difference ( Figure 6A ) . The defect suggests that tip-localization of one or more of the She3-associated transcripts may be required for the physical processes associated with epithelial cell damage or may be involved in sensing this particular niche . The she3 null strain showed normal virulence in a mouse model of disseminated candidiasis ( data not shown ) , suggesting that She3 is not required for this disease model [47] . Based on the sheer number of She3-transported mRNAs and the crucial functions predicted for some of the encoded proteins , one might have predicted that deletion of She3 would exhibit more severe phenotypes than those observed . It seems likely that She3 – mediated mRNA transport is one of several overlapping mechanisms to ensure proper protein localization . For example , the proteins encoded by She3-associated mRNAs could also contain localization signals ( 2 , 16 ) . Alternatively , some of these proteins may retain all or part of their function even if mislocalized . To further explore the role of She3-mediated transport in hyphal development , we analyzed the roles of individual transported mRNAs . We constructed homozygous deletion strains for 33 of the genes encoding transported transcripts and assessed their phenotypes after ten days on Spider agar plates and after 48 hours on Spider agar slabs under a cover slip , as previously performed with the she3Δ/she3Δ strain . Eleven of the 33 mutants displayed colony morphology defects on Spider agar plates , and , among these , nine displayed aberrant filamentation in the early stages of embedded colony growth on Spider slabs . Some of the mutants showed an overall increase in filamentous growth , while some showed an overall decrease ( Figure 7A and 7B ) . We tested the strains with aberrant colony morphology for the ability to form hyphae under strongly inducing conditions; i . e . exposure to serum at an elevated ( 37°C ) temperature . Three strains , those lacking CHT2 , orf19 . 6044 , or orf19 . 267 , displayed obvious defects . The phenotype of the orf19 . 6044 ( MOB2 ortholog ) deletion mutant is consistent with Mob2's established role in polarized growth [48] . The remaining strains showed normal hyphal morphology in liquid serum-containing medium , suggesting that the deleted genes are required for specialized hyphal function but not for hyphal formation per se . In summary , we analyzed deletion mutants corresponding to 33 transported mRNAs . Three strains exhibited severe defects in hyphal formation and an additional eight showed more subtle defects in hyphal growth regulation . These results support the conclusion that transport of specific mRNAs into the hyphal tip cell is needed for proper hyphal development and function .
In this paper , we describe an RNA transport system in C . albicans that localizes specific mRNAs to daughter cells in budding yeast and the tip cells of hyphae . When this RNA transport is inactivated by elimination of She3 ( a component of the transport system ) , hyphae display specific defects , including aberrant growth and decreased capacity to damage an epithelial cell monolayer . We identified mRNAs transported by this system through their tight association with She3 , and we used FISH to show that the transported mRNAs accumulate in yeast buds and in the tips of hyphae in a She3-dependent manner . We believe that this study represents the first description of a set of mRNAs that are specifically localized to hyphal tip cells of a filamentous fungus . Based on direct studies in C . albicans or characterization of orthologous genes in S . cerevisiae , the mRNAs bound by C . albicans She3 are predicted to encode several classes of proteins . Several ( orf19 . 3356 , MSS4 , CDC20 , orf19 . 267 , orf19 . 3071 , orf19 . 5537 , CHT2 , orf19 . 6044 and orf19 . 6705 ) encode proteins that function in mitosis , the cytoskeleton , cell wall dynamics , or cell polarity . Another group of associated mRNAs ( ASH1 , CTA9 , CTA9 , BCR1 , HAC1 , GLN3 ) encode transcriptional regulators . She3 also associates with mRNAs for cell-surface proteins , including predicted GPI anchored proteins ( PGA55 , YWP1 , PGA6 , and PGA54 ) and cell membrane-associated regulators of ion transport ( orf19 . 1582 and orf19 . 5406 ) . Finally , She3-associated RNAs encode known hyphal-specific virulence factors , RBT4 and SAP5 . Taken together , the identities of transported mRNAs suggest that the She3 machinery supports diverse functions , including localization of the basic machinery required for cellular growth and polarity , specification of transcriptional programs in daughter cells and in hyphal tip cells , and differential distribution of cell surface and secreted proteins , some of which function in virulence . We present several lines of evidence that She3-mediated RNA transport , although not required for hyphal formation per se , is required for normal hyphal growth and function . First , she3Δ/she3Δ strains display reduced ability form hyphae and to penetrate solid agar . Second , although she3Δ/she3Δ strains can form hyphal structures in certain conditions , these filaments are morphologically abnormal . Third , a she3Δ/she3Δ strain shows reduced capacity to damage an epithelial cell monolayer . Finally , we constructed and tested deletions of 33 genes whose transcripts are She3-bound . Approximately one third of the individual deletion mutants have filamentation defects on solid medium , and the aberrant morphologies vary considerably among the mutants . As might be expected , none of these strains displayed exactly the same defects as the she3Δ/she3Δ strains , indicating that the she3 mutant phenotype does not reflect the absence of a single transported mRNA in hyphal tip cells . Taken together , these observations support the idea that transport of multiple mRNAs to hyphal tip cells contributes to proper hyphal function . Our analysis of the She system in C . albicans allows for the first direct cross-species comparison of an RNA transport system . A surprising finding from our studies is the minimal apparent overlap between She-associated transcripts in S . cerevisiae and C . albicans: only two genes ( out of 40 in C . albicans and 24 in S . cerevisiae ) are bound in both species . These results suggest that specific mRNAs have moved in and out of the She3-dependent transport system relatively rapidly over evolutionary timescales . Based on existing data , two plausible mechanisms could account for the apparently rapid evolution of mRNAs transported by the She system . In one model , changes in the RNA-binding specificity of the modular She complex could account for this difference . In an alternate model , which we favor , the change in She3 cargo may have arisen via changes in the nucleotide sequences of the transported mRNAs , which have brought new transcripts under She3 regulation . The cis-acting elements that mediate localization of She-associated transcripts in S . cerevisiae have been characterized for a small subset of transported RNAs and are composed of short degenerate sequence motifs , as well as secondary structures that are largely sequence-independent [19] , [39] , [40] , [41] , [42] . It is plausible that , over evolutionary timescales , small sequence changes mediate rapid losses and gains of cargo mRNAs . Such a mechanism is analogous to evolutionary changes in transcription circuitry , where the basic transcriptional machinery and its regulators have been conserved over long timescales , but changes in cis-regulatory sequences have brought new sets of genes in and out of control of ancient regulators [49] , [50] . C . albicans and S . cerevisiae diverged from a common ancestor roughly 200 million years ago , and since that time they have adapted to distinct environmental niches . S . cerevisiae is widely distributed in the environment , whereas C . albicans is restricted to warm-blooded animals . We suggest that the differences in the She3-transported mRNA cargos likely reflect the distinct pressures of each organism's environmental niche .
Strains used in this study are listed in Table 1 and described in greater detail below and in Text S1 . CAI4 , CAF2-1 , SN87 , SN152 , and QMY23 have been described previously [43] , [51] , [52] . C . albicans transformations were performed according to standard lithium acetate methods . For cultivation of C . albicans hyphae , strains were grown to OD 10–12 , then diluted at least tenfold into YEPD containing 10% serum and grown at 37°C , unless otherwise indicated . Two methods were used for deleting orf19 . 5595 ( SHE3 ) . A modified Ura-blaster protocol [53] was used for construction of the heterozygous deletion strain SE6 and homozygous deletion strains SE4 and SE5 . Fusion PCR methods that avoided using the URA3 marker were used to produce the she3 null mutants SE30 and SE32 , in , respectively , SN87 and SN152 backgrounds [43] . For complementation studies , a construct containing SHE3 under the control of its own regulatory sequence was introduced to the region downstream of RPS1 ( orf19 . 3002 ) in SE32 to generate the SHE3-complemented strain SE64 . she3 heterozygote and null mutant strains with the same nutritional markers as SE64 were also generated . Strains SE18 and SE20 in which one copy of ASH1 contained an amino-terminal 6×MYC tag were generated in , respectively , wild type ( CAI4 ) and she3-null backgrounds . One copy of endogenous ASH1 was deleted , and the plasmid p DI-30 [35] , carrying 6MYC-ASH1 , was integrated into the region of the deleted ASH1 allele . A strain in which the single copy of SHE3 was TAP-tagged [37] was constructed using modified fusion PCR methods [43] to produce the SHE3-TAP strain ( SE25 ) , in which a TAP-URA3 cassette was added immediately upstream of the stop codon of the SHE3 allele . Deletion strains corresponding to individual She-associated transcripts were constructed in the SN152 background using fusion PCR methods [43] . Detailed methods of strain construction are provided in Text S1; primers used in strain construction are provided in Table S1 . Immunoprecipitation of She3-RNA complexes from SE25 were adapted from published methods [16] , [20] and are described fully in Text S1 . Briefly , exponentially growing yeast cells or hyphae produced by 30 minutes , one hour , or three hours of serum induction were lysed with glass beads in extraction buffer . Lysates were incubated with IgG-sepharose beads , the immunoprecipitate was released from the beads by cleavage with TEVprotease , and RNA was isolated by phenol-choloroform extraction followed by ethanol precipitation . For a mock RNA immunoprecipitation , the SHE3-TAP parental strain SE6 was subjected to the same methods . Total RNA from the SHE3-TAP strain was harvested from cultures prepared as above and was isolated by a hot phenol protocol [54] . Labeled cDNA was generated from each RNA sample , coupled to fluorescent dyes and hybridized to DNA microarrays essentially as described [38] . Microarray data were quantified using GENEPIX PRO version 3 . 0 or 5 . 0 and were further processed using NOMAD ( http://ucsf-nomad . sourceforge . net/ ) . Processed data were analyzed in Microsoft Excel; filters applied to the data are described in the Results . cDNA samples from generated from total RNA from SE25 and SE6 were directly compared on DNA microarrays using the above methods . Methods were adapted from published protocols [55] and are described in detail in Text S1 . Briefly , for each FISH probe , a digoxigenin-labeled antisense riboprobe was generated by in vitro transcription from a plasmid template containing a portion of the corresponding gene; primers used for template construction are provided in Table S2 . Yeast and hyphal cells were grown as described above , fixed in 5% formaldehyde , and spheroplasted in sorbitol buffer containing zymolyase 100 T . Probe hybridization and signal detection with the HNPP Fluorescent Detection Set ( Roche ) were performed essentially as described [24] . Mounted cells were imaged on the Axiovert-200 ( Carl Zeiss , Thornwood , NY ) . C . albicans strains were grown to approximately OD-12 in YEPD at 30° , diluted 1∶50 into YEPD/20% serum , and incubated with rotation for two hours at 37°C . Cultures were fixed in culture medium plus 4% formaldehyde for one hour , washed and resuspended in PBS , and sonicated in Branson Sonifier 450 for 30 seconds with power setting 1 . 5 , 40% duty cycle . Cells were adhered for ten minutes to cover slips that had been pretreated overnight with 1 mg/ml concanavalin A , washed twice in PBS , then stained with 1 µg/ml fresh Calcofluor White for ten minutes in the dark . Cover slips were washed five times with PBS , and then mounted on glass slides . Agar slabs were prepared by pouring molten agar media between glass plates separated by 1 mm spacers; rectangular pieces of the solidified media were places atop glass slides . The slabs were spread with 10 µl of C . albicans preparations ( exponentially growing C . albicans yeast cultures , diluted to approximately 1 cfu/µl in water ) , overlaid with glass coverslips , and kept in a humid chamber at 30°C . Endothelial and epithelial cell damage assays were performed as previously described [45] , [46] . | Generation of cellular polarity – asymmetry in shape , protein distribution , and/or sub-cellular function – is an essential feature of most eukaryotic cells and underlies such diverse processes as differentiation , mating , nutrient acquisition , and growth . Localization of specific mRNAs is one mechanism through which cells achieve polarity . We describe an RNA transport system in Candida albicans , a fungal pathogen of humans , that grows in both single cell ( budding yeast ) and filamentous ( hyphal and pseudohyphal ) forms . Hyphae are chains of elongated cells that remain attached after cell division and exhibit highly polarized growth at their tips . We show that the C . albicans She3-dependent RNA transport system binds to 40 mRNAs and transports these mRNAs to yeast buds and to the tips of hyphae . Both the transport system itself and many of the genes encoded by transported mRNAs are required for normal growth and function of hyphae . Although the basic transport mechanism appears conserved with that of the model yeast , Saccharomyces cerevisiae , the cargo mRNAs are largely distinct . The apparently rapid evolution of the transported mRNAs probably reflects distinct selective pressures acting on the two organisms . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology/microbial",
"physiology",
"and",
"metabolism",
"infectious",
"diseases/fungal",
"infections",
"cell",
"biology/developmental",
"molecular",
"mechanisms",
"molecular",
"biology/mrna",
"transport",
"and",
"localization",
"evolutionary",
"biology",
"genetics",
"... | 2009 | An RNA Transport System in Candida albicans Regulates Hyphal Morphology and Invasive Growth |
Phototrophic organisms such as cyanobacteria utilize the sun’s energy to convert atmospheric carbon dioxide into organic carbon , resulting in diurnal variations in the cell’s metabolism . Flux balance analysis is a widely accepted constraint-based optimization tool for analyzing growth and metabolism , but it is generally used in a time-invariant manner with no provisions for sequestering different biomass components at different time periods . Here we present CycleSyn , a periodic model of Synechocystis sp . PCC 6803 metabolism that spans a 12-hr light/12-hr dark cycle by segmenting it into 12 Time Point Models ( TPMs ) with a uniform duration of two hours . The developed framework allows for the flow of metabolites across TPMs while inventorying metabolite levels and only allowing for the utilization of currently or previously produced compounds . The 12 TPMs allow for the incorporation of time-dependent constraints that capture the cyclic nature of cellular processes . Imposing bounds on reactions informed by temporally-segmented transcriptomic data enables simulation of phototrophic growth as a single linear programming ( LP ) problem . The solution provides the time varying reaction fluxes over a 24-hour cycle and the accumulation/consumption of metabolites . The diurnal rhythm of metabolic gene expression driven by the circadian clock and its metabolic consequences is explored . Predicted flux and metabolite pools are in line with published studies regarding the temporal organization of phototrophic growth in Synechocystis PCC 6803 paving the way for constructing time-resolved genome-scale models ( GSMs ) for organisms with a circadian clock . In addition , the metabolic reorganization that would be required to enable Synechocystis PCC 6803 to temporally separate photosynthesis from oxygen-sensitive nitrogen fixation is also explored using the developed model formalism .
Flux balance analysis ( FBA ) has become a popular tool to analyze the metabolic function of organisms [1] . FBA assumes the cell is operating at a pseudo steady-state , wherein for each internal metabolite the sum of production fluxes must equal the sum of consumption fluxes . The steady-state assumption hinges upon the requirement that the time constants characterizing metabolic reactions are very rapid compared to the time constant of cell growth [2] . This time flux invariance places tight constraints on the feasible solution space and underpins the explanatory and predictive success of FBA [3–5] . However , for many organisms temporal and periodic variations in metabolism are part of their lifestyle [6] . This is the case for phototrophic organisms whose metabolism is tailored around light availability over a 24-hour cycle . Two distinct phases can be identified here: a light phase that centers around synthesis of metabolic precursors and storage compounds , and a dark phase that consumes those storage compounds to ensure survival in the absence of an energy source [7] . The transition between these two phases is driven by the circadian clock that choreographs the temporal expression of thousands of genes [6] . Highly varying gene expression levels over the 24hr cycle implies that the corresponding metabolic fluxes would also vary significantly and the biomass precursor production be dynamically shaped as the cumulative contribution by metabolism over 24 hours . FBA describes metabolic fluxes as the average over the 24hr period thus missing the opportunity to describe the ( i ) temporally varying nature of metabolism , ( ii ) time dependent inventory and remobilization of metabolites , and ( iii ) the time when different components of biomass are produced . This implies that FBA needs to be augmented so that it can accommodate temporally varying gene transcription information while still permitting the use of the pseudo steady-state assumption , by exploiting the difference in time-scales between metabolic reactions and cell growth . In their natural habitat , cyanobacteria are subject to a diurnal cycle of light and dark , leading to significant shifts and reorganization within their metabolic network . Although several studies , both experimental and computational [8–10] , have helped to illustrate this cyclic cyanobacterial lifestyle , metabolic studies have primarily focused on conditions of constant illumination or heterotrophic growth on externally-supplied carbon sources . Kinetic models of cyanobacterial metabolism can capture the temporal biochemical interactions in the system , but are only available for select subsystems , such as the cyanobacterial circadian clock [11 , 12] , photosystem II [13] , and the Calvin-Benson cycle [14 , 15] . These temporal transitions cannot be described using conventional FBA , and these limitations have been recognized before . Knoop et al . [16] augmented FBA by introducing a time varying biomass composition tailored around light availability . For instance , the ratio for pigments in the biomass reaction was increased two hours before sunrise and storage compound coefficients increased after noon . A set amount of glycogen was supplied to the model for fueling dark respiration instead of transferring the storage compounds generated during light to dark [16] . Optimal temporal allocation of resources has also been employed as a tool to model diurnal lifestyles using an approach called conditional FBA [17 , 18] . Conditional FBA limits the flux through a reaction by accounting for the abundance of enzymes ( or enzyme complexes ) and their catalytic turnover numbers [17 , 18] . Nonlinear constraints are used to maintain periodicity which makes the size of the model grow as the square of the number of time steps being simulated [17 , 18] . The LP problems solved at each iteration may also become ill-conditioned due to the orders of magnitude differences in fluxes . Finally , required inputs such as enzyme turnovers are often hard to determine accurately . CycleSyn alleviates these challenges by recasting diurnal growth as a single linear optimization framework , with solve times of the same order as that of a standard FBA . A 12h light/12h dark cycle is discretized into twelve intervals , each of which abides by the pseudo-steady state hypothesis , to provide twelve snapshots of metabolism ( see Table 1 for a comparison of CycleSyn with other published models of dynamic metabolism ) . By connecting these snapshots by metabolite transfer reactions , CycleSyn also provides insights into metabolite accumulation and consumption which are in line with published literature [19 , 20] . The only model input apart from model stoichiometry and biomass composition is transcriptomic data collected at 2-hour intervals , which is used to throttle back the upper bounds of corresponding reactions . Apart from modeling phototrophic metabolism , CycleSyn can be applied to functions typically served by conventional FBA such as testing in silico knockout mutants or identifying essential reactions/pathways under a varying light regime . In the absence of omics data from mutant strains , algorithms such as RELATCH [21] can be used in conjunction with CycleSyn to predict the effect of gene knockouts . The primary assumption employed there is that perturbed strains minimize relative metabolic changes and increase the capacity of previously active and inactive pathways in order to adapt to perturbations . By employing flux and gene expression data from wild-type strains , Kim et al . were able to successfully predict flux distributions in genetically and/or environmentally perturbed E . coli , S . cerevisiae , and B . subtilis strains . The current text demonstrates the ability of CycleSyn to guide the redesign of temporally-varying metabolism by identifying the metabolic shifts required to incorporate diazotrophy in a phototroph such as Synechocystis sp . PCC 6803 ( hereafter referred to as Synechocystis ) . In this paper , we seek to capture the temporal changes in phototrophic metabolism over a diurnal cycle by modelling a 24-hour day as twelve individual Time-Point Models ( TPMs ) , with each TPM spanning a two-hour period during 12 hours of light and 12 hours of dark . The pseudo-steady state assumption of standard FBA is imposed at every TPM . Metabolite balances are imposed at every TPM though accumulation and/or net consumption of metabolites is allowed . Any metabolite surplus in the cytosol or carboxysome is transferred to the next TPM . Metabolite levels are not allowed to drop below zero implying that all metabolite consumption within a TPM must not exceed the surplus provided by the previous TPM and the amount produced in the current TPM . Reaction flux upper bounds are set in proportion to the temporally varying transcriptomic data . The cascade of TPMs satisfies periodicity constraints by matching the output from the 12th TPM with the input to the 1st one . Comparisons with experimental observations for Synechocystis cultured under a 12h/12h light-dark cycle are used to ascertain biological fidelity . We find that CycleSyn correctly predicted the accumulation of metabolites such as glycogen , the primary storage compound in Synechocystis , and was able to replicate the temporal variations in metabolic pathways as seen in a diurnally cultured Synechocystis . We also found that upon constricting reaction fluxes using temporally segmented transcriptomic data , the primary bottlenecks in wild-type Synechocystis biomass production centered around pyruvate and 2-oxoglutarate metabolisms . Upregulating their production and/or diverting pyruvate flux selectively into the TCA cycle would lead to increased growth , as has been experimentally observed in Synechococcus elongatus PCC 7942 [27] . Subsequently , we used the 24hr model to address the metabolic flux rewiring needed to enable nitrogen fixation in a temporally segregated manner from photosynthesis in Synechocystis . We found that the added energy needed to fuel nitrogen fixation needs to be supplied by an enhanced TCA cycle turnover together with an upregulation of photosynthesis and glycolysis . The genes that need to be upregulated with respect to a non-diazotrophic wild-type Synechocystis are associated with pathways of energy metabolism , so as to meet the higher energy requirements posed by nitrogen fixation and amino-acid production .
The 24-hour model was created starting from the published Synechocystis genome-scale model ( GSM ) iSyn731 [28] as a reaction source , which was updated to include the latest Synechocystis genome annotation from CyanoBase ( http://genome . annotation . jp/cyanobase/Synechocystis ) . Twelve separate GSMs ( each called a Time Point Model or TPM ) ( Fig 1 ) , each spanning a two-hour period starting from the first light time point ( L0-L2 ) to the last dark time point ( D10- D12 ) were linked . Initially , all TPMs are the same except that TPMs 1 through 6 are allowed to take up light as photons and carbon as carbon dioxide whereas TPMs 7 through 12 are not . The maximum CO2 uptake rate was set to 1 . 1 mmol CO2 g-1 dry weight hr-1 for every light TPM , as cyanobacteria are known to not uptake carbon during dark [29–32] and no fixation occurs in the dark due to the lack of photons . This CO2 uptake flux corresponds 0 to 13 mg-1 dry weight hr-1 of carbon [33 , 34] . A basal ATP maintenance demand was also set for every TPM at 10 mmol g-1 dry weight h-1 [28] . The TPMs are connected by the unidirectional forward transfer of metabolites present in the cytosol and carboxysomes , thus only allowing for the consumption of a metabolite in a specific TPM if the metabolite was previously produced or is produced during the current TPM . Any metabolite surplus in the cytosol and carboxysome except photons and protons are transferred to the next TPM ( see Materials and Methods ) . The cyclic topology of the TPMs implies that metabolic flux can go around a closed loop forming a thermodynamically infeasible cycle [35] . To remedy this , the sum of flux through all transfer reactions ( between TPMs ) is minimized using modified parsimonious flux balance analysis ( pFBA ) [36] after constraining the biomass production flux to its theoretical maximum . This is implemented in CycleSyn as an additional model constraint . It should be noted that even though the transfer of energy metabolites such as ATP , NAD ( P ) , and NAD ( P ) H is allowed , CycleSyn results retain their qualitative trends when their transfer fluxes were set to zero . This is because transferring a single storage molecule such as glycogen to satisfy energy demands during the dark phase is preferred by CycleSyn as it is more in line with the model constraint of minimizing the sum of all transfer fluxes . Photosynthesis in the model is coupled to chlorophyll availability by only allowing flux through photosynthesis reactions in a TPM if chlorophyll is present in that TPM , as coupling chlorophyll production to photosynthetic flux places additional demands on chlorophyll synthesis outside of serving as a biomass precursor . This additional demand is corroborated by matching predicted photosynthetic oxygen evolution flux to experimental values ( 1 ) ( see Materials and Methods ) . A single biomass sink placed in the last dark TPM ( i . e . , TMP12 ) sequesters all biomass components in the experimentally measured ratios to model growth , although CycleSyn results were similar when the biomass drain was placed in TPM6 . TPM6 was chosen as a test case as it is known that very little biomass is produced in the dark [37] . It is important to note that the production of different biomass components is apportioned in a non-uniform manner over the twelve TPMs and only in the last TPM are their fluxes combined to form biomass ( Fig 1 ) . The periodicity in metabolism is captured by using transcriptomic data collected over two-hour intervals to constraint reaction fluxes [38] . Specifically , the upper flux bound of a reaction was scaled as a function of its associated gene expression value normalized by the maximum expression over all TPMs , thereby constricting the maximum allowable flux through it . Each reaction’s unscaled flux bounds are determined using only stoichiometric and thermodynamic constraints in order to determine the largest feasible flux range ( see Materials and Methods ) . This approach is often referred to as the valve approach of regulation [39] . The predicted biomass production flux before adding transcriptomic constraints was 0 . 03 hr-1 which corresponds to an average doubling time of 22 . 8 hours [40] . This closely aligns with the experimentally determined doubling time of wild-type Synechocystis under phototrophic conditions , which is approximately 24 hours ( ~0 . 0288 hr-1 ) [28] . Following the scaling of reaction fluxes using their corresponding transcriptomic ratios , the maximum biomass production flux was reduced by a tenth of its original unconstrained value . This corresponds to a doubling time of ~25 hours . Doubling times in the range of 20–40 hours have been seen for diurnally cultured Synechocystis [41] . CycleSyn uses reaction bounds informed from transcriptomic data to distinguish among the individual light and dark TPMs by relatively throttling reaction upper bounds based on gene expression . Implicit to this is the assumption that RNA levels track protein levels . This correlation between gene expression and protein levels has been shown both experimentally and computationally before [42–45] . Notably Zelezniak et al . [46] found that the correlations between gene expression and metabolite concentrations increases when considered against a background of a metabolic network . We assessed the effect of introducing normally-distributed white noise ( within one standard deviation of the mean of the gene expression normalized over time ) in the connection between gene expression and proteins . We found that CycleSyn predictions were within 19 . 28% of their unperturbed flux values , when averaged over all metabolites and all time points . Different light-sensing proteins help mediate external light cues to coordinate inner metabolic processes . Studies in cyanobacteria such as Cyanothece ATCC 51142 ( hereafter referred to as Cyanothece ) have shown that the abundance of many proteins change over the diurnal period [47] . 40 . 3% of these proteins are associated with central metabolism and energy pathways and 18 . 5% were associated with photosynthesis and respiration [47] . A majority of Synechocystis genes also show cycling , with the peak expression of cycling genes being during the transition from day to night , regulating energy supply and carbon metabolism [48] . Here , we examine if the diurnal nature of a cellular process was maintained from the gene to the metabolic level by incorporating gene expression data using the E-Flux method [39] ( see ‘Materials and Methods’ section ) . Using a modified parsimonious flux balance analysis ( pFBA ) [36] to minimize the total sum of all fluxes through the metabolite transfers between each TPM , the transfer flux between TPMs was predicted . The transfer flux for a metabolite from one TPM to the next can be interpreted as the accumulation/consumption of that metabolite in that particular TPM . This allows us to compare the model-predicted metabolite accumulation to the experimentally measured metabolite levels in a diurnally cultured Synechocystis . As the LP has multiple alternative optima , flux variability analysis [49] was used to determine the maximum and minimum possible transfer flux between TPMs and used to construct Fig 2 . This involves minimizing and maximizing the flux through the metabolite transfer reactions while ensuring that the model continues to produce biomass at the maximum value possible and the sum of all transfer fluxes does not exceed the value obtained using the aforementioned modified pFBA . The maximum and minimum transfer flux was taken for every transferring metabolite and summed over to assess the distribution across categories ( see supplementary S6 Table for a list of metabolites and their classifications ) . For metabolite classes such as amino acids , pigments , organic acids , and energy metabolites , there is very little variability between the maximum and minimum transfer flux across TPMs . This is due to the limitations placed on the maximum allowable flux through a reaction using transcriptomics data and the modified pFBA model constraint that further minimizes the sum of all transfer reactions , thus reducing variability . Metabolites are only allowed to accumulate so as to satisfy the demands placed on the system , such as biomass production and ATP maintenance . The amino acid flux profile , governed mainly by glucogenic amino acids such as proline and alanine , is found to be at its maximum during early light with a sharp decline at the transition between light and dark , i . e . between TPM6 and TPM7 . This points towards the role of proline as a carbon and nitrogen reservoir , as is known to occur in Cyanobacteria [50 , 51] . CycleSyn predicts that proline flux feeds into glutamate biosynthesis during the light TPMs by the action of glutamate dehydrogenase . This is corroborated in literature , where glutamate concentrations are known to increase during light in a diurnally cultured Synechocystis [41] and the NADP-specific glutamate dehydrogenase functions only during light in the phototroph C . sorokiniana [52] . The decrease in amino acid transfer flux during the transition from light to dark can be explained by the concurrent increase in accumulation of energy metabolites such as ATP , GTP and CTP . Glucogenic amino acids are degraded towards TCA cycle intermediates such as oxaloactetate , thereby producing energy equivalents in the form of ATP . Glucogenic amino acids levels are known to decrease right after the transition from light to dark incubation in wild-type Synechocystis [20] , alongside a substantial upregulation in genes involved in ATP synthesis [53] . This transition from light to dark is also accompanied by an increase in flux through the oxidative part of the pentose phosphate pathway ( OPPP ) . OPPP is the major pathway of glucose catabolism in heterotrophic and mixotrophic cultures of Synechocystis [54] and is known to be used in conjunction with the Calvin cycle to regulate carbon fixation in autotrophic conditions [54] . The initiation of photosynthesis after the transition from light to dark uses energy compounds such as NADPH , ATP , and Calvin cycle intermediates . As these intermediates are shared with glycolysis , it is expected that respiration during dark time periods be closely linked to the initiation of photosynthesis in diurnally-growing Synechocystis [56] . Glycogen is the primary respiratory substrate in Synechocystis and its degradation is initiated by glycogen phosphorylase transferring orthophosphate to the non-reducing end of the glucose residue in glycogen and releasing glucose-1-phosphate , which feeds into glycolysis . Experiments with Synechocystsis mutants deficient in glycogen phosphorylase ( ΔGlgP ) found that the amount of dark respiration was 25% lower than that in the wild-type , following which the photosynthetic oxygen evolution rate reached its steady-state value at a later time [56] , delineating a temporal dependence between glycolysis during the dark and photosynthesis during light . Our simulations also show accumulation of circulating Calvin cycle intermediates during the dark TPMs . Metabolites such as 3-phosphoglycerate , which is used to regenerate D-ribulose-1 , 5-bisphosphate ( RuBP ) during photosynthesis [57] , and other glycolytic metabolites such as 2-phosphoglycerate and fructose-6-phosphate exhibit this phenomena . These metabolites are fed into glycolysis during the dark and enter the Calvin cycle during the transition from dark to light , suggesting that glycolytic intermediates produced during respiration in the dark are used for regenerating RuBP via the Calvin cycle during the induction of photosynthesis . Interestingly , the total accumulation of RuBP was higher in light than in the dark , implying preferential production and degradation of metabolites guided by the organism’s circadian clock . This is in alignment with cyanobacteria upregulating the oxidative pentose phosphate pathway in the absence of light [58] . Metabolite classes such as fatty acids , porphyrins , nucleotides , and proteins were selectively produced in the light as opposed to the dark TPMs , in agreement with literature . Fatty acid biosynthesis is known to increase with increasing light intensity in Synechocystis [59] and the responsible enzymes involved show increased synthesis during the light in Cyanothece as well [60] . Furthermore , photo-oxidative stress during photosynthesis gives rise to reactive oxygen species and initiates redox signaling . Ansong et al . [61] showed that several proteins involved in fatty acid biosynthesis are redox controlled in Synechococcus elongatus 7002 , including acetyl CoA-carboxylase , which catalyzes the first step of fatty acid biosynthesis . Nucleotide and protein metabolism enzymes were shown to have the highest representation among all redox sensitive proteins , both of which show higher metabolite accumulation in light as opposed to dark in CycleSyn ( Fig 2 ) . Increased protein accumulation during light is also supported by an upregulation in the corresponding genes involved in protein synthesis in Synechocystis [38] . The increase in porphyrins such as heme during the light TPMs is the consequence of the increase in pigment production . Heme and chlorophyll are both tetrapyrrole pigments and hence share a common biosynthetic pathway [62] . Both chlorophyll and heme production is known to be upregulated during light in cyanobacteria [38 , 60] , due to their central role in photosynthesis [63] . Interestingly , terpenes such as lycopene and gamma-carotene are synthesized in the latter half of the day in CycleSyn , rising just before the transition to dark and maintaining these levels throughout the dark TPMs . We sought to investigate the source of this production using Metabolite-metabolite correlation analysis ( MMCA ) [64–67] ( see Fig 3 ) , which assesses metabolite concentration interdependencies using similarity metrics . These dependencies have been used before to explore the temporal organization of metabolic networks [68 , 69] . The analysis herein calculates pair-wise correlation coefficients between time-resolved transfer flux profiles for all metabolites using the non-parametric Spearman Test , employing a two-tailed test for hypothesis testing with a p-value cut-off of 0 . 05 . The transfer flux for every metabolite was normalized with respect to the maximum value recorded across all TPMs and used as a proxy for metabolite concentrations . MMCA is usually employed on experimental datasets of metabolite concentrations , but CycleSyn’s ability to predict metabolite accumulation levels under a FBA paradigm enables us to use MMCA to study possible correlations between metabolite transfer fluxes . MMCA showed that metabolites of the pentose phosphate pathway such as Ribulose-1 , 5-bisphosphate ( RuBP ) and dihydroxyacetone phosphate ( DHAP ) are negatively correlated with terpenes such as gamma-carotene , beta-carotene , and lycopene , the three terpenes showing maximum production during late light ( Fig 3 ) , thus indicating that a rise in the levels of terpenes is associated with a fall in the levels of RuBP and DHAP . A similar phenomenon has been observed earlier by Ershov et al . [70] for a phototrophically growing Synechocystis , where terpenoid biosynthesis was stimulated by the addition of DHAP and other compounds of the pentose phosphate pathway , such as glyceraldehyde-3-phosphate ( G3P ) and D-ribulose 5-phosphate . Isoprenoid synthesis in Synechocystis occurs via the 2-C-methyl-D-erythritol pathway ( MEP pathway ) , which starts with the condensation of glyceraldehyde 3- phosphate ( GA3P ) and pyruvate [71] . Thus , the substrates for terpenoid production are obtained from the metabolite products of photosynthesis such as DHAP and G3P . This also explains the temporal order of terpenoid production , wherein products of photosynthesis need to accumulate in order for the MEP pathway to be active , thus leading to terpene production in the late light , as replicated in CycleSyn . Furthermore , genes associated with the MEP pathway were seen to be upregulated in the dark in a diurnally cultured Synechocystis , such as those corresponding to phytoene dehydrogenase , phytofluene dehydrogenase , and the reaction CTP:2-C-Methyl-D-erythritol 4-phosphate cytidylyltransferase [38] , which catalyzes the conversion of MEP into 4- ( cytidine 5’-diphospho ) -2-C-methyl-D-erythritol and constitutes the second step of the MEP pathway . The metabolite class made up of RNA was found to be largely insensitive to the diurnal changes in metabolism , transfer fluxes ranging between 4 . 345 and 4 . 328 mmol gDW-1 hr-1 . The time invariant nature of RNA production is consistent with a study that determined the total amount of tRNAs is relatively constant over a diurnal cycle in Synechocystsis , with the major RNA variability originating from long transcripts such as 16S rRNA and 23S rRNA [72] which are not captured in the present metabolic reconstruction . Carbon fixation in Synechocystis during light exceeds needs ( luxury uptake [73 , 74] ) so as to catabolize those reserves in the dark to support growth and cellular maintenance . Glycogen is employed as such a reserve in Synechocystis [75 , 76] , providing maintenance energy for cellular functions during dark periods [77] . As seen in Fig 4 and S1 Table , glycogen accumulates during the day and is gradually consumed during the night in CycleSyn . The highest glycogen transfer flux was recorded at the transition between light and dark in our simulations . Fig 4 contrasts the experimentally observed glycogen concentrations with the simulated glycogen accumulation levels . This comparison enables us to approximate the amount of a metabolite shuttled across TPMs after all its production/consumption reactions have taken place , so as to examine its overall temporal dynamics and contrast it with experimental values . Model-predicted glycogen dynamics is in accordance with experimental data , with the total glycogen content increasing gradually during light and reaching its peak level just before the onset of dark . The dark TPMs see a progressive decrease in glycogen as it is utilized as a respiratory substrate [77] . In particular , CycleSyn glycogen accumulation during light matches with the experimental levels seen by Angermayr et al . [41] , but unlike Angermayr et al . we do not see a biphasic decline during late dark , which is also consistent with earlier studies [20 , 38 , 78] . Angermayr et al . [41] attribute the rapid decline in glycogen during the last two hours of dark to increased acetate accumulation and an upregulation of genes encoding the bidirectional NiFe-hydrogenase that is thought to help maintain the redox balance by reducing H+ [79] . CycleSyn did not predict a higher flux through NiFe-hydrogenase during the dark . In order to further ascertain the veracity of the model-predicted glycogen levels , we also compared the glycogen accumulation in the dark TPMs to data from Hanai et al . [20] ( Fig 4 ) . In this study Synechocystis was cultured under a 12hour light/12hour dark cycle and the glycogen concentration ( as nmol per gm fresh weight ( FW ) ) was measured at 0 , 2 , 6 , and 12 hours after the transition to dark ( Fig 4 ) . As seen in Fig 4 , CycleSyn predicted glycogen dynamics during the dark matches that seen by Hanai et al . [20] . Although CycleSyn predicts glycogen accumulation during light and degradation during the dark , the minimum possible transfer flux for glycogen is zero , indicating that other metabolites can serve as additional storage reserves . Isocitrate is found to be such a possible storage metabolite that is accumulated during light and degraded in the dark TPMs . Its catabolism occurs via isocitrate dehydrogenase , encoded by the icd ( slr1289 ) gene which has been found to be upregulated in the dark as compared to light in a diurnally cultured Synechocystis by Saha et al . [38] . Experiments with Synechocystis sp . PCC 6803 impaired in glycogen synthesis have displayed overflow of pyruvate and 2-oxoglutarate [80 , 81] , suggesting that carbon excess is directed preferentially into these compounds in the absence of glycogen . Isocitrate dehydrogenase catabolizes isocitrate to produce 2-oxoglutarate , whose central role in Synechocystis metabolism is discussed below and elucidated through metabolic control analysis . The biomass equation approximates the composition of dry biomass and is used to drain biomass precursors in their physiologically relevant ratios . Fig 5 describes the sum of the maximum and minimum transfer fluxes for classes of biomass precursors over a 24-hour diurnal cycle ( see S1 Table for the list of transfer fluxes of all biomass precursors ) . As seen in Fig 5 and S1 Table , biomass precursors such as carbohydrates and nucleotides are primarily produced during the day and sequestered in the night during the last TPM . Chlorophyll is synthesized during light as is known to occur in cyanobacteria [41] , with these levels remaining constant during the dark TPMs . By constraining reaction flux using transcriptomic data , we identified 110 reactions ( see S2 Table ) with active bounds , i . e . reactions whose flux constraints limit metabolism when the biomass objective function is maximized . This also resulted in a decrease in the biomass flux , which dropped by 10% ( compared to the unconstrained flux case ) corresponding to a doubling time of ~25 hours . As transcriptomic constraints induced a decrease in the biomass production flux , alleviating some or all of these bounds might allow for an increase in the maximal biomass produced . The 110 reactions with active bounds tend to maintain active bounds in multiple TPMs . After adjusting for this reoccurence , 33 unique reactions were identified—one exclusively in dark TPMs and the rest during light TPMs . The reactions with active upper bounds belong mainly to central carbon and amino-acid metabolism , alongside peripheral metabolic pathways such as purine , pyrimidine , aminosugars , and lipid metabolism , and can be broadly linked to pyruvate and 2-oxoglutarate metabolism . Reactions involved in central carbon metabolism such as glycolysis and the pentose phosphate pathway also have active constraints , such as the conversion of 3-phosphoglycerate to 1 , 3-bisphospho-D-glycerate and the reaction between ribose-5-phosphate and D-xylulose5-phosphate to produce glyceraldehyde-3-phosphate and sedoheptulose-7-phosphate . Furthermore , reactions belonging to glucogenic amino acid metabolism are also found to have active reaction bounds . Glucogenic amino acids such as lysine and aspartate yield through catabolism pyruvate or TCA cycle metabolites . As the TCA cycle is the primary source of ATP in cyanobacteria , upregulating these reactions during early light would allow for a greater TCA cycle turnover , leading to more biomass production and hence enhanced growth . The need to increase 2-oxoglutarate production and thus TCA cycle turnover is further evidenced by the constriction of reactions such as L-Phenylalanine:2-oxoglutarate aminotransferase , L-Aspartate:2-oxoglutarate aminotransferase , L-Valine:2-oxoglutarate aminotransferase , and L-Leucine:2-oxoglutarate aminotransferase in the direction of 2-oxoglutarate production . Enhanced pyruvate production has led to greater biomass production in cyanobacteria in an earlier study [27] . Modifying glycolytic pathways and the Calvin Benson cycle in Synechococcus elongatus PCC 7942 to redirect flux towards carbon fixation increased the intracellular pool of pyruvate , which led to about a three-fold increase in growth under both light and dark conditions [27] . In order to further investigate the central roles played by pyruvate and 2-oxoglutarate , we used metabolic control analysis [82] to identify reactions that most affect the biomass production upon a perturbation in their corresponding enzyme levels . A 1% enzyme perturbation was considered and the flux control coefficient ( FCC ) calculated for every reaction using transcript levels as proxies for the enzyme levels [83] ( see Materials and Methods ) . FCCs provide a quantitative measure of the degree of control a particular enzyme exerts on the reaction flux of interest . Interestingly , of all the reactions considered , only two affected the final biomass production flux and only during light TPMs . These included reactions L-Tyrosine:2-oxoglutarate aminotransferase and L-Phenylalanine:2-oxoglutarate aminotransferase with FCCs of 0 . 016 and 0 . 04 , respectively . Both these reactions are controlled by the same set of isozymes in Synechocystis which are not shared by any other reaction and both bidirectional reactions operate selectively in the direction of 2-oxoglutarate synthesis , alluding to the importance of 2-oxoglutarate in Synechocystis growth . The intracellular concentration of 2-oxoglutarate in Synechocystis has been implicated in the regulation of the coordination of carbon and nitrogen metabolism [84] . As Synechocystis lacks the traditional 2-oxoglutarate dehydrogenase complex , 2-oxoglutarate acts as the final carbon skeleton for nitrogen . It is used to sense changes in the cell’s nitrogen status [85] and provides the carbon backbones needed for synthesizing amino acids such as glutamate , glutamine , proline , and arginine biosynthesis via the GS-GOGAT cycle [86] . Nitrogen Fixation was introduced to the model by including the relevant reactions from iCyt773 , the GSM model for Cyanothece [28] , and constraints to ensure that anoxic conditions are maintained during nitrogen fixation ( see Materials and Methods ) , as the nitrogen-fixing enzyme nitrogenase is irreversibly inhibited by oxygen . Transcriptomic data from wild-type Synechocystis [38] was used to restrict reaction flux bounds , so as to identify the reactions that need to be regulated differently in order to fix nitrogen while maintaining growth . S3 Table lists the 672 reactions with flux ranges that change compared to the earlier-described diurnally growing wild-type Synechocystis ( i . e . the feasible flux range associated with these reactions do not overlap with the flux range associated with the wild-type strain ) . 236 reactions ( 107 unique reactions after adjusting for TPM multiple participation ) were upregulated and 436 reactions ( 225 unique ) were downregulated upon the introduction of nitrogen fixation . A total of 166 reactions ( 128 unique ) were found to be essential under diazotrophic conditions , i . e . had strictly positive or strictly negative flux profiles , out of which 44 reactions were non-essential under wild-type conditions . These reactions are indicators of the metabolic alterations required as a result of the introduction of diazotrophy in Synechocystis ( see Fig 6 and S3 Table ) . Reactions with a perturbed flux profile mainly belong to pathways of carbon and amino-acid metabolism , and target the two important modifications required for nitrogenase to function–increased ATP availability and an anaerobic environment . Upregulation of reactions such as L-Aspartic acid:oxygen oxidoreductase help maintain the latter , while the former is addressed by reactions belonging to glycolysis , TCA cycle , and photosynthesis in the light TPMs , i . e . TPMs 1 to 6 , glycogen synthesis , and oxidative pentose phosphate pathway in the dark TPMs , i . e . TPMs 7 to 12 ( Fig 7 and S3 Table ) . As nitrogen fixation is an energy-intensive process , requiring 16 ATP molecules and eight reducing equivalents for every molecule of dinitrogen fixed , this energy is being provided by the coordinated actions of these pathways , while the required reducing equivalents are being supplied by the upregulation of Ferredoxin:NADP+ oxidoreductase reaction , which converts NADP to NADPH while oxidizing ferredoxin . A similar phenomenon is seen in a diurnally cultured Cyanothece , where transcriptomic analysis shows simultaneous upregulation of entire pathways involved in respiration and energy metabolism , such as pentose phosphate pathway , TCA cycle , glycolysis , and amino-acid metabolism in the dark [87] . Cyanothece emerges as a natural comparison for the hypothesized diazotrophic Synechocystis as in order to accommodate both nitrogen fixation and photosynthesis , Cyanothece temporally separates the two incompatible processes [88] . A number of reactions with highly changed flux ranges are similar to the temporal distribution of flux seen in the nitrogen-fixing Cyanothece . For instance , there is an increase in flux in the reactions responsible for glycogen degradation during late dark , which is fed into glycolysis via glucose-1-phosphate , so as to fuel the higher energy demands associated with nitrogen fixation . Similar to the diazotrophic Cyanothece , glycogen degradation in CycleSyn proceeds via glycolysis , the oxidative pentose phosphate pathway , and the TCA cycle so as to provide the cell with ATP , cellular precursors , and pyrimidine nucleotides . Increased respiration of carbohydrate reserves in the dark produces NADPH and succinate , which transfers electrons via NADPH dehydrogenase and succinate dehydrogenase into the plastoquinone pool [90] , towards the terminal electron acceptor . This electron transport due to respiration sets up a proton gradient and drives ATP production . The increased flux towards succinate is evinced by the upregulation of ( S ) -Malate hydrolyase in the dark TPMs which reversibly converts malate into fumarate [47] . A similar upregulation is seen in Cyanothece BG 043511 as well , where an increase in nitrogen fixation at night coincided with a rise in respiratory electron transport [91] . Furthermore , as fixed nitrogen is incorporated via arginine and aspartate metabolic pathways , a number of reactions belonging to these pathways also show upregulation in the dark in a diazotrophic Synechocystis as opposed to the wild-type . Modelling phototrophic growth using constraint-based optimization techniques necessitates modeling contributions beyond conventional flux balance analysis . Since cyanobacteria show strong diurnal rhythms in its lifestyle , translating that phenotype into a metabolic model requires new approaches that enable us to incorporate and replicate those temporal metabolic reorganizations . Diurnal oscillations in Cyanobacteria have been the focus of many studies [6 , 17 , 53] but those have mainly been concentrated on the associated transcriptomic changes or a subset of its entire metabolism . In this work , we bridge that gap by developing a diurnal model of Synechocystis metabolism . We accommodated the cyanobacterial circadian clock and its influence on the underlying metabolic machinery by employing temporally-resolved transcriptomic data . The developed formalism was able to replicate the changes in metabolism observed in Synechocystis over a diel light-dark cycle . It should be noted here that CycleSyn does not assume that metabolite concentrations , enzyme activities , or reaction rates are governed solely by mRNA expression levels . It is well known that the true flux through a reaction depends on the enzyme kinetics and expression , alongside metabolite concentrations ( Michaelis-Menten kinetics ) . The biological rationale behind CycleSyn is that expression data provides measurements of the level of mRNA available for each gene . If there was a limited accumulation of an enzyme in a particular TPM with respect to the others , the ( relative ) level of mRNA can be used as an approximate upper bound for the maximum protein available . This can then be used to constrict the maximum permissible flux through a reaction , effectively reshaping the metabolic flux cone . This enables a systematic extension of flux balance analysis by making use of temporal changes in expression levels to predict the metabolic capacity of Synechocystis over a 24-hour period . The correspondence between transcript levels , reaction fluxes and metabolite levels has been elucidated before [92] where accuracy in predicting the direction of change ( increase/decrease ) in metabolite levels increased by 90% when constraints derived from transcriptomic data were included in the metabolic model of a maize leaf . The 24hr model provides a time-course for all reaction fluxes and metabolite levels . The model predictions are aligned well with several known phenomena in a diurnally-controlled cyanobacterial phototrophic metabolism . We predicted that glycogen was accumulated during light and degraded steadily during dark , as is seen in Synechocystis [38 , 41] . Different pathways were upregulated during the light and dark phases , highlighting the variations in metabolism occurring due to light availability . Glycolysis intermediates produced during respiration in the dark were being used for regenerating RuBP via the Calvin cycle during the induction of photosynthesis . Levels of amino acids produced from glycolytic precursors such as pyruvate and alpha-ketoglutarate decreased at the transition from light to dark incubation in wild-type Synechocystis alongside a substantial upregulation in genes involved in ATP synthesis , as is seen experimentally [20] , [53] . CycleSyn also predicted pyruvate metabolism as a bottleneck in biomass synthesis . Redirecting flux towards pyruvate synthesis can increase carbon fixation and hence biomass formation , as is seen in Synechococcus elongatus PCC 7942 [27] . We also treated Synechocystis as an example cyanobacterium to predict the various metabolic pathways that would need to be regulated in a photosynthetic organism so as to incorporate the two inherently incompatible processes of photosynthesis and nitrogen fixation . The introduction of nitrogen fixation drew parallels from the non-heterocystous cyanobacteria Cyanothece ATCC 51142 , which temporally separates the two antagonistic processes . In doing so it fixes glycogen during the day and uses it as a respiratory product in the dark , thus also achieving the anoxic conditions required for nitrogenase activity . This necessitates a reorganization of the cellular metabolic processes , with dominant flux-carrying pathways differing during the light and dark periods . The model predicted changes in pathways of carbon fixation and amino acid synthesis upon introduction of diazotrophy in Synechocystis . The dark phase in Cyanothece is known to have a high protein turnover , with upregulation of amino-acid biosynthesis pathways due to the increased nitrogen sequestration . Pathways such as arginine and aspartate metabolism were consequently upregulated in the dark , due to the need to sequester the fixed nitrogen . CycleSyn predicted the high-energy demands associated with nitrogen fixation to be met by increased flux through TCA cycle and the pentose phosphate pathway , maintained by higher glycogen synthesis and remobilization . Furthermore , oxygen scavenging reactions such as L-Aspartic acid:oxygen oxidoreductase were upregulated across dark TPMs , due to their oxygen-scavenging role which is required to maintain the anaerobic conditions required for nitrogenase to function . The developed framework enables analyzing a time-variant GSM model while preserving the fundamental time-invariant assumption of conventional flux balance analysis . It improves upon existing techniques of diurnal simulations of metabolism while maintaining a linear programming problem resulting in low computational costs . The formulation can readily be customized to accommodate quantitative measurements of reaction fluxes over a 24-hour cycle . This will also constrain the feasible solution space and thereby improve the precision and accuracy of model predictions [93] . We expect this and similar methods to become instrumental in understanding , analyzing , and predicting temporal metabolic flux variations .
We constructed CycleSyn using the iSyn731 genome-scale metabolic ( GSM ) model for Synechocystis sp . 6803 [28] as a scaffold . iSyn731 was updated to reflect the latest annotations made to the Synechocystis genome as present in CyanoBase ( see supplementary S5 Table ) . Additions include the Entner-Doudoroff pathway , the phosphoketolase pathway , and the light-independent serine biosynthesis pathway [94–98] , among others . Metabolite and reaction IDs were borrowed from ModelSEED [99] wherever possible . From 1 , 156 reactions and 1 , 003 metabolites , the model increased to 1 , 165 reactions and 1 , 008 metabolites . Flux variability analysis was performed on this model to ensure that it is free of any thermodynamically infeasible loops which can carry unbounded flux . The 24-hour model consists of 12 individual Time Point Models ( TPMs ) , each approximating reaction fluxes over a 2-hour period . The first TPM covers the first two hours of the light period ( L0-L2 ) , the second TPM covering the next two hours of light ( L2-L4 ) , with the pattern continuing until TPM12 which contains the last two hours in the dark period ( D10-L0 ) . Each time-point model was made by duplicating reactions ( by appending ‘_tpmX’ to reaction name , where X is the TPM number ranging from 1 to 12 ) and metabolites ( denoted by appending ‘[tpmX]’ to the metabolite name ) in the base model . A single biomass reaction occurs in TPM12 to account for organism growth . All metabolites except photons and protons that are present in the cytosol and carboxysome are transferred unidirectionally from the n to the n+1 TPM using transfer reactions . A transfer reaction j for a metabolite i in a TPM k always operates in the forward direction from TPM k to TPM k+1 such that metabolitei , k→transferj , kmetabolitei , k+1 Photosynthesis is only allowed to occur in a TPM k if it contains the necessary chlorophyll . The chlorophyll balance on a TPM equates the difference between total chlorophyll synthesis and degradation fluxes to the difference between the chlorophyll transfer fluxes that exit and enter the TPM . This is mathematically represented as: ∑vchlorophyllsynthesis , k−∑vchlorophylldegradation , k=vpigmenttransfer , ktok+1−vPigmentTransfer , k−1tok Where vPigment Transfer , k to k+1 refers to the flux through the chlorophyll transfer reaction from TPM k to TPM k+1 . The chlorophyll transfer flux from a TPM k to TPM k+1 represents the cumulative chlorophyll accumulation ( i . e . , from TPM 1 to TPM k ) . We approximate this as a linearly increasing function with respect to time for a single TPM . Hence , the average amount of chlorophyll made in a TPM k can now be calculated as <vchlorophyll , k>=12 ( vPigmentTransfer , ktok+1−vpigmenttransfer , k−1tok ) The constraint levied on the model that couples photosynthesis to chlorophyll availability is expressed as: vPhotosynthesis , k≤12 ( vPigmentTransfer , ktok+1−vpigmenttransfer , k−1tok ) MC ( 1 ) where MC is a constant large enough not to constrain flux through photosynthesis reactions and k the Time Point Model . vPigment Transfer , k to k+1 refers to the flux through the chlorophyll transfer reaction from TPM k to TPM k+1 . This implies that amount of chlorophyll available to carry out photosynthesis is equal to the difference between the amount transferred in to time point k and the amount transported out to point k+1 , divided by two . A value of 1000 for MC predicts photosynthetic oxygen evolution ( estimated by the output flux through the oxygen exchange reaction during the light TPMs ) ranging between 173 to 169 mmol O2 ( gm chlorophyll ) −1 which is in the same order as that of wild-type Synechocystis . A wide range of values have been reported experimentally ranging from 225 mmol O2 ( gm chlorophyll ) −1 ( hr ) −1 [100] to 380 mmol O2 ( gm chlorophyll ) −1 ( hr ) −1 [101] , precluding the matching of a single value . The set of photosynthetic reactions included in this constraint are cytochrome b6/f complex , cytochrome c oxidase cytochrome oxidase bd , Mehler reaction , photosystem I ( plastocyanin ) , photosystem I ( ferrocytochrome ) , photosystem II , and succinate dehydrogenase ( periplasm ) ( see S4 Table for model reaction identifiers and reaction descriptions for reactions belonging to this constrained set ) . The optimization formulation used to determine maximum biomass production flux is maximizevbiomass , TPM12 ( 2 ) subject to ∑j∈JSijkvjk=0 , ∀i∈I , k∈K ( 3 ) vCO2uptake , k≤1 . 1 , k=1 , … , 6 ( 4 ) vPhotonuptake , k≤60 , k=1 , … , 6 ( 5 ) vATPmaintenance , k≥10 , ∀k∈K ( 6 ) vjkLB≤vjk≤vjkUB , ∀j∈J , k∈K ( 7 ) 0≤vj , k≤10 , 000 , ∀j∈JTransfer , k∈K ( 8 ) vPhotosynthesis , k≤12 ( vPigmentTransfer , ktok+1−vpigmenttransfer , k−1tok ) MC , ∀k∈K ( 9 ) where Sijk is the stoichiometric coefficient for metabolite i in reaction j and TPM k , vjkLB and vjkUB are the upper and lower flux bounds for reaction j in TPM k . I , J , and K denote the sets of total metabolites , reactions , and Time Point Models ( TPMs ) , respectively , and JTransfer the set of all transfer reactions . Constraints ( 4 ) and ( 5 ) refer to carbon ( as CO2 ) and photons being supplied to only the light TPMs . The maximum amount of photons supplied to a TPM are such that it is not growth-limiting but at the same time not in excess so as to not trigger light-sensitive reactions . vCO2uptake , k and vPhonton uptake , k represent the carbon and photon uptake reactions , and vATP maintenance , k is the ATP maintenance requirement for a TPM k . Every transfer reaction j in TPM k is constrained to have a non-negative flux by Eq ( 8 ) . As all the individual metabolites are transferred ( only forward ) throughout all TPMs , they may give rise to cycles that can carry unbounded flux . Hence , to prevent this cycling , the sum of transfer fluxes was set to a scalar f that was identified by solving a modified pFBA formulation . Minimizef=∑k∈K∑j∈JTransfervjk ( 10 ) subject to ∑j∈JSijkvjk=0 , ∀i∈I , k∈K vCO2uptake , k≤1 . 1 , k=1 , … , 6 vPhotonuptake , k≤60 , k=1 , … , 6 vATPmaintenance , k≥10 , ∀k∈K vjkLB≤vjk≤vjkUB , ∀j∈J , k∈K 0≤vj , k≤10 , 000 , ∀j∈JTransfer , k∈K vPhotosynthesis , k≤12 ( vPigmentTransfer , ktok+1−vpigmenttransfer , k−1tok ) MC , ∀k∈K vBiomass , TPM12=vBiomass , TPM12max ( 11 ) where JTransfer is the set of all transfer reactions and vBiomass , TPM12max the maximum biomass production flux as determined by ( 2 ) . In photosynthetic organisms such as Synechocystis , a proton gradient is generated during photosynthesis across the thylakoid membrane that drives ATP formation . Transferring this gradient across time points would result in an untenable way for storing energy outside of storage compounds . To this end , protons and photons were not transferred across TPMs . The following optimization model formulation is used to carry out flux variability analysis ( FVA ) : Maximize/Minimizevjk ( 12 ) subject to ∑j∈JSijkvjk=0 , ∀i∈I , k∈K vCO2uptake , k≤1 . 1 , k=1 , . . , 6 vPhotonuptake , k≤60 , k=1 , . . , 6 vATPmaintenance , k≥10 , ∀k∈K vjkLB≤vjk≤vjkUB , ∀j∈J , k∈K vPhotosynthesis , k≤12 ( vPigmentTransfer , ktok+1−vpigmenttransfer , k−1tok ) MC , ∀k∈K 0≤vj , k≤10 , 000 , ∀j∈JTransfer , k∈K vBiomass , TPM12=vBiomass , TPM12max ∑k∈K∑j∈JTransfervjk≤f ( 13 ) CPLEX solver ( version 12 . 1 , IBM ILOG ) was used in the GAMS ( version 23 . 3 . 3 , GAMS Development Corporation ) environment for solving all optimization models . All computations were carried out on dual 10-core and 12-core Intel Xeon E5-2680 and Intel Xeon E7-4830 quad 10-core processors that are the part of the ACI cluster of High Performance Computing Group of The Pennsylvania State University . Numerical scaling issues were not observed when solving CycleSyn . | Phototrophic organisms such as cyanobacteria harvest the sun’s energy to convert atmospheric CO2 into organic carbon , due to which their metabolism is heavily influenced by light availability . The strongly diurnal nature of their metabolism is reflected in the presence of two distinct metabolic phases–a light-dependent anabolic phase tailored around the synthesis of storage compounds and metabolic precursors and a light-absent catabolic period that metabolizes the previously manufactured compounds to release energy in the absence of an external energy source . Due to these considerations , the analysis of phototrophic growth using constraint-based optimization methods is insufficient and needs to be extended beyond time-invariant descriptions . Here , we introduce CycleSyn which models the periodic nature of metabolism in Synechocystis sp . PCC 6803 . Our approach enables us to account for temporal metabolic shifts tailored around light availability while still allowing for the use of the pseudo steady-state assumption used in conventional flux balance analysis . This is achieved by exploiting the large difference in time-scales between metabolic reactions and cell growth . We first validate the biological fidelity of CycleSyn predictions by comparing them to experimental observations for a diurnally cultured Synechocystis sp . PCC 6803 and to identify the major temporal variations in its metabolic processes . Next , we demonstrate the ability of CycleSyn to describe a temporally-varying metabolism by introducing diazotrophy in Synechocystis and evaluating the genes that need to be upregulated/downregulated to enable nitrogen fixation in a photosynthetic organism . Our study lays the foundation for subsequent analysis of systems with temporal variations in metabolism using a constraint-based optimization approach . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] | [
"light",
"plant",
"physiology",
"electromagnetic",
"radiation",
"plant",
"science",
"metabolites",
"genome",
"analysis",
"photosynthesis",
"bacteria",
"genomics",
"metabolic",
"pathways",
"cyanobacteria",
"physics",
"biochemistry",
"plant",
"biochemistry",
"nitrogen",
"fixa... | 2019 | A diurnal flux balance model of Synechocystis sp. PCC 6803 metabolism |
CD8 T cells are necessary for the elimination of intracellular pathogens , but during chronic viral infections , CD8 T cells become exhausted and unable to control the persistent infection . Programmed cell death-1 ( PD-1 ) blockade therapies have been shown to improve CD8 T cell responses during chronic viral infections . These therapies have been licensed to treat cancers in humans , but they have not yet been licensed to treat chronic viral infections because limited benefit is seen in pre-clinical animal models of chronic infection . In the present study , we investigated whether TLR4 triggering could improve PD-1 therapy during a chronic viral infection . Using the model of chronic lymphocytic choriomeningitis virus ( LCMV ) infection in mice , we show that TLR4 triggering with sublethal doses of lipopolysaccharide ( LPS ) followed by PD-1 blockade results in superior improvement in circulating virus-specific CD8 T cell responses , relative to PD-1 blockade alone . Moreover , we show that the synergy between LPS and PD-1 blockade is dependent on B7 costimulation and mediated by a dendritic cell ( DC ) intrinsic mechanism . Systemic LPS administration may have safety concerns , motivating us to devise a safer regimen . We show that ex vivo activation of DCs with LPS , followed by adoptive DC transfer , results in a similar potentiation of PD-1 therapy without inducing wasting disease . In summary , our data demonstrate a previously unidentified role for LPS/TLR4 signaling in modulating the host response to PD-1 therapy . These findings may be important for developing novel checkpoint therapies against chronic viral infection .
CD8 T cells are critical for controlling intracellular infections , but during chronic viral infections , CD8 T cells undergo functional exhaustion . Immune checkpoint blockade therapies can restore the function and proliferative capacity of exhausted CD8 T cells . In particular , PD-1 blockade therapy is now licensed to treat human cancers . This therapy was initially found to improve exhausted CD8 T cells in the chronic lymphocytic choriomeningitis virus ( LCMV ) infection model in mice , and subsequently , it was demonstrated to also improve exhausted CD8 T cells in other models of chronic infection [1–6] . PD-1 therapies targeting either the receptor ( PD-1 ) or the ligand ( PD-L1 ) can partially improve exhausted CD8 T cells , but these therapies have clinical limits that are not fully understood . In particular , their limited efficacy in models of chronic viral infection has precluded their licensing for treating chronic viral infections , and significant efforts are aimed at improving efficacy using combined regimens [7–12] . Interestingly , it has been suggested that certain products of the microbiota can improve clinical responses to PD-1 therapy [13 , 14] , but the specific microbial products that underpin this positive effect remain unknown . There is a growing interest in modulating innate immune responses to improve immune checkpoint therapies . For example , TLR9 activation can improve cancer immunotherapy [15 , 16] . Another study showed that stimulation of the STING pathway , which senses cytosolic dinucleotides , can improve PD-1 therapy during cancer [17] . However , it is unknown if TLR4 , a potent activator of innate immune responses , affects PD-1 therapy during chronic viral infection . We first evaluated whether lipopolysaccharide ( LPS ) , a natural component of the Gram-negative microbiome , was able to affect the host response to PD-1 therapy during a chronic LCMV infection in mice . Treatment of chronically infected mice with LPS alone did not rescue exhausted CD8 T cells . However , combined treatment with LPS and PD-1 therapy resulted in one of the most impressive synergistic effects that we have ever observed . These results identify for the first time a specific microbial product that augments the efficacy of PD-1 immunotherapy . Systemic LPS administration has obvious safety concerns that may preclude its clinical use , but we demonstrate that adoptive transfer of ex vivo activated DCs also synergizes with PD-1 therapy without inducing wasting disease , suggesting the potential translatability of our findings . Overall , we demonstrate novel strategies to harness the TLR4 pathway to improve PD-1 therapy during chronic viral infection .
Prior studies have shown that LPS improves memory CD8 T cell responses [18] , but whether LPS improves exhausted CD8 T cell responses remains unknown . To answer this simple question , we utilized the model of lifelong LCMV infection in mice ( LCMV Cl-13 ) . At day 45 post-infection , we treated mice with sublethal doses of LPS administered throughout a PD-L1 blockade therapy , and after 15 days , mice were sacrificed to analyze total activated CD8 T cells , as well as virus-specific CD8 T cells ( Fig 1A ) . As expected , LPS alone induced a significant increase in total activated CD8 T cells ( Fig 1B ) , consistent with its known adjuvant effect , but it did not increase virus-specific CD8 T cells ( Fig 1C ) . These data demonstrate that LPS alone does not exert adjuvant effects on exhausted virus-specific CD8 T cells that sense persistent viral antigen . However , LPS rendered exhausted virus-specific CD8 T cells more responsive to PD-1 therapy . Combining LPS with PD-L1 blockade resulted in synergistic expansion of DbGP276+ virus-specific CD8 T cells in blood ( Fig 1C ) . This unprecedented synergy was also evident in tissues ( Fig 1D–1G ) , and was also observed for other responses , such as DbGP33-41 ( Fig 1H ) . Virus-specific CD8 T cells also exhibited more significant functional improvement ( Fig 2A ) , enhanced expression of granzyme B ( Fig 2B ) and the proliferation marker Ki67 ( Fig 2C ) , and showed increased survival following combined therapy ( Fig 2D and 2E ) . In addition , the combined therapy resulted in improved antiviral control in sera and tissues relative to PD-L1 blockade alone ( Fig 2F–2H ) . We also evaluated long-term viremia in separate experiments and we observed improved virologic control for several weeks ( Fig 2I ) . Although there were profound effects on virus-specific CD8 T cell responses , virus-specific antibody responses were not affected by the therapy ( Fig 3 ) . Taken together , these results demonstrate that bacterial LPS renders exhausted CD8 T cells more responsive to PD-L1 blockade therapy . To understand the mechanism of how LPS renders exhausted CD8 T cells more responsive to PD-1 therapy , we performed gene expression analyses . At day 15 post-treatment , virus-specific CD8 T cells were FACS-sorted and used for RNA-Seq analyses ( Fig 4A and 4B ) . Principal Component Analyses ( PCA ) showed differential clustering , suggesting differences in gene expression following combined treatment ( Fig 4C ) . Heat map analyses showed the granzyme B gene ( Gzmb ) as one of the top upregulated genes in the combined treatment group , consistent with our flow cytometric data from Fig 2B . Revigo analyses showed enrichment in metabolic , cell division , and immune processes , such as cell killing and defense response , in the combined treatment ( S1 Fig ) . In particular , genes that are normally induced by interferon type I signaling ( IFN-I ) ( Fig 4E and 4F ) and CD28 costimulation ( Fig 4G and 4H ) were highly enriched in virus-specific CD8 T cells in the combined treatment by Gene Set Enrichment Analyses ( GSEA ) and radar plot analyses . At the genome-wide level , the two most differentially enriched pathways were CTLA-4 and CD28 signaling by Ingenuity Pathway Analyses ( IPA ) ( S2 Fig ) and Molecular Activity Predictor ( MAP ) Analyses ( S3 Fig ) . These data indicated that CTLA-4 and CD28 signaling were primarily triggered on CD8 T cells of mice receiving combined LPS and PD-L1 blockade . It is important to highlight that the CTLA-4 and CD28 receptors on CD8 T cells are triggered by the same B7 . 1/B7 . 2 molecules expressed on antigen presenting cells ( APCs ) . Our transcriptional data suggested a possible a role for IFN-I signaling and B7/CD28 costimulation in promoting the synergy between LPS and PD-1 therapy . Consistent with the gene expression data , LPS treatment in chronically infected mice resulted in high levels of IFN-I in sera ( Fig 5A ) . Immune activation following TLR4 triggering is thought to be dependent on IFN-I responses [19–21] , leading us to hypothesize that IFN-I responses mediated the synergistic effect of LPS on PD-1 therapy . To interrogate a possible role for IFN-I responses , we treated chronically infected mice with an antibody that blocked IFN-I responses , administered together with LPS and PD-L1 blocking antibodies ( Fig 5B ) . We utilized an antibody that binds to interferon α/β receptor subunit 1 ( IFNAR1 ) and precludes its binding to interferons α/β , abrogating downstream IFN-I signaling [22–26] . This is a widely characterized antibody that has been previously shown to reduce the expression of interferon-stimulated genes ( ISGs ) during chronic viral infection in vivo [22] . We also confirmed in vivo blockade of the IFNAR1 receptor by this antibody clone ( MAR1-5A3 ) ( Fig 5C ) . Moreover , we confirmed that this antibody abrogated IFNα-driven PD-L1/MHC-I upregulation in vitro using tumor cell lines ( S4 Fig ) , confirming that this antibody precludes IFN-I signaling . However , contrary to our hypothesis , IFN-I blockade with this antibody did not abrogate the synergistic effect of LPS and PD-L1 blockade , suggesting an IFN-I independent mechanism ( Fig 5D and 5E ) . We previously demonstrated that T regulatory cell ( Treg ) ablation induces a similar synergistic effect on PD-1 therapy [27] . Tregs suppress CD8 T cells via an APC-dependent mechanism; CTLA-4 molecules on Tregs bind to and remove B7 molecules on APCs , a process referred to as trans-endocytosis [27–29] . High PD-1 expression on Tregs is especially associated with increased Treg suppressive function [30] . We hypothesized that the synergistic effect of LPS on PD-1 therapy was associated with attenuated Treg responses . However , LPS did not attenuate Treg responses in terms of their frequency ( Fig 5F ) or expression of inhibitory receptors associated with Treg suppressive function [29 , 31] ( Fig 5G ) . Taken together , the synergy between LPS and PD-1 therapy was likely not caused by downregulation of Treg responses . It must be noted , however , that PD-L1 blockade itself increased Treg frequencies and expression of inhibitory markers on Tregs ( Fig 5G ) . During normal conditions , LPS is known to induce the expression of costimulatory molecules on DCs [32 , 33] . During chronic infection , however , DC function is severely impaired [34–38] , and it is unclear if TLR4 activation with LPS can override DC dysfunction and induce the expression of costimulatory molecules on DCs . To answer this question , we phenotyped DCs 24 hours after treatment of chronically infected mice with LPS . This treatment did not increase DC numbers in chronically infected mice ( S5A Fig ) , but it induced upregulation of MHC molecules on DCs from spleen ( S5B and S5C Fig ) . LPS also resulted in significant upregulation of B7 . 1 and B7 . 2 molecules ( S5D and S5E Fig ) , suggesting that LPS improved DC costimulatory function during chronic viral infection . However , other TLR4 agonists ( Neoseptin-3 and MPLA ) and TLR2 agonists ( LAM ) did not improve costimulatory molecule expression during chronic viral infection ( S5F Fig ) . Note that LPS , Neoseptin-3 and MPLA are TLR4 agonists , but these molecules are biologically distinct . In particular , LPS has been shown to be substantially more pro-inflammatory and induce distinct types of downstream responses [18 , 39–41] . Notwithstanding the upregulation of costimulatory B7 molecules , LPS treatment also upregulated inhibitory PD-L1 molecules on DCs , suggesting a negative feedback loop ( S5G and S5H Fig ) . A similar pattern of dual B7 and PD-L1 upregulation was observed on other APC subsets in chronically infected mice ( S6 Fig ) . Dual upregulation of B7 and PD-L1 was also observed in DCs from naïve mice ( S7 Fig ) . Overall , our data demonstrate that LPS has more profound phenotypic effects on DCs , relative to other TLR4 agonists . B7 upregulation on APCs by flow cytometric analyses and enrichment in CD28-driven genes on virus-specific CD8 T cells by gene expression analyses led us to hypothesize a potential role for B7/CD28 costimulation . To test this hypothesis , we treated chronically infected mice with LPS , PD-L1 blocking antibodies , and B7 . 1/B7 . 2 blocking antibodies ( Fig 6A ) . We utilized anti-B7 antibodies that have been previously shown to block the B7/CD28 costimulatory pathway [42] . Interestingly , blockade of B7/CD28 costimulation abrogated the synergy between LPS and PD-L1 blockade in terms of CD8 T cell rescue ( Fig 6B ) and viral control ( Fig 6C ) , demonstrating a B7/CD28 costimulation-dependent mechanism . It is important to clarify that B7 blockade did not affect APC numbers or subset distribution in spleen , but PD-1 therapy significantly reduced the frequencies of myeloid DCs ( Fig 6D–6G ) . To evaluate whether DCs mediated the synergistic effects of LPS on PD-1 therapy , we generated bone marrow derived DCs from infected mice , and stimulated these cells ex vivo with LPS for 24 hours , followed by extensive washing to remove LPS . We adoptively transferred 107 DCs into chronically infected mice , followed by PD-L1 blockade ( Fig 7A ) . Injection of LPS-activated DCs resulted in greater improvement of CD8 T cells , relative to PD-L1 blockade alone ( Fig 7B ) . Accordingly , antiviral control was also improved by this combined therapy ( Fig 7C ) . No effect was observed by transferring LPS-activated DCs alone ( S8 Fig ) . Altogether , these data demonstrate a DC-intrinsic mechanism for the synergy . TLR4 is a receptor for LPS , but various intracellular molecules can also sense LPS and mediate potent inflammatory responses [43–45] . To ascertain if TLR4 was critical for the potentiation of PD-1 therapy by LPS , we repeated the above experiment , but using Tlr4-/- DCs ( Fig 7D ) . However , Tlr4-/- DCs did not improve PD-1 therapy ( Fig 7E and 7F ) . We compared the phenotype of wild type and Tlr4-/- DC responses following LPS stimulation . Wild type DCs exhibited profound phenotypic and morphological changes upon LPS stimulation , but this was not observed in Tlr4-/- DCs ( S9A and S9B Fig ) . Taken together , these data indicate that the synergistic effect of LPS is mediated strictly by a TLR4-dependent mechanism and not by alternative LPS sensors . We showed that systemic LPS administration improves PD-1 therapy , but systemic LPS administration resulted in wasting disease . Even at our low sublethal dose ( 25 μg ) , mice exhibited hunched posture and lethargy within 24 hours of systemic LPS treatment . In addition , systemic LPS administration resulted in significant weight loss ( S10A Fig ) . On the other hand , the specific transfer of DCs activated ex vivo with LPS was safer and did not induce overt wasting ( S10B Fig ) . These data demonstrate that adoptive transfer of LPS-activated DCs was substantially safer compared to systemic LPS administration . The prior experiments demonstrate that TLR4/LPS signaling improves the efficacy of PD-1 therapy , but an unanswered question is if TLR4 is mechanistically required for PD-1 therapy . To answer this question , we infected wild type or Tlr4-/- mice with chronic LCMV , and at day 45 post-infection , mice received PD-L1 blocking antibodies ( S11A Fig ) . Both Tlr4-/- and wild type mice exhibited improvement in CD8 T cells ( S11B Fig ) and enhanced viral control ( S11C Fig ) following PD-L1 blockade , demonstrating that TLR4 is dispensable during PD-L1 blockade therapy . We then interrogated whether LPS could improve PD-1 therapy in Tlr4-/- mice . Importantly , we demonstrate that LPS treatment in Tlr4-/- mice does not improve PD-1 therapy ( S12 Fig ) . Collectively , these experiments highlight an important distinction: The TLR4 pathway can improve PD-1 therapy , but the TLR4 pathway is not mechanistically required for PD-1 therapy to rescue CD8 T cells .
CD8 T cell exhaustion is a hallmark of chronic viral infections , such as HIV and HCV , which kill millions of people every year . CD8 T cells are critical for controlling chronic viral infections , but they become exhausted and unable to clear the persistent antigen due to upregulation of inhibitory molecules , including PD-1 . Blockade of either the receptor ( PD-1 ) or the ligand ( PD-L1 ) can rescue exhausted CD8 T cells during chronic infection , but these therapies have clinical limitations that are not well-understood . Recent reports using cancer models have shown intriguing associations between gut microbiota and the host response to PD-1 therapy , but the specific microbial products that improve responses remain unknown [13 , 14] . Our data using the chronic LCMV model demonstrate a potent synergism between bacterial LPS ( a ubiquitous component of the microbiome ) and PD-1 therapy . Treatment of mice with sublethal doses of LPS followed by PD-L1 blockade resulted in a more striking CD8 T cell rescue relative to PD-L1 blockade alone , identifying for the first time a specific bacterial product that affects PD-1 therapy . It is important to highlight that the effects of combined LPS and PD-1 therapy are not additive , but synergistic , as LPS alone did not induce rescue of exhausted CD8 T cell responses . Such synergy was surprising and difficult to theoretically predict based on prior studies with LPS treatment alone . Systemic LPS administration potentiated PD-1 therapy , but this adjuvant also caused excessive weight loss and inflammation . In light of this , we aimed to develop a safer approach by stimulating DCs ex vivo with LPS , followed by extensive washing to remove LPS prior to adoptive transfer . LPS-activated DCs also synergized with PD-1 therapy and were substantially safer . DC immunotherapy has been used previously in clinical trials to revert CD8 T cell exhaustion , but it has not shown significant clinical benefits [46–49] . Our studies are novel , because we show for the first time that DC immunotherapy using LPS-activated DCs improves the efficacy of PD-1 therapy . LPS is one of the most immunogenic adjuvants ever discovered , and LPS treatment induces potent activation of naïve and memory CD8 T cells [18 , 50 , 51] , but the effect of LPS on exhausted CD8 T cells has remained understudied . Our initial prediction , based on those prior reports , was that LPS alone would rescue exhausted CD8 T cells , but this was not the case . Such result could be explained by the PD-1/PD-L1 pathway . LPS not only upregulates costimulatory B7 , but it also upregulates inhibitory PD-L1 , and since exhausted CD8 T cells overexpress PD-1 , they are more susceptible to PD-1/PD-L1 mediated inhibition relative to naïve or memory CD8 T cells . Taken together , both costimulatory and inhibitory ligands are co-regulated by LPS/TLR4 signaling , and thus , LPS/TLR4 signaling can only improve exhausted CD8 T cells when inhibitory PD-L1 signals are concomitantly blocked . B7/CD28 costimulation is required for CD8 T cell rescue after PD-1 blockade [42 , 52] , but during chronic viral infection , B7 expression by DCs is limited , resulting in decreased costimulation of exhausted CD8 T cells [27] . Therefore , B7/CD28 costimulation is a critical rate-limiting step that determines the success of PD-1 therapies , and we demonstrate that it is plausible to reinforce B7/CD28 costimulation during chronic viral infection by triggering TLR4 with LPS , rendering exhausted CD8 T cells more responsive to PD-1 therapy . Altogether , we demonstrate that the synergy between microbial LPS and PD-1 therapy is dependent on TLR4 signaling and B7/CD28 costimulation . We also demonstrate that this synergy is at least partially due to DC-intrinsic effects , but we cannot rule out the contribution of other APC subsets . Moreover , we show that the synergy is not dependent on IFN-I responses and does not seem to be caused by attenuated Treg responses . A critical mechanistic insight from our studies is that the TLR4 pathway improves PD-1 therapy , but it is not itself required for PD-1 therapy . This is an important distinction , since a fraction of humans has genetic deficiencies in the TLR4 pathway , which results in a plethora of opportunistic infections [53 , 54] . Our studies suggest that PD-1 therapy could still be effective in these patients . TLR4 triggering with LPS also upregulated MHC on DCs , but dissecting the role of MHC is technically difficult , since exhausted CD8 T cells require continuous TCR recognition to persist in vivo , a process colloquially referred to as “antigen-addiction” [55] . We also reason that during chronic viral infection , CD8 T cells are not critically regulated by limited MHC/TCR interactions , since the persistent antigen is expressed and presented ubiquitously . The discovery of PD-1 therapies to rescue exhausted virus-specific CD8 T cells was initially made using the chronic LCMV model [7] . PD-1 therapies were later demonstrated to also improve exhausted CD8 T cells in other models of chronic viral infection [1–6] . However , PD-1 therapies have not been yet licensed to treat chronic viral infection , because of their limited capacity to induce clinically relevant antiviral control . In light of this , our data make a compelling case for evaluating whether TLR4 signaling can improve PD-1 therapy in other models of chronic viral infection . Interestingly , low LPS levels can be detected in plasma under normal conditions , and the translocation of LPS from the intestinal microbiota to the circulation increases during chronic infection [56–58] . Moreover , injection of low LPS doses has been shown to be safe in humans , suggesting that there is a threshold of tolerance to this naturally occurring molecule , with the maximum tolerated dose in humans shown to be ~4 ng/kg [59 , 60] . Future studies will identify whether this safe dose of LPS is still sufficient to improve PD-1 therapy in humans , and whether plasma or intestinal LPS levels could serve as biomarkers to predict the efficacy of PD-1 therapy . Regimens that reinforce B7/CD28 costimulation on exhausted CD8 T cells using CD28 agonistic antibodies or recombinant B7 proteins may also be of interest to improve PD-1 therapies . In summary , our study has various novel findings . We show that exhausted CD8 T cells cannot be re-activated by LPS alone , but LPS sensitizes exhausted CD8 T cells to respond more vigorously to PD-1 therapy . We also show for the first time a specific microbial product that augments the efficacy of PD-1 therapy . These findings could have important implications for the development of novel treatments for chronic viral infections , and for understanding how natural components of the microbiome affect responses to immune checkpoint therapies .
6-8-week old female and male C57BL/6 wild type mice were used in all experiments . We also used Tlr4-/- mice in the C57BL/6 background . All mice were from Jackson laboratories . All infections were intravenous ( i . v . ) via the lateral tail vein and using a mouse restrainer . Mice were infected with 2x106 PFU of LCMV Cl-13 . Before starting treatments , we randomized mice in terms of viremia and number of virus-specific ( DbGP276+ ) cells in PBMCs . All Antibodies for in vivo treatments were purchased from BioXCell , and were diluted in sterile PBS . To induce lifelong uncontrolled multiorgan infection , CD4 T cells were depleted at the time of infection with 500 μg of a CD4 depleting antibody ( GK1 . 5 ) administered intraperitoneally ( i . p . ) , as described previously [61] . PD-L1 blocking antibodies ( 10F . 9G2 ) were administered i . p . at 200 μg , every three days , five times , as previously shown [7] . B7 . 1 and B7 . 2 blocking antibodies ( 16-10A1 and GL-1 , respectively ) were administered at 500 μg ( 250 μg each ) , every three days , five times . IFNAR1 blocking antibodies ( MAR1-5A3 ) were administered at 500 μg , every three days , five times . This MAR1-5A3 antibody binds to interferon α/β receptor subunit 1 ( IFNAR1 ) and blocks binding to interferons α/β , abrogating the induction of ISGs in vivo [22–26] . IgG isotype controls were used in all experiments . The first dose of B7 and IFNAR1 blocking antibodies were administered at least 6 hours before the first injection of LPS and PD-L1 blocking antibodies . Sublethal LPS doses were administered i . p . at 25 μg per mouse , on days 0 and 7 of the PD-L1 blockade regimen . Other TLR4 agonists were also administrated at 25 μg per mouse , DCs were harvested and activated in vitro as shown before [62 , 63] . In brief , bone marrow cells from infection-matched mice were cultured for 5 days in GM-CSF ( Sigma ) at 20 ng/mL , and stimulated for 1 day with 100 ng/mL of E . coli derived LPS ( Sigma ) . DCs were washed 5 times prior to injection to remove traces of LPS , and 107 DCs were injected i . v . at day 0 and 7 of the PD-L1 blockade regimen . Quantification of LCMV titers was performed on Vero E6 cell monolayers as previously described [64] . In brief , Vero E6 cells ( ATCC ) were seeded onto 6-well plates , and once they reached ~95% confluency , the media was removed and 200 μL of serial viral dilutions were slowly pipetted on top of the monolayers . Plates were rocked every 10 min in a 37°C , 5% CO2 incubator . After 1 hr , 200 μL of media was aspirated out of each well , and the monolayers were overlaid with a 1:1 mixture of 2x199 media and 1% agarose . After 4 days of culture in a 37°C , 5% CO2 incubator , a second overlay was added , consisting of a 1:1 solution of 2x199 media and 1% agarose and 1:50 of neutral red dye . Overlay was removed with forceps on day 5 and plaques were counted using a transluminator . All mouse experiments were performed with approval of the NU Institutional Animal Care and Use Committee ( IACUC ) . ELISA to measure LCMV‐specific IgG responses in sera was performed using lysates of BHK‐21 cells infected with LCMV Cl-13 . ELISA plates ( Nunc MaxiSorp , 439454 ) were coated with infected BHK-21 lysates for 48 hr at room temperature . Twelve serial 30-fold dilutions of sera were added onto each well , followed by incubation with goat anti-mouse IgG HRP ( SouthernBiotech , 1030–05 ) . Sure Blue TMB peroxidase ( KLP , 52-00-03 ) was added onto each well , and after 8 minutes , TMB Stop solution ( KLP , 50-85-06 ) was added . ELISA plates were immediately read at 490 nm in a Spectramax 384 plus ( Molecular Devices ) . Single cell suspensions were obtained from PBMCs and various tissues as previously described [65] . Live cells were gated using Live/Dead fixable dead cell stain ( Invitrogen ) . LCMV MHC class I tetramers were obtained from the NIH tetramer facility ( Emory University ) . Cells were stained with anti- CD8α ( 53–6 . 7 ) , -CD44 ( IM7 ) , -Granzyme B ( MHGB04 ) , -Ki67 ( B56 ) , -PD-L1 ( MIH5 ) , -B7 . 1 ( 16-10A1 ) , -B7 . 2 ( GL-1 ) . Anti-mouse flow cytometry antibodies were purchased from BD Pharmingen , except for CD44 ( Biolegend ) and Granzyme B ( Invitrogen ) . Flow cytometry samples were acquired with a Becton Dickinson LSRII and analyzed using FlowJo ( Treestar ) . For confocal microscopy , spleen OCT sections were stained with rat anti-mouse PD-L1 ( 10F . 9G2 ) at 1:200 dilution from a 2 mg/mL stock . Slides were then stained with a secondary anti-rat IgG antibody conjugated to Cy3 . Microscopy slides were acquired using an Evos digital inverted microscope ( Advanced Microscopy Group ) . Gene expression profiling was performed as described previously [66 , 67] . In brief , splenic CD8 T cells were MACS-sorted at day 15 after treatment , using a MACS negative selection kit ( STEMCELL ) . Purified CD8 T cells were stained with DbGP276 tetramer , live dead stain , and flow cytometry antibodies for CD8 and the CD44 activation marker . ~20 , 000 live , CD8+ , CD44+ , DbGP276+ cells were FACS-sorted to approximately 97% purity using a FACS Aria ( BD Biosciences ) . FACS-sorted cells were collected in 10% FBS RPMI , and were then spun at 2000 rpm for 10 minutes at 4°C . The supernatant was aspirated , and cell pellets were resuspended in 1 mL of TRIzol ( Life Sciences ) inside a fume hood . All samples were stored at -80°C . The next day , RNA extraction was performed using the RNAdvance Tissue Isolation kit ( Agencourt ) on a plate magnet , following the manufacturer’s instructions . RNA quality assessment and HiSeq sequencing ( Illumina ) were performed at the NUSeq core at Northwestern University ( Chicago , IL ) . Revigo pathway analyses were generated by pasting ranked genes in the Gene Ontology enRIchment anaLysis and visuaLizAtion ( Gorilla ) tool [68 , 69] , and the output was plotted as GO terms to generate ranked cellular pathways [70] . Ingenuity Pathway Analyses ( IPA ) were performed using the Molecular Activity Predictor software ( Qiagen ) . Gene Set Enrichment Analysis ( GSEA ) were performed with GSEA ( Broad Institute ) using C7 databases . Data included 3 control mice , 3 mice treated with PD-L1 blockade , and 4 mice treated with LPS and PD-L1 blockade . RNA-Seq data were deposited in the GEO database ( GSE123153 ) , titled “Gene expression comparison of exhausted CD8 T cells after PD-L1 blockade alone or PD-L1 blockade combined with LPS , ” at https://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE123153 . Statistical analyses were performed using the test indicated in each figure legend . Data were analyzed using Prism software ( Graphpad ) . Statistical significance was established at p≤0 . 05 . Mouse studies were reviewed and approved by the Institutional Animal Care and Use Committee ( IACUC ) at Northwestern University ( protocol number IS00003258 , IS00008785 , IS00003324 ) . All mouse experiments were performed minimizing distress . Mouse experiments were conducted in accordance with recommendations listed in the Guide for the Care and Use of Laboratory Animals of the NIH . | Although PD-1 therapies have revolutionized cancer treatment , these therapies have not yet been licensed to treat chronic viral infections . This is because limited benefit is seen in pre-clinical models of chronic viral infection . Interestingly , recent reports in cancer models have suggested that certain microbes can affect the efficacy of PD-1 therapies , but the specific microbial products that modulate host responses to therapy remain unknown . We utilized a model of chronic viral infection to evaluate if bacterial lipopolysaccharide ( LPS ) , a major constituent of the microbiome , influences the efficacy of PD-1 therapy . Interestingly , we demonstrate that TLR4 triggering with low doses of LPS combined with PD-1 blockade induced a synergistic rescue of exhausted virus-specific CD8 T cell responses . Moreover , we demonstrate that adoptive transfer of LPS-activated DCs also results in similar improvement of PD-1 therapy without inducing overt immunopathology . Mechanistically , the synergy was DC-intrinsic , IFN-I independent , and B7/CD28 dependent . Taken together , our data may be important for understanding how components of the microbiome modulate the efficacy of PD-1 therapy , and may result in novel combined regimens for treating chronic viral infections . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immune",
"physiology",
"viral",
"transmission",
"and",
"infection",
"spleen",
"cancer",
"treatment",
"immunology",
"microbiology",
"cancer",
"immunotherapy",
"oncology",
"clinical",
"medicine... | 2019 | TLR4 signaling improves PD-1 blockade therapy during chronic viral infection |
Biological systems often detect species-specific signals in the environment . In humans , speech and language are species-specific signals of fundamental biological importance . To detect the linguistic signal , human brains must form hierarchical representations from a sequence of perceptual inputs distributed in time . What mechanism underlies this ability ? One hypothesis is that the brain repurposed an available neurobiological mechanism when hierarchical linguistic representation became an efficient solution to a computational problem posed to the organism . Under such an account , a single mechanism must have the capacity to perform multiple , functionally related computations , e . g . , detect the linguistic signal and perform other cognitive functions , while , ideally , oscillating like the human brain . We show that a computational model of analogy , built for an entirely different purpose—learning relational reasoning—processes sentences , represents their meaning , and , crucially , exhibits oscillatory activation patterns resembling cortical signals elicited by the same stimuli . Such redundancy in the cortical and machine signals is indicative of formal and mechanistic alignment between representational structure building and “cortical” oscillations . By inductive inference , this synergy suggests that the cortical signal reflects structure generation , just as the machine signal does . A single mechanism—using time to encode information across a layered network—generates the kind of ( de ) compositional representational hierarchy that is crucial for human language and offers a mechanistic linking hypothesis between linguistic representation and cortical computation .
Here we describe how representations that maintain relational structure can be composed and decomposed in a layered neural network . A simple computational mechanism , time-based binding , in which time is used to carry information about the relationships between representations in the input , is one way to achieve such an architecture . It is a truism and/or a biological principle of cortical organisation that "neurons that fire together , wire together" [28] . As such , some time-based binding systems use synchrony of firing to link representations together in the network for processing [e . g . , 29] . Conversely , neurons that do not fire in synchrony can stay independent , and the proximity in time between firings can be exploited to carry information . Discovery of Relations by Analogy ( DORA; a symbolic-connectionist model of relational reasoning; the full computational specifics of the model can be found in Doumas et . al [19] , operating procedure available in Appendix A ) exploits the synchrony principle in order to keep representations separable in the limit while binding them together for processing . This situation means that the system can be said to have variable-value independence when the representation of a given variable and its particular value at a moment in time are explicitly , independently represented [20 , 30] . A representation of a variable and its value must be able to function separately in any system performing a computation that requires relationality or ( de ) compositionality , such as during analogical reasoning , language processing , and likely many other higher-level cognition functions [31] . In a system with variable-value independence , statistics about the association between representations can still play an important role , but those statistics are not the sole basis of the representational architecture . In other words , variable-value independence allows the system to represent a variable , its value , and also to compute statistics about their association without changing the core representations . In brief , DORA's primary computational assumptions are ( 1 ) a neural network with layers of units , ( 2 ) lateral inhibition , ( 3 ) separate banks of units , ( 4 ) Hebbian learning , and ( 5 ) sensitivity to time . The layered structure of the network , combined with sensitivity to time as carrying information about the relations between the nodes in the layers of the network , is one solution to preserving the structure in ( 1 ) . After learning ( please see Appendix A in [19] and pp . 8–13 , 16–21 in the main text ) , the network represents words , phrases , and sentences across layers of the network , such that words and brackets in ( 1 ) correspond to nodes on different layers of the network . Nodes on higher layers code for the composition ( a phrase ) of two sublayer nodes ( words ) and fire when either of the subnodes below it fire in time . Asynchrony of unit firing ( in this case , of a node , but in theory , of a neural population or assembly , see [32] ) allows the network to bind representations together for processing on a higher layer while maintaining independent codes for the input representations on a lower layer of the network . The combination of layers and time-based binding via asynchrony is what allows the network to have ( de ) compositionality—representations of compositional product and the decomposed inputs on different layers of the network . To represent ( 1 ) , the network encodes the adjective phrase {funadj{gamesn}NP}ADJP over two layers of the network; on the lower layer , one node codes for the word {funadj} and another for the word {gamesn} ( see Fig 1 ) . On the next layer above that , the phrasal node will activate when the nodes {funadj} and {gamesn} fire . These word nodes fire staggered in time , or at an asynchrony , and still activate the node that codes for the phrase {funadj{gamesn}NP}ADJP . A similar configuration codes the phrasal binding between {wastev} and {timen} ( a node that codes the verb phrase {wastev{timen}NP}VP ) . The relationality between the phrases {funadj{gamesn}NP}ADJP and {wastev{timen}NP}VP is represented by encoding information about being an agent ( e . g . , "the waster" ) or a patient ( e . g . , "the wasted" ) in a two-argument predicate . One way to represent predicate argument relationships in a neural network is to code role information in a separate node from the particular argument that fills that role at a given processing moment . When the role slot in a predicate is represented separately from the given input , predicate–argument relationships can be generatively applied to any input that predicate is associated with in the dataset ( see Fig 1 ) . These argument role nodes , which code for the role-filler binding relation , can be learned in the same way the word nodes are ( see [19] for details ) . Lastly , a sentence node that represents the relation between {funadj{gamesn}NP}ADJP and {wastev{timen}NP}VP as a sentence ( a node that fires when the AdjP and VP units fire , and thus codes for the whole sentence {{funadj{gamesn}NP}ADJP{wastev{timen}NP}VP}IP ) is on the highest layer . The layered structure of the network is a core assumption and is necessary for time-based binding to function and for predicates to be represented in a connectionist network . We base the assumption that the system has layers on the broadest notion of cortical organisation; however , the representational codes for a given layer are learned from the input ( see [19] for detailed explanation of the learning process , which itself is not central to the results reported here , nor to the theoretical claim made here ) . The main advantage of time-based binding is that it avoids the superposition problem that a system that only uses synchronous firing alone would face . In sum , the essential difference between DORA and other connectionist models is that DORA binds representations using asynchrony of unit firing and thus can vary the level of representation at which the asynchrony is maintained . Slight asynchrony in unit firing leads to independent , discriminable sequences of representation across layers . Units on the next layer of the network then code for or fire when two or more subunits fire within a certain time of each other , which results in a hierarchical representation that can discriminate sequences with the same inputs and represents input values independently . In this way , DORA learns structured representations of relations from unstructured ( holistic flat feature vector ) inputs and is based on traditional connectionist computing principles ( i . e . , layers of interconnected nodes passing activation via weighted connections that are modified via Hebbian learning ) . However , in contrast with most connectionist networks , DORA effectively learns and implements hierarchical symbolic representations [20 , 33] . As summarised above , DORA does this by using time to encode relational information and uses comparisons of distribution of activity in time to subset units into function representations and to continue refining them throughout the learning process [19] . DORA accounts for numerous phenomena from relational learning , as well as its development ( e . g . , [19 , 34 , 35 , 36 , 37] ) . Although there are seldom clear physical boundaries in speech input that directly correspond to higher-level representations ( i . e . , phonemes , words , phrases , and sentences ) , we perceive and experience complex discrete representations in continuous input . A growing body of evidence suggests that this perception is based in the entrainment of cortical brain rhythms to speech [3 , 5 , 6 , 7] . While such neuroimaging evidence is suggestive and its implications tantalising , there is a mechanistic gap between such cortical signals and the representational output of the system ( as described by formal theories , or as might be formalised in any computational system ) . An adequate mechanistic linking hypothesis between linguistic and cortical computation would explain how the system goes from input of perceptual features to a hierarchy of structured representations and would describe how the computational mechanism gives rise to the observed cortical activation states . Furthermore , such a hypothesis would shed light on whether the observed cortical signal reflected “mere” tracking of hierarchical linguistic representations or whether it , in fact , reflects the online generation of hierarchical linguistic structure . To carve the computational problem at its joints , we turn to Ding et al . [6] , who report cortical tracking of hierarchical linguistic structures ( i . e . , words , phrases , and sentences ) in oscillatory data in both electrocorticography and magnetoencephalography recordings . Ding et al . presented auditory strings of synthesised speech in Mandarin Chinese and American English . Their stimuli were isochronous but manipulated the structural relationship between the syllables such that , in one condition , there was no meaningful relationship between the string of syllables/words ( “walk egg nine house” ) , in a second condition , phrases were formed from adjacent syllables ( “flat table angry birds” ) , and in a third condition , sentences emerged ( “new plans gave hope” ) . An increase in power at frequencies in the oscillatory response on the timescale of syllabic/lexical rate ( 4 Hz ) , phrasal rate ( 2 Hz ) , and sentential rate ( 1 Hz ) tracked the hierarchy of linguistic representation . Importantly , Ding et al . showed that the signal could not be attributed to entrainment to acoustic information , transitional probability , or word predictability [6] . Finding cortical entrainment at these slower oscillations ( 1 Hz , 2 Hz ) suggests that cortical populations are entraining to abstract linguistic representations like phrases and sentences . This is remarkable because it is unclear what , if any physical instantiation of these higher-level stimuli are present in the speech signal—at least , there is no ( known ) set of acoustic cues in speech that reliably or diagnostically signal word , phrase , and sentence structures . That Ding et al . [6] found cortical entrainment to such higher-level linguistic structures suggests that the brain is , nonetheless , sensitive to the presence of these abstract linguistic representations . Strikingly , DORA predicts such a representational pattern in oscillatory unit firing . DORA is a model of how information is represented—consequently , the simulations we present here are in no way intended as a fully articulated model of parsing . Furthermore , we want to emphasise that DORA represents a form of role-filler binding predicate calculus , in which expressions can be , but are not always , nor even usually , formally equivalent to the natural language expression . Future work is needed to derive representations in DORA with one-to-one correspondence to natural language; however , for the sentence stimuli tested herein , the difference between natural language form of the expression and the DORAese predicate logic does not bear on the theoretical conclusion that we draw from the results of our simulations . However , if the structure of information in DORA turned out to resemble how language appears to be represented in the human brain , that finding would indicate mechanistic synergy between the two computational systems . We present simulations of sentence processing within a model of relational concept learning that provides a mechanistic explanation for how representational structure emerges from unstructured input . We show that the model not only represents sentences and exhibits oscillatory activation patterns that are strikingly similar to human cortical oscillatory brain activity during exposure to the same stimuli [6] . We further demonstrate that the model , as a result of processing the sentence stimuli , forms representations of sentence meaning that reliably discriminate between semantically composable grammatical sentences and syntactically intact but semantically non-composable sentences . Our results , though not a model of parsing nor of a cortical microcircuit , are an existence proof that using time to carry information ( “time-based binding , ” or generating explicit hierarchical representations by encoding structural relations in time ) addresses two hard problems for cognition that have so far been investigated entirely separately: sentence processing and analogy ( see [19] for human-level simulations of relational and analogical reasoning ) . In order to determine if the oscillatory pattern from Ding et al . [6] can arise from the serial presentation of a stimulus at 4 Hz alone , we ran the same simulations in a RNN . The RNN simulations test whether signals like the cortical oscillations observed by Ding et al . [6] can arise from a system without time-based binding or representational hierarchy , in which they might arise from seriality alone . DORA processed the English sentence stimuli from Ding et al . [6] , as well as three control conditions wherein either only syntactic relationships , but no compositional meaning , were present ( Jabberwocky condition ) , no syntactic relationships existed between the words in the input ( Word List condition ) , or only phrases existed ( a version of the phrase-only condition from Ding et al . , Phrases condition; please see [6] for the list of Grammatical Stimuli and S1 Text for a list of our additional stimuli and Fig 2 for a schematic of Grammatical sentence representations in DORA's predicate calculus ) . We observed whether DORA represented the phrases or sentences correctly , and recorded the oscillatory pattern of unit firing in layers of DORA’s network during processing . Finally , in contrast to available brain data , we assessed the content of the representations DORA generated during processing of Grammatical and Jabberwocky sentences . We plotted the activation of existing predicates in memory in response to the hierarchical representations that parsing generated in those two conditions . To discount the hypothesis that the oscillatory pattern stems only from serialised processing , we repeated the four simulations in a RNN for comparison . Please see the Methods section for a description of the RNN that we trained . In brief , this was a standard RNN with one hidden layer that was trained on the same stimuli as in the DORA simulations .
The results of the simulations are presented in Figs 3 , 4 and 5 . Interestingly , the oscillatory pattern of firing in the various layers of DORA during processing of sentences , closely mirrored the patterns observed by Ding et al . [6] . Specifically , just like the cortical signals , for Grammatical sentences , DORA showed an activation burst that lasted throughout the processing of the sentence ( i . e . , firing in the 1 Hz range ) , activation bursts at twice the rate of the whole sentence burst ( i . e . , firing in the 2 Hz range ) , aligned with phrase-level processing , and activation bursts at four times the rate of the whole sentence burst ( i . e . , firing in the 4 Hz range ) , corresponding to word-level processing . However , the activity in the Word List condition also resembled the human data , in which in both cases , there was only spiking at 4 Hz but not at slower frequencies , which indicates that larger constituents were neither tracked nor formed in this condition . The Jabberwocky condition , which has no analogue in the available human data , resulted in oscillations that resembled the Grammatical condition , suggesting that the model is sensitive to something akin to syntactic structure or category during processing of Jabberwocky . Finally , we had the model process the Phrases condition , which contained strings of phrases , but not sentences , and it showed a power increase at 2 Hz and 4 Hz , resembling the human data to a phrase-only condition in Ding et al . [6] . To assess the nature of the representations DORA generated during parsing , we plotted activation maps of the contents of DORA’s memory over 100 trials ( instances of processing a sentence in the simulation ) in the Grammatical and Jabberwocky conditions . Note that the oscillatory firing pattern from the Word List condition indicated that DORA did not compose role-filler bindings to compare with memory in that condition . The difference in activation of existing representations in DORA’s memory between the Grammatical condition ( Fig 3 ) and the Jabberwocky condition ( Fig 4 ) are illustrated below ( Fig 5 ) . We plot only the activation of propositions that contain a word that was present in the stimuli in these particular trials ( 0–100 ) . The darker bars indicate that the hierarchical representations formed during processing more strongly activated representations that were already in the semantic memory of the model . Jabberwocky sentences activated fewer existing token units and activated these units to a lesser extent than the Grammatical sentences did , suggesting the model is generating novel , syntactically licensed representations that are "representationally unusual , " comparable to the human experience of reading syntactically intact but semantically anomalous Jabberwocky sentences . We performed a t-test comparing the number of units above threshold across the 100 runs in the Grammatical and Jabberwocky conditions and found that more units were active while processing a Grammatical sentence ( mean units active = 70 . 78 ) than a Jabberwocky sentence ( mean units active = 31 . 95 ) ; t = 58 . 312 , p < 2 . 2e−16 . We then repeated the four simulations in an RNN . In order to determine the activation level of the RRNs at each frequency , we attempted to identify units in the trained RRNs output layer that were active above a threshold of 0 . 7 consistently at 1 Hz , 2 Hz , 3 Hz , and 4 Hz across all sentences in each condition ( Fig 6 ) . Activation in the hidden layer corresponds to representations of the statistical patterns that the network learns , so finding patterns here would indicate that the RNN learned hierarchical linguistic structures from the input . For the Grammatical condition , zero units in the hidden layer were active at the 1 Hz , 2 Hz , and 3 Hz rates; five units ( out of 50 hidden units in the network ) were active above the threshold at the 4 Hz rate . For the Jabberwocky condition , zero units were active at the 1 Hz , 2 Hz , and 3 Hz rates; nine units were active above the threshold at the 4 Hz rate . For the Word List condition , zero units were active at the 1 Hz , 2 Hz , and 3 Hz rates; six units were active above the threshold at the 4 Hz rate . For the Phrase condition , one unit was active at the 2 Hz rate; five hidden units out of 30 nodes were active above the threshold at the 4 Hz rate . Fig 6 shows the proportion of units ( ranging from 0–0 . 2 proportion of units ) active in the recurrent layer at each rate in each condition . We not that the RNN achieved perfect performance , that is , it learned the sentences such that for any input word , the RNN could tell you the next n words with 100% accuracy . The difference between the RNN and DORA is that , in doing so , the RNN did not do anything in a way that resembles what humans do , that is , it neither formed ( symbolic ) hierarchical representations , nor oscillated , nor oscillated in a way that resembles the cortical signals to the same stimuli .
We have presented simulations from a computational model that learns and generates hierarchical structure from an unstructured string of lexical input . The model , DORA [19] , was built for a completely different purpose ( learning relational concepts to perform analogical reasoning ) but achieved the current outcome for sentence processing without any formal or structural changes from its original state . DORA learns and generates structured , symbolic representations using time-based binding in a layered neural network [19]; it is this computational architecture that enables the model to process sentences , to form hierarchical representations in general , and is what gives rise to the oscillatory pattern that resembled cortical oscillations . DORA is a model of how information is represented in the human mind , and perhaps more speculatively , how information might be macroscopically represented in cortical networks . It is not a model of parsing or cortical microcircuits . Nonetheless , when processing sentences like fun games waste time , the English sentence stimuli from Ding et al . [6] , the oscillatory firing pattern of units in various layers of DORA closely resembled the cortical oscillations observed by Ding et al . [6]—an activation burst lasting throughout the processing of the sentence ( 1 Hz range ) , activation bursts at phrase-level processing , or twice the rate of the whole sentence burst ( 2 Hz range ) , and activation bursts at word-level processing , or around four times the rate of the whole sentence burst ( 4 Hz range ) . Furthermore , the representations that the model generated during parsing reliably reflected whether the sentence was grammatical , jabberwocky , a list of phrases , or a list of words . This result indicates that DORA's architecture generates representations that are both semantically rich and structurally sensitive , two properties that are essential to human language . In contrast , a traditional connectionist network ( an RNN ) failed to show an oscillatory pattern that resembled the cortical signal , but a few nodes coded words at 4 Hz . Thus , there was no evidence of the RNN representing hierarchical linguistic structures . This contrast showed that seriality of processing alone is insufficient to produce the oscillatory pattern observed in humans and in DORA . Our results naturally beg the converse question , as to whether any system with representational hierarchy could produce the oscillatory pattern of activation that [6] and DORA show . For example , could natural language processing ( NLP ) parsers , which feature representational hierarchy and were developed to specifically parse natural language in a machine , produce oscillations and the pattern seen in [6] and in our simulations ? In principle , any system of hierarchy that is ( de ) compositional has the representational ingredients to encode units that could be fired in a sequence . However , we would argue that any given representational hierarchy could only produce oscillations if it were combined with time-based binding , which , as far as we know , no NLP system features . In DORA , time-based binding is the oscillation of activity throughout the network , which is part of the reason why the RNN did not show oscillatory activity nor the specific pattern from [6] . The representational structure of DORA ( a ( de ) compositional role-filler binding predicate logic ) is what makes the oscillations take the form that the data from [6] have . In terms of the specifics of the observed 1-2-4 Hz pattern , both [6] and our simulations are highly shaped by the word presentation rate of 250 ms/4 Hz . But without time-based binding , there is no mechanism to produce oscillations in a network , even in a NLP parser or other system with representational hierarchy . In sum , we remain relatively agnostic about the specific details of the required representational hierarchy because we do not yet know how to link the predicate calculus representations we use to natural language mental and cortical representations . What we are not agnostic about is the need for asynchronous time-based binding in order to produce oscillations in a neural network , as well as the need for representational hierarchy to produce the particular pattern of oscillations observed here and in [6] . Furthermore , another difference between DORA , RNNs , and NLP parsers is that DORA is a model of how information is represented and computed in the human mind . In other words , DORA makes a specific theoretical claim about how the human mind represents ( some kinds of ) information—that is , it does so in a relational , structured , hierarchical manner , which is realised via time-based binding . RNNs and NLP parsers do not make such theoretical claims about the architectures and mechanisms of mental and cortical computation . In DORA , structured representations form during learning and become activated during later processing . Our simulations and theoretical claims concern the way in which information is activated during processing . However , we note that it is the same computational mechanism that can generate representations during learning , which also performs processing: the use of time to carry information about the relations between inputs , implemented as systematic temporal asynchrony of unit firing across layers of the network , which we have called "time-based binding . " Through temporal asynchrony of firing , the model can use separable populations of units to maintain activation of different levels of representation as they occur in time . This ability to maintain hierarchy is the computational feature that turns out to be crucial for representing and processing the kinds of relations that are necessary to represent human language and is likely to be necessary for generating hierarchical levels of representation in any cortical network . We note here that our argument that temporal asynchrony is the mechanism that gives rise to hierarchical representation finds traction also in the conceptual terminology of neural oscillations , namely that temporal asynchrony corresponds to neural desynchronisation . Importantly , synchrony and asynchrony of unit firing in time are not orthogonal mechanisms; in fact , they are the same function or variable with different input values ( e . g . , sin ( x ) and sin ( 2x ) ) that can carry different information ( see [38] ) . Our mechanistic claim is that it is asynchrony that allows the system to bind information for processing while maintaining ( de ) compositionality and generating hierarchical representations . Binding or forming representations through synchrony alone would fail at the superposition problem , effectively superimposing a variable and its value onto a single , undecomposable representation . The generation of structured representations in DORA is a form of predication [19] . DORA’s representations are symbolic predicates that code for the invariant relations between features and objects ( e . g . , between the features “fun” and the object “games , ” the feature “adjective” and the object “fun , ” or the feature “noun” and the object “games” ) , or between objects and each other ( e . g . , between the object “fun games” and the object “waste” [19 , 39] ) . The resulting computational architecture thus explicitly represents invariant relations between input representations ( or variables ) and output representations ( or values ) , as a function . This architecture is in contrast with RNNs and current deep-learning algorithms , which associate input and output via statistical association and , therefore , do not ( currently ) preserve compositionality or represent relational structures [20 , 25 , 26 , 30 , 40 , 41] . Because DORA explicitly codes invariance as a function , it can combine novel inputs with existing predicate structures , leading to productive , combinatorial generativity of representation that is neither maligned by violations of statistical regularity , and not reliant , in principle , although not demonstrated here , on hard coding of representations ( cf . [42] ) . Though not directly relevant to the claims we make here , DORA is able to learn hierarchical structure from unstructured input—a feature that is very important for any developmentally plausible model of cognition and to models that seek to explain cortical and biological system organisation and plasticity . The ability to learn structured representations contrasts with current Bayesian models , which assume structured representations a priori or fit them from a specified space of possible representations predefined by the modeller ( cf . [42] ) . Although Bayesian models have a powerful descriptive application , they do not currently offer a mechanistic explanation for how a biological system came to be the way it is . What are the computational origins of predication , and where does it fall in the taxonomy of cortical computations ? What might those origins tell us about why a model of analogy happens to be able to process sentences ? Perhaps communicating information across time and space or needing to code for a relationship between representations that goes beyond , or even violates , statistical regularity ( such as encountering novel objects and interacting with them , or interacting with old objects in new ways ) were challenges that were sufficient to recruit a latent computational mechanism from existing neurocomputational subroutines , in response to the common computational requirement that these problems entailed . One hypothesis is that the underlying computation behind predication is a domain-general abstraction mechanism , such that language processing and analogical reasoning are instances of processing that call upon the same abstraction subroutine , one that might also underlie other “binding”-like phenomena in cognition and perception [13 , 27] , in any neural system that requires relationality or representational hierarchy . An abstraction sub-routine might be a cognitive or computational mechanistic “kind” that is at work in much of human cognition [9 , 11 , 22 , 23 , 32 , 43] . Our results suggest a formal and mechanistic synergy between how representational structures are computed , and how energy is expended in both cortical and machine oscillators . The pattern of oscillatory activation observed in the layers of DORA arises directly from the online processing of hierarchical sentence representations—by inductive inference , the cortical signal reflects generation too , rather than “mere” tracking , of hierarchical linguistic representation . Time-based binding via asynchrony , the computational mechanism in our model , links the generation of hierarchical structure to the observable “read-out” in the machine and cortical signal , with broad implications as a computational first principle of cognition in the human brain . Minimally , it explains how a computation gives rise to hierarchical representation and why cortical signals stemming from said computation appear the way they do . As such , our results can make the broad prediction that there ought to be temporally dissociable populations oscillating asynchronously; that is , desynchronisation between neural assemblies should occur as a function of the level of linguistic analysis that is being represented in a cortical network at a given time step . In other words , DORA's representational architecture predicts desynchronisation between assemblies oscillating at different frequencies in order to represent , for example , the speech envelope , acoustic phonetic features , syllables/phonemes , morphemes , words , phrases , and sentences . Whether the cross-frequency desynchronisation signal is more strongly observable at stimulus onset and offset or whether representational ( de ) synchronisation signals should be thought of as the dynamics of phase-to-power cross-frequency coupling over time , as well as a myriad of other important complications , are difficult to predict concretely at this stage . But , it is likely that functional characterisation of the dynamic recruitment or entrainment of cell assemblies during information processing could reveal powerful mechanistic first principles of cortical representation and computation [32 , 38] . Our results support a view where the organisational principles of cognitive , cortical , and biological systems arise from the nature of the mechanisms that carry out the computations that the system must perform . Our results provide a mechanistic explanation for ( 1 ) how our brains form discrete hierarchical representations from holistic unstructured input , as in language comprehension and other domains of cognition , ( 2 ) how patterns in cortical oscillations relate to representational structure building , and ( 3 ) how the system exploits the fact that time carries information to achieve a representational system that is generative and ( de ) compositional . The mechanism that gives rise to this state of affairs is time-based binding—the asynchrony of unit firing across layers of the network that allows the model to represent information independently at multiple timescales . The class of possible processing mechanisms that accounts for the output of the cortical computation signal must correspond in some fundamental way to the mechanism through which DORA forms representations . This computational mechanism gives rise to a generative representational hierarchy , as observed in machine and cortical oscillations , serving as an “abstraction engine” for representation in the human brain . Our results suggest that the identification of biological mechanisms , circuits , and subroutines that can perform computations beyond the domain in which they arose , or from which they were derived—a form of computational "recycling"—offers an approach to understanding biological systems , that , through modelling , can derive mechanistic explanations for why systems are the way they are .
In DORA , before learning ( although again , not demonstrated here ) , objects are represented as flat feature vectors ( see Fig 7 ) . After learning , relational structures are represented by a hierarchy of distributed and localist codes ( see Figs 1 , 2 and 8 ) . At the bottom , “semantic” units represent the features of objects and roles in a distributed fashion . At the next level , these distributed representations are connected to localist units ( called POs in DORA ) representing individual predicates ( or roles ) and objects; in these simulations , these units represent words . Localist RBs link object and predicate units into role-filler binding pairs for processing; these units represent phrases in these simulations . At the top of the hierarchy , a localist P-unit that represents the sentence links RBs ( phrases ) into a whole relational proposition or sentence ( see Figs 1 and 2 ) . The token units at the various layers of DORA represent information at a progressively more conjunctive level . While independence of roles and filler is maintained in the semantic and PO units , RB units code conjunctively for specific phrases or role-filler bindings , and P-units code for conjunctions of role-filler sets into full relational propositions , or sentences in this case . Conjunctive binding is sufficient for long-term storage but violates role-filler independence and so fails fundamentally for any tasks that require independent representation of roles and fillers [20 , 44] , as processing hierarchical and symbolic structure requires that representational elements in a system can be composed into structures in a manner that does not violate the independence of those elements ( see [26 , 40] ) . Consequently , during structured processing , DORA must maintain binding information in those units ( semantics and POs ) that maintain role-filler independence . DORA is composed of a number of sets or banks of units ( see Fig 8 ) . The driver is the current focus of attention , or what DORA is thinking about , at any given moment . The recipient is a set of units representing propositions that are available for comparison with the propositions in the driver . Hummel and Holyoak [45] have described the recipient in terms of Cowan’s [46] active memory . Finally , LTM is the set of propositions that DORA has encountered in the past that are not currently active . All banks of units are connected via a common pool of semantic units . During processing , activation flows from units in the driver , to the semantic units , and then to units in the recipient and LTM . The flow of activation from driver to recipient is fundamental for a number of DORA’s operations , including analogical mapping , inference , and predicate learning and refinement . The flow of activation from driver to LTM is important for retrieval . None of the current simulations rely on operations involving flow of activation between driver and recipient , so we do not discuss the recipient set any further . The driver , as the focus of DORA’s attention , is the starting point for all of DORA’s processing . When DORA performs any structured processing , role-filler bindings must be maintained in the units that maintain role-filler independence ( see above ) . This dynamic binding ( bindings that do not violate role-filler independence and can be created and destroyed on the fly [20] ) information is carried in DORA using time . Specifically , DORA uses systematic asynchrony of firing to maintain role-filler bindings . In DORA , binding information can be carried either by synchrony ( as in the symbolic-connectionist model Learning and Inference with Schemas and Analogies [LISA] [21] ) or by systematic asynchrony of firing , with bound role-filler pairs firing in direct sequence . Asynchrony-based binding , or what we call "time-based binding , " allows roles and fillers to be coded by the same pool of semantic units , which allows DORA to learn representations of relations from representations of objects ( Doumas et al . [19] ) . During asynchronous binding , in which a proposition like waste ( games , time ) becomes active in the driver ( see Fig 8 ) , the units representing waster fire ( along with units conjunctively coding for waster+games and for the waste [games , time] proposition; see Fig 1A ) , followed directly by the units representing games ( along with units conjunctively coding for waster+games and for the waste [games , time] proposition; see Fig 1B ) , representing the binding of waster to games . Then , the units representing wasted fire ( along with units conjunctively coding for wasted+time and for the waste ( game , time ) proposition; see Fig 1C ) , followed directly by the units representing time ( along with units conjunctively coding for wasted+time and for the waste [time , games] proposition; see Fig 1D ) , representing the binding of wasted to time . In short , bound role-filler pairs fire in direct sequence and out of synchrony with other bound role-filler pairs . These patterns of sequential oscillation dynamically code role-filler bindings in DORA and underlie DORA’s capacity to use the representations that it learns to support relational reasoning ( e . g . , analogical mapping , schema induction , and relational induction; see [19] ) and to learn structured relational representations from unstructured object representations . While establishing time-sharing patterns of firing in a connectionist model might seem complicated , as demonstrated by Hummel and Holyoak [21 , 45] and Doumas et al . [19] , it is actually rather simple . In DORA , each token unit is actually a coupling of two units , an exciter and a yoked inhibitor . The exciter unit behaves like a conventional node in a neural network , taking and passing input to units at higher and lower levels and laterally inhibiting and being inhibited by units in the same layer . Each exciter is also yoked to an inhibitor unit that integrates input over time and , when a threshold is reached , forces the yoked exciter unit to inactivity , allowing other units to become active . Continuing the above example , when the phrase/RB unit coding for waster+games fires , the two word/PO units connected to that phrase/RB—waster and games—compete to become active . Due to noise in the system , one of these tokens will become slightly more active and inhibit the other to inactivity . For example , waster might become slightly more active and inhibit games to inactivity . After some time firing , waster’s inhibitor unit will fire , forcing it to inactivity and allowing the word/PO representing games to fire . Similarly , after some time firing , the inhibitor yoked to the phrase/RB unit coding waster+games will fire , forcing that unit to inactivity and allowing another phrase/RB ( e . g . , wasted+time ) to fire . Establishing the pattern of firing described above requires units in different layers firing at different timescales . This pattern can be achieved in any number of ways , the simplest being setting the threshold of the inhibitor units appropriately ( e . g . , word/PO inhibitors have half the firing threshold of phrase/RB inhibitors ) . In DORA , inhibitor units of words/POs integrate input both from their yoked exciter and from the exciters of all units in the above layers ( e . g . , phrases/RBs ) . Consequently , words/POs naturally oscillate at twice the frequency of phrases/RBs . Crucially , sequential firing of related constituent elements is a necessary property of binding via synchrony and systematic asynchrony . When DORA performs any structured processing , a pattern will invariably emerge wherein bound elements within a larger compositional proposition will fire in direct sequence and at a different timescale than units coding for conjunctions of independently bound elements and full propositional compounds . In the following section , we show that the pattern of activation produced by DORA as it processes compositional structures very closely matches the temporal pattern of spike activity observed in Ding et al . [6] when people process sentences . We simulated the Ding et al . [6] studies using the same English sentences used in their Experiments 5 and 6 ( with native English speakers ) . All of these sentences took the form modifier-noun-verb-noun , forming sentences like “new plans give hope , ” “fun games waste time , ” and “dry fur rubs skin . ” DORA can represent hierarchical propositions by representing propositional structures as arguments of other propositional structures . For example , to represent “dry fur rubs skin , ” the modified noun phrase “dry fur” can be represented explicitly by the propositional structure dry ( fur ) , which can then serve as the argument of the agent role of the rubs relation ( see Fig 2; details of higher-order structure representation in LISA , from which DORA is descended , and from DORA can be found in Hummel & Holyoak [21] and in Doumas et al . [19] respectively ) . To simulate Ding et al . ’s main experimental procedure , we allowed DORA to process Ding et al . ’s English sentences one at a time , using the representations of those sentences . Representations of the sentence structures ( e . g . , Figs 1 and 2 ) entered the driver ( i . e . , were attended to ) . We then activated these sentence structures one word at a time . That is , we activated the semantic units encoding the first word for 110 iterations , then the second word for 110 iterations , and so forth . As semantic units became active , they passed activation to token units in the driver . The units in the driver responded to the pattern of firing in the semantic units ( i . e . , the units fired to represent and encode binding information; see Fig 8 ) . To simulate Ding et al . ’s control condition we allowed DORA to process Ding et al . ’s random word sequences one at a time . In this Word List condition , there were no syntactic relationships between words . Representations of the sentence word sequence entered the driver ( i . e . , were attended to ) . DORA processed the word as it normally would ( i . e . , the units coding each term fired , but because the sequence of words included no propositional structure , only word/PO units fired during processing ) . We tracked firing rate of all the nodes in the driver as DORA processed the sentences . The results of the simulation and the comparison to the patterns observed by Ding et al . are presented in Fig 3 . Interestingly , the pattern of firing of the nodes in the various layers of DORA very closely mirror the patterns observed by Ding et al . Specifically , just like the human participants , DORA showed an activation burst only at the rate of word representation , or four times the rate of the whole sentence burst ( i . e . , the word/PO units firing in the 4 Hz range ) . DORA , like LISA , performs memory retrieval by firing propositions in the driver and allowing activation to flow to LTM via the shared semantic units . Units in LTM respond to the pattern of activation in the semantic units imposed by the units in the driver and are retrieved into active memory ( the recipient ) via a Luce [47] choice function ( see Doumas et al . [19] ) . For example , when DORA is "thinking about" how games waste time , and the proposition wastes ( games , time ) becomes active in the driver , activation will flow through the semantic units to units in LTM that share semantic overlap with the word/PO units becoming active in the driver ( i . e . , wasters , wasted-things , games , time , and things like them ) . We used this property of the model to further test whether the model is representing syntactic structure . We had DORA to process versions of Ding et al . ’s word sequences in the Word List condition ( please see S1 Text ) , and we created a Jabberwocky condition where there were only syntactic relationships between words but no typical compositional semantic relationships . Representations of the sentence word sequence entered the driver ( i . e . , were attended to ) . DORA processed the sentences one word at a time ( i . e . , the units fired to represent and encode binding information , as above ) . We tracked firing rate of all the nodes in the driver as DORA processed the sentences . The results of the simulation and the comparison to the patterns observed in the experimental conditions are in Figs 3 , 4 , 5 and 6 . Ding et al . also manipulated a form of constituency—the linguistic relationship between discrete units and larger units , in this case , between syllables and words—to determine if there was evidence for cortical tracking of these various units . They found that words with multiple syllables elicited oscillations that tracked with the duration of the phrase boundary , not just syllables , which were fixed at 250 ms or 4 Hz . Ding et al . found 2 Hz and 4 Hz activity for two disyllabic words that together formed a phrase ( a stimulus stream of ( xx ) ( xx ) , but no 1 Hz activity . Since DORA does not have the perceptual apparatus to process auditory signals , we created a text analogue of the phrase condition ( Phrases condition ) . To test whether the oscillatory pattern observed by Ding et al . and in our DORA simulations can be observed in a system without time-based binding and without representational hierarchy , we repeated all four simulations in a RNN implemented in Theano [48] . The network had one hidden layer ( see Fig 9 ) . | Human language is a fundamental biological signal with computational properties that differ from other perception-action systems: hierarchical relationships between sounds , words , phrases , and sentences and the unbounded ability to combine smaller units into larger ones , resulting in a "discrete infinity" of expressions . These properties have long made language hard to account for from a biological systems perspective and within models of cognition . We argue that a single computational mechanism—using time to encode hierarchy—can satisfy the computational requirements of language , in addition to those of other cognitive functions . We show that a well-supported neural network model of analogy oscillates like the human brain while processing sentences . Despite being built for an entirely different purpose ( learning relational concepts ) , the model processes hierarchical representations of sentences and exhibits oscillatory patterns of activation that closely resemble the human cortical response to the same stimuli . From the model , we derive an explicit computational mechanism for how the human brain could convert perceptual features into hierarchical representations across multiple timescales , providing a linking hypothesis between linguistic and cortical computation . Our results suggest a formal and mechanistic alignment between representational structure building and cortical oscillations that has broad implications for discovering the computational first principles of cognition in the human brain . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"learning",
"neurolinguistics",
"linguistics",
"neural",
"networks",
"engineering",
"and",
"technology",
"signal",
"processing",
"social",
"sciences",
"neuroscience",
"learning",
"and",
"memory",
"sentence",
"processing",
"cognitive",
"psychology",
"cognition",
"memory",
... | 2017 | A mechanism for the cortical computation of hierarchical linguistic structure |
Genome stability is jeopardized by imbalances of the dNTP pool; such imbalances affect the rate of fork progression . For example , cytidine deaminase ( CDA ) deficiency leads to an excess of dCTP , slowing the replication fork . We describe here a novel mechanism by which pyrimidine pool disequilibrium compromises the completion of replication and chromosome segregation: the intracellular accumulation of dCTP inhibits PARP-1 activity . CDA deficiency results in incomplete DNA replication when cells enter mitosis , leading to the formation of ultrafine anaphase bridges between sister-chromatids at “difficult-to-replicate” sites such as centromeres and fragile sites . Using molecular combing , electron microscopy and a sensitive assay involving cell imaging to quantify steady-state PAR levels , we found that DNA replication was unsuccessful due to the partial inhibition of basal PARP-1 activity , rather than slower fork speed . The stimulation of PARP-1 activity in CDA-deficient cells restores replication and , thus , chromosome segregation . Moreover , increasing intracellular dCTP levels generates under-replication-induced sister-chromatid bridges as efficiently as PARP-1 knockdown . These results have direct implications for Bloom syndrome ( BS ) , a rare genetic disease combining susceptibility to cancer and genomic instability . BS results from mutation of the BLM gene , encoding BLM , a RecQ 3’-5’ DNA helicase , a deficiency of which leads to CDA downregulation . BS cells thus have a CDA defect , resulting in a high frequency of ultrafine anaphase bridges due entirely to dCTP-dependent PARP-1 inhibition and independent of BLM status . Our study describes previously unknown pathological consequences of the distortion of dNTP pools and reveals an unexpected role for PARP-1 in preventing DNA under-replication and chromosome segregation defects .
DNA replication is a fundamental cellular process that ensures duplication of the genetic information and subsequent transfer to daughter cells . The accuracy of DNA replication can be hampered by various exogenous and endogenous stresses , threatening genome integrity . It has become clear that replication stress , due to disturbance of the DNA replication program , is a major source of genome instability early during cancer development . Replication stress is defined as any phenomenon that alters the fulfillment of the DNA replication program . These phenomena include alteration of the initiation and elongation steps of DNA replication , conflicts between DNA replication and metabolic pathways such as transcription and mRNA processing , nucleotide pool disequilibrium , and overexpression or activation of oncogenes [1–4] . Some loci in the human genome are particularly difficult to replicate . They include common fragile sites ( CFSs ) that have a high A-T content and origin-poor regions , and such loci are prone to the formation of secondary structures and late replication [5 , 4] . Such "difficult to replicate" regions are very sensitive to replication stress . Such stress can jeopardize the completion of their replication , with the possibility of the formation of intertwined sister chromatid bridges during mitosis [6] . There are two types of sister-chromatid anaphase bridges: chromatin bridges , consequences of defective sister chromatid segregation [7] , can be stained with conventional dyes , such as DAPI; and ultrafine anaphase bridges ( UFBs ) , which cannot be stained with conventional DNA dyes or antibodies against histones [8 , 9] . UFBs were discovered only recently , and are generally identified by the detection of the helicase-like protein , PICH ( Plk1-interaction checkpoint “helicase” ) . They have been found in all cultured normal cells tested , and are therefore probably physiological structures [10] . However , their prevalence often increases in constitutive or induced replication stress conditions , such as Bloom syndrome and the inhibition of replication progression , respectively . [9 , 11 , 12] As cells progress through anaphase , UFBs are progressively stretched and decrease in number , as they are resolved . Most UFBs detected in normal cells are of centromeric origin: they are thought to contain unresolved DNA catenations between the centromeres separating during anaphase [9] . However , some UFBs , of common fragile site origin [11 , 13] ( CFS-associated UFBs ) , are induced by treatment with the replication inhibitor aphidicolin and are detected through the binding of FANCD2/FANCI protein complexes to bridge ends . Defects in either FANCD2 or its partner FANCI are involved in Fanconi anemia syndrome , and these proteins have been reported to co-localize with fragile sites [14] ) . UFBs , the origin and function of which remain unclear , do not result from recombination intermediates [15 , 16] . They may correspond to replication intermediates persisting on entry into mitosis , reflecting a failure to complete DNA replication or to resolve sister chromatid catenanes fully [6 , 17] . As CFS are replicated late and associated with loci poor in replication initiation events [18 , 19] , it has been suggested that CFS-associated UFBs originate from DNA incompletely replicated when the cell enters mitosis [11] . In cells from Bloom syndrome ( BS ) patients , chromatin bridges and PICH-positive UFBs are abnormally frequent [9] . The correlation between chromosomal instability and an increased risk of malignancy at an early age is stronger in BS than other pathological or physiological situations [20] . BS is a consequence of mutations in both copies of the BLM gene which encodes a 3’-5’ DNA helicase identified as a member of the RecQ family [21] . A characteristic of BLM-deficient cells is the frequency of sister chromatid exchange ( SCEs ) [22] . The BS cellular phenotype also includes chromosome breaks , slow replication fork speed , and high frequencies of both blocked replication forks and anaphase bridges [23 , 24] . These cellular features reflect the presence of endogenous DNA damage , replication stress and chromosome segregation defects and implicate BLM in favoring faithful duplication of the genome . However , we and others have reported that BLM deficiency leads to the downregulation of cytidine deaminase ( CDA ) [25 , 26] , an enzyme of the pyrimidine salvage pathway . CDA catalyzes the hydrolytic deamination of cytidine and deoxycytidine ( dC ) to uridine and deoxyuridine ( dU ) , respectively [27] . CDA deficiency causes excess dCTP leading to nucleotide pool disequilibrium . Some of the genetic instability associated with BLM deficiency , including the slow replication fork progression and about 30% of the increased sister chromatid exchange ( SCE ) frequency , results from a pyrimidine pool imbalance due to the CDA defect [25] . Whether other cellular defects associated with BS phenotype are due to CDA deficiency remains to be established . Here , we investigated the mechanism underlying the increase in UFB frequency in BS cells and its possible relationship to pyrimidine pool disequilibrium . We demonstrate that the formation of supernumerary UFBs , but not chromatin bridges , is fully due to defective CDA and not to the BLM defect per se . The increase in the dCTP pool resulting from CDA defect leads to a significant reduction of basal PARP-1 activity . PARP-1 is a multifunctional protein involved in diverse physiological processes and in the response to DNA damage [28] . Lower basal PARP-1 activity causes higher frequencies of unreplicated centromeres , foci of mitotic DNA synthesis and ( UFB ) -containing unreplicated DNA , without replication fork speed being affected . Our investigations reveal a novel mechanism of UFB formation and new pathological consequences of the distortion of dNTP metabolism fluxes . These findings imply that the effects of nucleotide pool disequilibrium may be more far-reaching than previously thought , and include jeopardizing genome stability not only through the regulation of fork progression , but also by reducing PARP-1 activity . We also provide evidence that PARP-1 has a previously unsuspected role in preventing replication stress and chromosome segregation defects .
We investigated whether the higher frequency of anaphase bridges in BLM- and CDA-deficient cells derived from BS patients ( BS cells , BLM-/CDA- ) was due to CDA deficiency itself . The characteristics of all the cell lines used in this study are presented in Table 1 . The expression of GFP-BLM in BS cells ( BS-BLM ) restores the expression of both BLM and CDA ( BLM+/CDA+ ) , as well as nucleotide pool equilibrium [25] . We analyzed the frequencies of chromatin bridges and UFBs in these cells by DAPI and PICH antibody staining , respectively ( Fig 1A ) . The numbers of both chromatin bridges and UFBs were lower in BS-BLM cells ( BLM+/CDA+ ) than in control cells ( BS-Ctrl ( BLM ) , BLM-/CDA- ) , as previously reported [9] . The stable overexpression of CDA in BS cells ( BS-CDA , BLM-/CDA+ ) restored the nucleotide pool equilibrium but did not restore BLM expression [25]; the frequency of UFBs in these cells was similar to that in BS-BLM cells ( BLM+/CDA+ ) , whereas the frequency of chromatin bridges remained similar to that observed in BS control cells ( BS-Ctrl ( CDA ) , BLM-/CDA- ) ( Fig 1A and 1B ) . UFBs originating from the centromere and CFS were similarly affected in CDA-deficient cells ( BS-Ctrl ( BLM ) and BS-Ctrl ( CDA ) ) ( Fig 1A and S1A Fig ) . CDA expression in BS cells ( BS-BLM and BS-CDA ) significantly increased the percentage of cells with no UFBs and substantially decreased the frequency of cells with more than three UFBs , indicating that the entire BS cell population is affected by CDA deficiency ( S1B and S1C Fig ) . These results suggest that supernumerary UFBs in BS cells result from the CDA defect and that the increase in chromatin bridge frequency may result directly from BLM deficiency . To verify these findings , we constructed cells stably expressing both BLM and CDA under the control of the CMV promoter ( BS-BLM-CDA ) ; this construct allowed the subsequent downregulation of BLM without downregulating CDA . In this model , CDA overexpression had no effect on the frequencies of chromatin bridges and UFBs ( S1D Fig ) ; however , siRNA-mediated CDA downregulation ( BS-BLM-CDA-siCDA , BLM+/CDA- ) increased UFB frequency to the levels observed in BS control cells ( BS-Ctrl ( BLM/CDA ) -siCtrl , BLM-/CDA- ) , without affecting chromatin bridge frequency ( Fig 1C ) . Also , siRNA-mediated BLM downregulation ( BS-BLM-CDA-siBLM , BLM-/CDA+ ) increased the chromatid bridge frequency to that in control BS cells ( BS-Ctrl ( CDA ) -siCtrl , BLM-/CDA- ) , without affecting UFB frequency ( Fig 1C ) . Thus , BLM prevents increases in chromatin bridge frequency , whereas CDA prevents increases in UFB frequency . We used tetrahydrouridine ( THU ) , a potent CDA inhibitor [29] that does not modify CDA levels , to confirm these findings ( S1E Fig ) . THU treatment induced a 33% increase in UFB frequency in CDA-expressing cells ( BS-BLM , BLM+/CDA+ ) , without affecting chromatin bridge frequency , but did not affect frequencies of UFBs and chromatin bridges in CDA-deficient cells ( BS-Ctrl ( BLM ) , BLM-/CDA- ) ( Fig 1D ) . Thus , the effect of THU , increasing UFB prevalence , was dependent on CDA , confirming that THU is a specific inhibitor of CDA , as previously reported [29] . BS cells ( BS-Ctrl ( BLM ) , BLM-/CDA- ) were cultured in the presence of dU ( dU is the final product of CDA , so its addition mimics CDA activity [25] without modifying CDA expression; S1E Fig ) : UFB frequency was 29% lower in BS cells cultured with than without dU , whereas chromatin bridge frequency did not differ ( Fig 1E ) . Supplementation with dU had no effect on UFB or chromatin bridge frequencies in CDA-expressing cells ( BS-BLM , BLM+/CDA+ ) ( Fig 1E ) . These findings demonstrate that the abnormally high frequency of chromatin bridges in cells from BS patients is caused by the BLM deficiency whereas that of UFBs is due to the CDA deficiency . We confirmed these findings in another cellular model based on an adenoviral short hairpin RNA ( shRNA ) specific for CDA to downregulate CDA stably in HeLa cells ( HeLa-shCDA ) . The CDA downregulation was highly efficient , with no detectable CDA protein and significantly less CDA mRNA than in controls ( S1F Fig ) . The CDA depletion in this model resulted in dCTP accumulation , replication fork slowing and an increase in SCE frequency , reproducing the part of the BS phenotype associated with CDA deficiency [25] ( S1G , S1H and S1I Fig ) . UFB frequency was 41% higher in HeLa-shCDA cells than in control cells , whereas chromatin bridge frequency was unaffected ( Fig 1F ) . These results were confirmed in four additional independent HeLa-shCDA clones , to check that they were not due to a clonal artifact . The UFB frequency in CDA-depleted HeLa cells was decreased by 25% by dU treatment , whereas that in control HeLa cells was increased by 34% by THU treatment ( S1J Fig ) . These experiments further confirm the role of CDA in preventing UFB formation in BLM-expressing cells . Thus , all supernumerary UFBs in BS cells result from the CDA deficiency , and CDA prevents supernumerary UFB whether or not BLM is expressed . Our data indicate that a CDA defect , and thus pyrimidine pool disequilibrium , results in an increase in the prevalence of centromere and CFS-associated UFBs . Arlt and Glover ( 2010 ) showed that very low doses of camptothecin ( CPT ) greatly decreased the frequency of CFS gaps and breaks upon replication stress , and the amount of single-stranded DNA at stalled replication forks . On the basis of these results , the authors suggested that polymerase-helicase uncoupling is a key initial event in CFS instability . They suggested that the slowing of the replication fork might cause such uncoupling , generating genetic instability [30] . We , therefore , first investigated whether very low doses of CPT could rescue UFB formation in the absence of CDA . We treated CDA-deficient cells and their corresponding CDA-expressing control cells with 2 pM CPT for 10 hours ( S and G2 phases of the cell cycle , Fig 2A ) or 3 hours ( G2 phase; Fig 2D ) [31]; we verified that this treatment did not affect CDA expression or the cell cycle ( S2A and S2B Fig ) . In CDA-deficient cells , but not CDA-expressing cells , UFB frequency was significantly decreased by CPT treatment but only if it was administered during both the S and G2 phases ( Fig 2B , 2C and 2E ) . Thus , as reported for CFS instability , treatment with 2 pM CPT during S-phase abolished the excess of UFBs in CDA-deficient cells . We then investigated whether replication uncoupling could be responsible for UFB formation . Replication uncoupling is thought to result in long stretches of single-stranded DNA ( ssDNA ) [30] potentially detectable by electron microscopy ( EM ) . We analyzed replication intermediates from CDA-depleted and control HeLa cells , treated or not treated with CPT , but observed no significant differences in the lengths of ssDNA at replication forks between the four conditions tested ( S2C–S2E Fig ) . Thus , neither CDA depletion nor 2 pM CPT treatment induced detectable changes in the length of ssDNA stretches at replication forks . However , CDA-deficient cells displayed a marked accumulation of ssDNA gaps , visible on both replicated and parental duplexes ( S2C and S2F Fig ) , consistent with impaired ssDNA gap repair in these cells . This is also consistent with our results showing a constitutive activation of both γH2AX and Chk2 in CDA-depleted HeLa cells ( S2G Fig ) , as previously reported in BS cells [24] . Moreover , as for other conditions affecting the integrity of the replication template [32] , the accumulation of ssDNA gaps in untreated CDA-depleted cells was associated with an increase in the frequency of reversed forks . As described for more acute CPT treatments [33 , 34] , 2 pM CPT treatment increased the frequency of both ssDNA gaps and reversed forks , regardless of CDA status ( S2F and S2H Fig ) . These findings suggest that the slowing of replication in CDA-depleted cells is not accompanied by generalized replication fork uncoupling , but rather with marked accumulation of ssDNA gaps and the associated increase in fork reversal . The accumulation of UFBs in CDA-deficient cells thus appears to be a consequence of other phenomena , which is suppressed by mild CPT treatment , through a molecular mechanism unrelated to fork uncoupling . We thus investigated the alternative molecular mechanism by which CDA deficiency contributes to supernumerary UFB formation . UFBs may be derived from unreplicated DNA or from double-stranded DNA catenanes [9] . UFBs cannot be stained by conventional DNA dyes , so we directly labeled DNA during the S and G2 phases of the cell cycle [31] . Cells were incubated for 10 hours with the thymidine analog 5-ethynyl-2’-deoxyuridine ( EdU ) and studied for bridges in anaphase ( Fig 3A ) . In both CDA-depleted HeLa cells and control cells , EdU incorporation was detected in all chromatin bridges , but not in UFBs ( Fig 3B and 3C ) . We cannot exclude the possibility that the thinness of UFB DNA structures limits EdU detection , but these data suggest that UFBs do not contain replicated DNA and might be derived from unreplicated DNA . We explored this hypothesis with other approaches . The replication of some chromosomal loci , such as CFS , may be completed during late G2 or early mitosis [35 , 36] . We thus investigated whether CDA deficiency could lead to detectable DNA synthesis during mitosis that would reflect under-replication of some chromosomal loci by the end of S phase . We therefore labeled CDA-depleted HeLa cells and control cells with EdU for 1 hour and analyzed the following mitosis ( Fig 3D ) . EdU foci in mitosis were significantly more numerous in CDA-depleted than control HeLa cells ( Fig 3E and 3F ) . About 55% of the EdU foci colocalized with CREST foci and about 24% of the EdU foci colocalized with FANCD2 foci , independently of the presence or absence of CDA ( S3 Fig ) . Using the Image J macro “confined displacement algorithm” [37] , we confirmed that the percentage of colocalization between EdU foci and CREST or FANCD2 foci was significantly different from that expected if the colocalization occurred by chance ( p<0 . 05 ) . As the excess of UFBs was abolished by CPT treatment in CDA-depleted cells , we analyzed the effect of CPT treatment on mitotic DNA synthesis . We found that the CPT treatment of CDA-depleted HeLa cells also prevented the accumulation of mitotic cells presenting EdU foci ( Fig 3F ) . Thus , both centromeres and CFSs are replicated particularly late in the absence of CDA , raising the possibility that some of these loci remain unreplicated when the cell enters mitosis . Importantly , CPT treatment abolished the excess of both mitotic DNA synthesis and UFBs in CDA-deficient cells , establishing a molecular link between the late replication of centromeres and CFSs , and UFB prevalence . We then hypothesized that not all centromeres and CFS are fully replicated in cells lacking CDA when entering mitosis . We used CREST staining and FISH ( fluorescence in situ hybridization ) -based assay with a centromeric probe specific for chromosome 8 ( Cen-8 ) to investigate whether centromere replication was impaired in CDA-deficient cells ( double-dotted and single-dotted CREST or Cen-8 foci , indicating fully replicated and unreplicated centromeres , respectively ) . The frequency of unreplicated centromeres was significantly higher in prometaphase/metaphase CDA-depleted HeLa cells than in the corresponding control cells , further evidence that UFBs arise from unreplicated DNA ( Fig 3G–3J ) . The CPT treatment of CDA-depleted HeLa cells also prevented the accumulation of unreplicated centromeres , indicating that defective centromere replication is likely a source of mitotic DNA synthesis and UFBs . Altogether , these data suggest that in CDA-deficient cells , some “difficult-to-replicate sites” , such as centromeres and CFS , are replicated particularly late , or even left unreplicated , leading to UFB formation . DNA synthesis events during mitosis may help rescue the duplication of some of these loci . However , many may not be processed by mitotic DNA synthesis , resulting in UFB formation . Importantly , CPT treatment rescued the defect in centromere replication , mitotic DNA synthesis and the excess of UFBs , providing a molecular link between compromised replication at “difficult to replicate sites” and UFB formation . We investigated how CPT treatment counteracts the effects of CDA deficiency . CPT and other genotoxic agents activate PARP-1 ( Poly ( ADP-ribose ) polymerase-1 ) , a multifunctional protein preventing genetic instability in response to DNA damage and replication stress [28 , 38] . PARP-1 catalyzes poly ( ADP-ribosyl ) ation ( PARylation ) by transferring ADP ribose units from nicotinamide ( NAD+ ) onto diverse acceptor proteins [39]: this post-translational modification is important in a wide array of physiological processes and in response to DNA damage [40–42] . We tested whether PARP-1 activity is compromised in CDA deficient cells . We assayed basal levels of PARylation by immunofluorescence microscopy and customized software for automatic counting of PAR foci . At least 500 cells per condition were counted for each experiment ( S4A Fig and materials and methods ) . Basal PARylation levels were significantly lower in CDA-deficient HeLa and BS cells than in control cells ( Fig 4A and S4C Fig ) . This was not due to a weaker PARP-1 expression in CDA-deficient cells ( Fig 4B ) . There were twice as many cells without PAR foci and fewer cells with three or more foci in the absence than presence of CDA ( S4B Fig ) . Therefore , basal PARP-1 activity is lower in the absence of CDA . CPT treatment abolished the mitotic defects observed in CDA-deficient cells ( Fig 2B and 2C ) , so we tested whether CPT treatment restored PARP-1 activity . CPT treatment activated PARP-1 in both CDA-deficient cells and control cells ( Fig 4C and S4C Fig ) . The frequency of PAR foci ( Fig 4C and 4D and S4C Fig ) and numbers of UFBs ( Fig 2B and Fig 2C ) in CPT-treated CDA-deficient cells were similar to those in untreated control cells . We treated cells with H2O2 , another PARP-1 activator [43] . The treatment of CDA-depleted cells with H2O2 increased PARylation ( S4D Fig ) , and normalized the UFB frequency ( Fig 4E ) . We assessed the dependence on PARP-1 activity of the suppressor effect of CPT , using the PARP-1 inhibitor olaparib [44] . Olaparib treatment ( S4E Fig ) reduced the frequency of PAR foci in both CDA-proficient and CDA-deficient cells , and abolished the effect of CPT on PARylation levels without changing CDA or PARP-1 levels ( S4F and S4G Fig ) . Olaparib treatment also increased UFB frequency in CDA-expressing cells , but not in CDA-deficient cells ( Fig 4F and S4H Fig ) . As expected , CPT treatment did not prevent supernumerary UFB formation in CDA-deficient cells treated with olaparib to inhibit PARP-1 ( Fig 4G and S4H Fig ) . These experiments show how CPT treatment alleviates the mitotic defects observed in the absence of CDA . There is a deficiency in the basal PARP-1 activity in CDA-deficient cells , and CPT treatment restores PARP-1 activity to levels similar to those in untreated control cells , rescuing the mitotic defects . To further confirm these data , we used HeLa cells stably depleted of PARP-1 with a specific shRNA ( HeLa-shPARP-1 ) [45 , 46] . We checked that BLM and CDA levels were unaffected in these cells ( Fig 4H ) . PARP-1-depleted cells displayed higher than control frequencies of unreplicated centromeres , metaphase cells presenting EdU foci , and UFBs , thereby mimicking CDA deficiency ( Fig 4I–4K ) . As expected , these increases were not abolished by CPT treatment ( S4I and S4J Fig ) . Thus , regardless of how PAR synthesis was suppressed , through the inhibition or depletion of PARP-1 , we observed an increase in the frequencies of unreplicated centromeres , EdU focus formation and UFBs . Moreover , these results also indicate that the effect of CPT in the prevention of excess UFB formation is strictly dependent on PARP-1 activity , with no additive effects of CDA deficiency and PARP-1 inhibition . CDA and PARP-1 must , therefore , act in the same pathway to prevent UFB formation . These results indicate that optimal PARP-1 activity is required for full centromere replication , which in turn prevents abnormal DNA synthesis and UFB formation during mitosis . CDA deficiency leads to a slowing of the replication fork [25] . Decreases in PARP-1 activity due to CDA deficiency compromise the completion of DNA replication and lead to UFB formation . We therefore investigated whether PARP-1 depletion affected replication fork speed: PARP-1 depletion accelerated fork progression ( S4K Fig ) , whereas CDA-deficiency slowed replication ( S1H Fig ) . Thus , the completion of DNA replication , which prevents the formation of UFB-containing unreplicated DNA , is not directly related to the overall rate of fork progression . Our results indicated that long-term dU supplementation rescued the excess of UFBs ( Fig 1E and S1J Fig ) , suggesting a possible influence of dUTP pool on PARP-1 activity . We therefore investigated whether CDA deficiency was associated with lower levels of cellular dUTP , as expected [27] . We detected no decrease in cellular dUTP levels in the absence of CDA ( S5A Fig ) , suggesting that dU supplementation did not act by restoring the dUTP pool . These results prompted us to check whether culturing CDA-deficient cells in the presence of dU decreased UFB frequency through PARP-1 activation , rather than by mimicking CDA activity , as previously proposed [25] . We found that long-term dU treatment significantly increased the PAR signal ( S5B Fig ) , indicating that the decrease in UFB frequency observed in dU-treated CDA-depleted cells was probably due to PARP-1 activation . These results suggest that the rescue by dU treatment involves a cellular metabolic process creating DNA lesions , similarly to CPT treatment . Indeed , we suggest that the addition of dU led to dUTP incorporation into DNA , activating the base excision repair pathway , and , thus , PARP-1 , to ensure that dUTP is removed from the DNA [47 , 48] . We conclude that the lower basal PARP-1 activity in CDA-deficient cells is probably not related to the dUTP pool , because it is unaffected in these cells . We then investigated whether excess dCTP impaired PARP-1 activity , because dCTP accumulation is an established consequence of CDA deficiency ( S1G Fig ) . The addition of dC to the culture medium decreased PARylation levels by 23% in CDA-expressing HeLa cells , and 18% in BS-BLM cells . Moreover , dC treatment during S phase ( S5C Fig ) increased UFB frequencies in these cells to levels similar to those in the corresponding CDA-deficient cells ( Fig 5A and 5B and S5D and S5E Fig ) . The addition of dC did not affect PARylation levels or UFB frequency in CDA-deficient cells ( Fig 5A and 5B and S5D and S5E Fig ) . The presence of dC also significantly increased the frequency of unreplicated centromeres in CDA-expressing HeLa cells ( Fig 5C and S5F Fig ) . These unexpected results led us to check whether the partial cellular inhibition of PARP-1 by dC treatment , which is known to induce dCTP accumulation [25] , reflected PARP-1 inhibition by dCTP . We used an in vitro colorimetric PARP assay kit ( Trevigen ) , with 3-aminobenzamide ( 3-AB ) as a control PARP-1 inhibitor . The inhibition of PARP-1 activity by dCTP was dose-dependent , and was 95% at 10 mM dCTP ( Fig 5D and 5E ) . Interestingly , 10 mM dATP , dGTP , dTTP or dUTP had no effect on PARP-1 activity ( Fig 5E ) , indicating that the only dNTP to inhibit PARP-1 in vitro is dCTP . This inhibition was fully reversed by the addition of 100 μM NAD+ ( NAD+ is the PARP-1 substrate used as a donor of ADP-ribose units for PARylation reactions [28] ) indicating that dCTP is a poor inhibitor of PARP-1 in vitro ( Fig 5F ) . Therefore , PARP-1 inhibition by dCTP in vivo is unlikely in the presence of the concentrations of dCTP and NAD+ found in cells ( see the discussion ) . Nevertheless , these analyses demonstrate that increasing the amount of intracellular dCTP by culturing cells in the presence of dC is sufficient to reduce PARP-1 activity .
We previously showed that CDA deficiency slows the progression of the replication fork and contributes to the high levels of SCEs in BS cells [25] . We report here that supernumerary UFBs in BS cells are entirely and solely due to CDA deficiency . In contrast , high chromatin bridge frequency results directly from BLM deficiency , and not from CDA deficiency . Our study demonstrates that these two types of sister-chromatid bridges have distinct genetic causes and presumably therefore different origins . This conclusion contrasts with a previous report suggesting that UFBs are derived from chromatin bridges through chromatin unraveling by PICH and BLM [49] . We also show that CDA deficiency makes a major contribution to several aspects of the cellular phenotype associated with Bloom syndrome . In view of this , all observations relating to or following BLM depletion should be interpreted with care: BLM depletion leads to CDA depletion [25 , 26] , and some of the cellular abnormalities observed may result from nucleotide pool disequilibrium associated with CDA deficiency , rather than BLM depletion per se . The precise nature and structure of UFBs remains to be established . Our data indicate that both centromeric and CFS-associated UFBs likely contain unreplicated DNA , suggesting that , like CFS-associated UFBs , centromeric UFBs originate from unresolved replication intermediates . The frequency of mitotic DNA synthesis foci that co-localize with centromeres and CFS was abnormally high in CDA-deficient cells . The nature of these DNA synthesis events is still unclear , but these observations are consistent with centromeres and CFS being late-replicated loci in the absence of CDA , the duplication of which may be completed into mitosis . The persistence of mitotic DNA synthesis may allow the completion of DNA replication/repair of sites that were not fully replicated by the end of S phase; however , it is possible that only few sites are able to benefit from this mitotic processing [36 , 12] , such that many remain unreplicated , leading to UFB formation ( Fig 5G ) . In support of this , the duplication of centromeres was impaired in the absence of CDA when cells enter mitosis . CDA deficiency caused a lowering of basal PARP-1 activity . The reactivation of PARP-1 activity by exposing cells to genotoxic agents that activate PARP-1 was sufficient to restore the completion of DNA replication and suppress the formation of supernumerary UFBs . In addition , both depletion and chemical inhibition of PARP-1 resulted in partially replicated centromeres and under-replication-induced UFBs . These findings reveal a previously unsuspected role for PARP-1 in contributing to the robustness of DNA replication and preventing abnormal segregation of unreplicated DNA . In contrast to CDA deficiency , stable down-regulation of PARP-1 accelerated replication fork progression . In previous studies , transient or chemical inhibition of PARP-1 did not affect replication fork speed [33 , 50] . Similarly , while PARP-1 inhibition was reported to limit fork reversal upon genotoxic stress [33 , 34 , 51] , the partial reduction in PARP-1 activity reported here in CDA-defective cells was associated with an increase in the rate of fork reversal , probably due to the observed increased in the frequency of discontinuities on the replication template [32] . The reasons for these discrepancies are unclear but could be related to the way PARP-1 activity was inhibited . Nonetheless , the completion of DNA replication , which prevents the formation of UFB-containing unreplicated DNA , is not directly related to the rate of fork progression . Our analyses show that the deleterious effects of nucleotide pool disequilibrium on genome stability extend further than previously anticipated , and include inhibition of PARP-1 activity . PARP-1 defects may therefore be a source of endogenous replication stress . A decrease in basal PARP-1 activity of about 30% led to the formation of a number of supernumerary UFBs similar to that observed following total PARP-1 inhibition or depletion . This suggests that there is a critical basal threshold of PARP-1 activity for full replication of the genome before the cells enter mitosis . Although PARP-1 promoted the recovery of stalled replication forks through homologous recombination ( HR ) [52] inhibition of RAD51 , and thereby of HR does not affect UFB frequency [15 , 16] . Therefore , PARP-1 may also contribute to the completion of replication at AT-rich DNA sequences through a mechanism independent of HR . Thousands of proteins , involved in diverse biological functions , are substrates of PARP-1 [39] . Most PARP-1 targets have been identified on the basis of responses to genotoxic stresses and basal PARylation levels of specific targets are difficult to determine [39] . In the absence of CDA , exposure of cells to genotoxic stress restored the activation of PARP-1 masking abnormal PARylation of individual substrates . Thus , further work is required to identify PARP-1 substrates that prevent UFB formation . CDA deficiency leads to an excess dCTP . We report that increasing the intracellular dCTP pool in vivo by culturing cells in the presence of dC was sufficient to reduce PARP-1 activity , leading to the under-replication-induced formation of UFBs . We also showed that dCTP was the only dNTP that inhibited PARP-1 activity , although this inhibition was observed only at high dCTP concentrations ( between 6 . 5 and 10 mM ) and was fully reversed by 100 μM NAD+ . The physiological intracellular concentrations of dCTP and NAD+ have been estimated to be about 30 μM and 1 mM , respectively [53 , 54] , and CDA-deficient cells contain about twice as much dCTP as control cells [25] ( S1G Fig ) . As 100 μM NAD+ is sufficient to reverse the PARP-1 inhibition induced by 10 mM dCTP in vitro , it is unlikely that dCTP compete with NAD+ in vivo , to inhibit PARP-1 . However , we cannot formally exclude the possibility that a particular environment within cells influences the local NAD+/dCTP ratio during replication , making it possible for dCTP to inhibit PARP-1 directly . The analysis of local concentrations of nucleotides at sites of DNA replication or repair remains difficult and this issue has been little explored , but such analyses in the future would undoubtedly shed light on new mechanisms . In conclusion , our findings indicate that in CDA-deficient cells , the presence of excess intracellular dCTP caused a lowering of basal PARP-1 activity , impeding replication at some chromosomal loci , including centromeres and CFS , on entry into mitosis . Unreplicated DNA that is not processed by mitotic DNA synthesis results in the formation of supernumerary UFBs ( Fig 5G ) .
Cell lines were cultured in DMEM supplemented with 10% FCS . BS-Ctrl ( BLM ) , BS-BLM cells , BS-Ctrl ( CDA ) and BS-CDA were obtained and cultured as previously described [25] . The BS-BLM-CDA cell line was obtained by transfecting BS-BLM cells with a vector containing the full-length CDA cDNA ( NM001785 ) , with JetPEI reagent . After 48 h , selection was carried out with 0 . 2 μg . ml− 1 puromycin ( Invivogen ) and 500 μg . ml− 1 G418 ( Euromedex ) . Individual colonies were isolated and maintained in culture with 0 . 1 μg . ml− 1 puromycin and 500 μg . ml− 1 G418 . HeLa-Ctrl ( CDA ) and HeLa-shCDA cells were obtained by transfecting cells with an empty pGIPZ vector or with the same vector encoding a short hairpin RNA sequence directed against CDA ( Open Biosystems , clone V3LHS_369299 ) , respectively , with JetPEI reagent . After 48 h , transfectants were selected on 1–5 μg . ml− 1 puromycin ( Invivogen ) . Individual colonies were isolated and cultured in medium containing 1 μg . ml− 1 puromycin . HeLa-Ctrl ( PARP-1 ) and HeLa-shPARP-1 cells were cultured as previously described [46] . For siRNA transfection assays , 3×105 HeLa cells or 8×105 BS-Ctrl ( BLM ) , BS-BLM , BS-Ctrl ( CDA ) , BS-CDA , or BS-BLM-CDA cells were used to seed the wells of a six-well plate . Cells were transfected with an siRNA specific for BLM or CDA ( ON-TARGETplus SMARTpool , Dharmacon ) or negative control siRNAs ( ON-TARGETplus siCONTROL Non Targeting Pool , Dharmacon; 100 nM final concentration ) for 48 h for BLM , or twice successively , for a total of 120 h for CDA , in the presence of DharmaFECT 1 ( Dharmacon ) . Deoxyuridine ( dU ) , deoxycytidine ( dC ) , deoxyuridine triphosphate ( dUTP ) , deoxycytidine triphosphate ( dCTP ) , deoxyadenosine triphosphate ( dATP ) , deoxyguanosine triphosphate ( dGTP ) and thymidine triphosphate ( dTTP ) were provided by Sigma Aldrich ( D5412; D0779; D4001 , D4635 , D6500 , D4010 and T0251 respectively ) ; tetrahydrouridine ( THU ) was provided by Calbiochem ( 584222 ) ; camptothecin ( CPT ) was provided by Sigma Aldrich ( C9911 ) and olaparib was provided by SelleckChem ( S1060 ) . THU and dU were added to the cell culture medium at a final concentration of 100 μM , for 96 h ( 2x48 h ) . Other drugs were added to the cell culture medium at the following concentrations: dC , 1 mM; H2O2 , 30 μM; camptothecin , 2 pM; olaparib , 1 μM . All cells were routinely checked for mycoplasma infection . Cells were lysed in 8 M urea , 50 mM Tris HCl , pH 7 . 5 and 150 mM β-mercaptoethanol , sonicated and heated at 75°C for 10 minutes . Samples ( equivalent of 2 x 105 cells ) were subjected to electrophoresis in NuPAGE Novex 4–12% Bis-Tris pre-cast gels ( Life Technologies ) . The procedures used for gel electrophoresis and immunoblotting have been described elsewhere [16] . Primary and secondary antibodies were used at the following concentrations: rabbit anti-BLM antibody ( 1:5 , 000; ab2179 from Abcam ) ; rabbit anti-CDA antibody ( 1:500; ab56053 from Abcam ) ; rabbit anti-β-actin antibody ( 1:10 , 000; Sigma ) ; rabbit anti-PARP-1 antibody ( 1:4 , 000; ALX-210-302 from Enzo Life Sciences ) ; rabbit anti-Chk2 ( 1/500; 2662 from Cell Signaling; rabbit anti-Chk2 T68 ( 1/500; 2661 from Cell Signaling; rabbit anti-H2AX ( 1/500; 2595 from Cell Signaling ) ; rabbit anti-H2AX S139 ( 1/500; 2577 from Cell Signaling ) ; horseradish peroxidase-conjugated goat anti-rabbit IgG ( 1:5 , 000; Santa Cruz Biotechnology ) . Immunofluorescence staining and analysis were performed as previously described [17] . Details of the protocol are provided in SI Appendix , SI Materials and Methods . At least three independent experiments were carried out to generate each dataset and the statistical significance of differences was calculated with Student’s t-test , Kruskal-Wallis tests or Mann and Whitney tests , as indicated in the figure legends . We distinguished true colocalization from the random colocalization of EdU foci and CREST or FANCD2 foci , by analyzing the images with the Image J macro “confined displacement algorithm” [37] . | The maintenance of genome stability is essential for the accurate transmission of genetic information , to ensure the successful duplication of chromosomes and their even segregation during mitosis . Errors occurring during DNA replication may affect both the accuracy of chromosome duplication and the balance of chromosome segregation during mitosis . Accurate DNA replication is strongly dependent on deoxynucleotides ( dNTP ) concentrations . Distortions in dNTP pool affect the rate of replication fork progression and compromise genetic stability . In the work presented here , we identified a novel mechanism by which dNTP pool disequilibrium compromises the completion of DNA replication and thus chromosome segregation , independently of the rate of fork progression . This mechanism involves the intracellular accumulation of deoxycytidine due to cytidine deaminase ( CDA ) deficiency , inhibiting PARP-1 activity . These results have direct implications for Bloom syndrome ( BS ) , a rare genetic disease combining susceptibility to cancer and genomic instability . BS cells also have a CDA defect , resulting in a high frequency of ultrafine anaphase bridges due entirely to dCTP-dependent PARP-1 inhibition . These data highlight new pathological consequences of the distortion of dNTP pools and reveal an unexpected role for PARP-1 in preventing the accumulation of excessive amounts of unreplicated DNA and chromosome segregation defects . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Pyrimidine Pool Disequilibrium Induced by a Cytidine Deaminase Deficiency Inhibits PARP-1 Activity, Leading to the Under Replication of DNA |
Mutations in the glucosylceramidase beta ( GBA ) gene are strongly associated with neurodegenerative diseases marked by protein aggregation . GBA encodes the lysosomal enzyme glucocerebrosidase , which breaks down glucosylceramide . A common explanation for the link between GBA mutations and protein aggregation is that lysosomal accumulation of glucosylceramide causes impaired autophagy . We tested this hypothesis directly by measuring protein turnover and abundance in Drosophila mutants with deletions in the GBA ortholog Gba1b . Proteomic analyses revealed that known autophagy substrates , which had severely impaired turnover in autophagy-deficient Atg7 mutants , showed little to no overall slowing of turnover or increase in abundance in Gba1b mutants . Likewise , Gba1b mutants did not have the marked impairment of mitochondrial protein turnover seen in mitophagy-deficient parkin mutants . Proteasome activity , microautophagy , and endocytic degradation also appeared unaffected in Gba1b mutants . However , we found striking changes in the turnover and abundance of proteins associated with extracellular vesicles ( EVs ) , which have been proposed as vehicles for the spread of protein aggregates in neurodegenerative disease . These changes were specific to Gba1b mutants and did not represent an acceleration of normal aging . Western blotting of isolated EVs confirmed the increased abundance of EV proteins in Gba1b mutants , and nanoparticle tracking analysis revealed that Gba1b mutants had six times as many EVs as controls . Genetic perturbations of EV production in Gba1b mutants suppressed protein aggregation , demonstrating that the increase in EV abundance contributed to the accumulation of protein aggregates . Together , our findings indicate that glucocerebrosidase deficiency causes pathogenic changes in EV metabolism and may promote the spread of protein aggregates through extracellular vesicles .
Mutations in the gene encoding the lysosomal enzyme glucocerebrosidase , glucosylceramidase beta ( GBA ) , are associated with neurodegeneration and brain protein aggregation [1 , 2] . Homozygous mutations in GBA cause the lysosomal storage disorder Gaucher disease , which in some cases includes devastating neurological symptoms [3] , while heterozygous GBA mutations are the strongest risk factor for both Parkinson disease ( PD ) and the related disorder dementia with Lewy bodies [1 , 2 , 4] . Up to 10% of individuals with nonfamilial PD carry a GBA mutation [5] . In addition , PD patients with a GBA mutation have faster progression of both motor and cognitive symptoms [6] . To study the mechanisms underlying the association between GBA mutations and neurodegeneration , we created a Drosophila model of glucocerebrosidase ( GCase ) deficiency . Drosophila has two GBA homologs , designated Gba1a and Gba1b . The Gba1a gene is expressed exclusively in the midgut [7] , and deletion of this gene does not appear to confer deleterious phenotypes [8] . By contrast , the Gba1b gene is ubiquitously expressed [7] , and Gba1b deletion causes marked abnormalities . We previously reported that Gba1b null mutants exhibit phenotypes including shortened lifespan , locomotor and memory deficits , neurodegeneration , accumulation of the autophagy adaptor Ref ( 2 ) P ( p62/SQSTM1 ) , and accumulation of ubiquitinated protein aggregates [9] . Similar phenotypes were subsequently seen in an independently generated Gba1b null mutant [8] . The protein aggregation and elevated Ref ( 2 ) P levels in Gba1b mutants suggested that they had impaired autophagy , as did morphological changes in the autolysosomal system noted by Kinghorn et al . [8 , 9] . These findings are consistent with previous reports of autolysosomal impairment upon loss of GCase activity [1 , 10–15] . Based on such findings , we and others hypothesized that lysosomal accumulation of glucosylceramide , the normal substrate of GCase , leads to impairment of autophagy [12 , 16–18] . However , none of the work implicating autophagy in the pathogenic effects of GCase deficiency has yet established that GCase loss of function causes global impairment of autophagic degradation . To investigate the autophagy failure model of GBA pathogenesis , we used proteomics-based techniques to measure protein turnover and abundance in Gba1b mutants and controls , as well as in flies with mutations in key autophagy ( Atg7 ) or mitophagy ( parkin ) genes [19 , 20] . While Atg7 mutants showed marked and widespread slowing of autophagy substrate turnover , Gba1b mutants did not . The effects of Gba1b mutation on the turnover and abundance of autophagy substrates also failed to correlate with those of Atg7 or parkin mutations . Moreover , we detected no deficits in turnover mediated by the proteasome , microautophagy , or endocytosis . However , we found high incidences of faster turnover and increased abundance among proteins associated with extracellular vesicles ( EVs ) , which have been previously suggested as a mechanism for the spread of protein aggregates in neurodegenerative disease . Biochemical studies confirmed increased abundance of EV marker proteins in isolated EVs from Gba1b mutants , and nanoparticle tracking analysis showed that the mutants had markedly increased numbers of EVs . Genetic manipulations to reduce EV production decreased the accumulation of ubiquitinated protein aggregates and Ref ( 2 ) P in Gba1b mutants , supporting the model that excessive EV abundance promotes the accumulation of protein aggregates . Together , our findings suggest that the most important pathological consequence of Gba1b loss of function is not failure of autophagic protein degradation but excessive production of extracellular vesicles .
To test the hypothesis that GCase deficiency causes impaired autophagic turnover , we compared protein degradation rates in heads from Gba1b mutants and controls using stable isotope labeling . In brief , our method involves feeding flies a stable heavy isotope of leucine and then using mass spectrometry to monitor the rate at which unlabeled proteins are degraded and replaced with labeled proteins [21] . We measured the influence of Gba1b loss of function on all proteins with data that met quality standards in both Gba1b mutants and controls ( 1297 proteins for turnover analysis , 4221 for abundance; S1 Data ) . We analyzed turnover data with Topograph [22] , software specifically designed for measurement of protein turnover via stable isotope labeling . We also compared protein abundance in Gba1b and control flies using Skyline [23] and MSstats [24] . Fold change in turnover and fold change in abundance were then calculated for every protein . Fold change for a protein was calculated as the value in Gba1b mutants divided by the value in controls . We predicted that autophagy substrates would show slower turnover ( longer half-lives ) in Gba1b mutants , and that they might show increased abundance if synthesis did not decrease to match the slower degradation rate ( Fig 1A ) . We defined autophagy substrates as proteins from mitochondria , cytosolic ribosomes , endoplasmic reticulum ( ER ) , and peroxisomes , all previously identified as targets of autophagy [25–30] . We validated our prediction using turnover and abundance data from autophagy-deficient Atg7 mutants , which we characterized in previous work [21] ( S1 Data ) . We first plotted Atg7 fold change in turnover against fold change in abundance for autophagy substrates to observe the overall pattern of proteostasis changes ( Fig 1B ) . Atg7 mutants showed changes consistent with our prediction: the vast majority of autophagy substrates ( 72% ) had slower turnover ( fold change in half-life >1 ) and increased or unchanged abundance . We therefore used Atg7 mutant data as a reference for the effects of autophagy impairment . When we plotted fold change in turnover against fold change in abundance for Gba1b mutants , the pattern of changes was markedly different; only 15% of autophagy substrate proteins had slower turnover and increased or unchanged abundance ( Fig 1C ) . Proteostasis of autophagy substrates in Gba1b mutants thus did not overall resemble the pattern seen in Atg7 mutants . To compare in more detail the effects of Gba1b and Atg7 mutations on protein turnover and abundance , we performed several additional analyses , beginning by calculating Gba1b and Atg7 mean fold change in turnover ( half-life ) for the autophagy substrate proteins mentioned above . Each mutant was compared to its own control . Turnover of proteins from all three classes of autophagy substrates was significantly slowed in Atg7 mutants ( p < 0 . 001 by nested ANOVA ) , but in Gba1b mutants there was no overall change in the half-lives of ribosomal or ER/peroxisomal proteins ( Fig 1D; p = NS by nested ANOVA ) and only a very mild slowing of mean mitochondrial protein turnover ( mean fold change 1 . 15 ± 0 . 32; p = 0 . 02 by nested ANOVA; Fig 1D ) . To test further for evidence of impaired autophagy in Gba1b mutants , we compared the effects of Atg7 and Gba1b mutations on individual proteins . We began with turnover , plotting the fold change in half-life for Gba1b mutants ( Gba1b mutant half-life/Gba1b control half-life ) against the fold change for Atg7 mutants ( Atg7 mutant/Atg7 control ) . We compared Gba1b and Atg7 effects on individual proteins from each of the three autophagy substrate categories . There was no statistically significant relationship between the effects of Gba1b and those of autophagy ablation for mitochondrial , ribosomal , or ER/peroxisomal proteins ( Fig 2A–2C ) . We also tested for a relationship between Gba1b and Atg7 effects on protein abundance ( Fig 2D–2F ) and found no significant correlation for any of the three autophagy substrate groups . The effects of Gba1b loss of function on protein turnover and abundance thus do not resemble the effects of autophagy ablation , and we find no evidence that Gba1b mutation causes global impairment of autophagic protein degradation . One reported consequence of GBA loss of function is accumulation of dysfunctional mitochondria due to defective mitophagy [31 , 32]; the slight but statistically significant slowdown of mitochondrial protein turnover in Gba1b mutants therefore raised the possibility of a mild mitophagy deficit . We had previously found a mitochondrial protein turnover deficit in flies with mutations in the mitophagy factor parkin [21] , and we now compared the effects of Gba1b mutation on mitochondrial proteostasis with those of parkin . In parkin mutants , turnover was slowed for the vast majority of mitochondrial proteins ( Fig 3A ) . In Gba1b mutants , changes in mitochondrial protein turnover were both milder and less consistent ( Fig 3A , S1 Data ) . We considered the possibility that Gba1b mutants had a mitophagy defect that was obscured by compensatory upregulation of other mitochondrial protein turnover mechanisms , as we previously found in PINK1B9 mutants , which lack a mitophagy factor upstream of Parkin [21] . In PINK1B9 mutants , while the mean fold change in mitochondrial protein half-life was not significantly altered , the effects of PINK1B9 mutation on individual proteins correlated strongly with those of parkin mutation . We therefore tested whether the effect of Gba1b on mitochondrial proteostasis would also correlate with the effect of parkin . However , we detected no significant correlation between Gba1b and parkin effects on mitochondrial protein turnover ( Fig 3B ) or abundance ( Fig 3C ) . Our findings therefore do not support either globally impaired autophagy or selectively impaired mitophagy in Gba1b mutants . The lack of evidence for autophagy failure led us to consider alternative explanations for the accumulation of ubiquitin-positive protein aggregates in Gba1b mutants . We first considered the possibility that these aggregates could arise because of reduced proteasome function , which is known to lead to the formation of large ubiquitin-positive protein aggregates called aggresomes [33] . We tested proteasome activity using fluorescent substrates , and found that all three enzyme activities were normal ( Fig 4A ) . We also considered the possibility that delivery of substrates to the proteasome might be impaired [34] , and used our proteomic data to determine whether actual proteasome substrates were degraded normally in Gba1b mutants . We identified cytosolic proteasome substrates based on data from Wagner et al . [35] ( S2 Data ) and compared turnover and abundance changes in this group of proteins to the changes in all other cytosolic proteins . If proteasomal degradation were impaired , we would expect substrates of the process to have slowed turnover and possibly increased abundance , as we predicted for autophagy impairment . In fact , however , the percentage of proteins with slowed turnover in Gba1b mutants was significantly lower for proteasome substrates than for other cytosolic proteins , and proteasome substrates did not show a greater incidence of increased abundance in Gba1b mutants ( Fig 4B ) . Together , these results indicate that proteasome dysfunction does not underlie the accumulation of ubiquitin-positive aggregates in Gba1b mutants . We next examined whether the protein aggregation in Gba1b mutants could be the result of altered endosomal functioning . As Gba1b mutants have markedly increased levels of glucosylceramide ( S1 Fig , [8] ) and moderately decreased levels of ceramide ( S1 Fig ) , abnormal membrane composition could compromise functioning of the endosomal system [36 , 37] . We therefore tested for impairment of endosomal microautophagy , an Hsc70-4–dependent process that degrades cytosolic proteins with specific targeting sequences ( “KFERQ-like motifs” ) [38] . To test whether microautophagy is impaired in Gba1b mutants , we searched the Drosophila proteome for proteins with KFERQ-like motifs , and compared the effects of Gba1b on cytosolic proteins with and without such motifs . Compared to proteins without KFERQ-like motifs , proteins with one or more KFERQ-like motifs did not have an increased incidence of proteins with slower turnover or increased abundance ( Fig 4C , S2 Data ) . We also tested whether overexpression of Hsc70-4 , which has been shown to increase microautophagy in Drosophila [39] , would influence the accumulation of insoluble ubiquitinated protein . However , this manipulation had no effect on the abundance of ubiquitinated protein aggregates ( Fig 4D ) . We thus found no evidence that impaired endosomal microautophagy is responsible for the accumulation of ubiquitinated protein aggregates in Gba1b mutants . We then investigated whether Gba1b mutations impaired the functioning of another endosomal degradation pathway , endocytic turnover . Using FlyBase [40] and other annotation resources , we identified typical substrates of this pathway , primarily integral cell membrane proteins ( n = 90 in turnover data , 437 in abundance data; S2 Data ) . We also identified a separate group of “endosomal machinery” proteins , which reside in endosomes or are required for endocytosis ( n = 32 in turnover data , 102 in abundance data; S2 Data ) . We found no evidence that degradation of endocytic turnover substrates was compromised; compared to all other proteins , endocytic turnover substrates did not have a higher frequency of significantly slowed turnover or increased abundance ( Fig 5A ) . When we examined endosomal machinery , however , we found a higher prevalence of proteins with increased abundance ( p < 0 . 0001 vs . all other proteins by Fisher exact test; Fig 5B ) . Thus , Gba1b mutants had no evidence of compromised endocytic turnover , but proteostasis of the endosomal machinery was clearly altered . Many endosomal machinery proteins also play roles in the creation and release of extracellular vesicles ( EVs ) , a heterogeneous population of membrane-delimited structures originating from the multivesicular endosome and plasma membrane [41 , 42] . EVs transport varied cargoes of protein and nucleic acids from cell to cell and play roles in signaling , waste disposal , and intercellular resource transfer [43–45] . EVs have also been implicated in the spread of protein aggregates in neurodegenerative disease [41] . Given that Gba1b mutants have altered turnover and abundance of endosomal machinery proteins but not endocytic turnover substrates , we considered the alternative possibility that GCase deficiency influences EV biology . To explore this hypothesis , we first tested whether proteins known to be associated with EVs showed significant alterations in turnover or abundance in Gba1b mutants . We compiled a list of proteins detected in EVs from Drosophila cultured cells [46–49]; the resulting list contained 544 nonredundant proteins ( S3 Data ) , 329 of which were found in the Gba1b turnover data and 499 in the abundance data . Compared to all other proteins in the dataset , a smaller percentage of EV-associated proteins had slowed turnover , and a higher percentage had faster-than-normal turnover ( p < 0 . 0001 by Fisher exact test; Fig 5C ) . In addition , a greater proportion of EV proteins had increased abundance in Gba1b mutants ( p < 0 . 0001 by Fisher exact test; Fig 5C ) . To confirm that EV-associated proteins had faster turnover and increased abundance in Gba1b mutants , we repeated our analysis using an independent list of EV proteins . We obtained the ExoCarta [47] “top 100” list of proteins most frequently identified in mammalian EVs and identified their Drosophila orthologs using DIOPT v6 . 0 [50] ( n = 97; S3 Data ) . Once again , compared to the rest of the dataset , EV-associated proteins had higher frequencies of faster turnover and increased abundance in Gba1b mutants ( Fig 5D ) , suggesting that GCase deficiency may cause dysregulation of EV biology . To test whether faster turnover and increased abundance of EV-associated proteins are specifically associated with Gba1b loss of function , we investigated whether these proteins were also disproportionately affected by other conditions that alter protein turnover . We evaluated the pattern of changes , as we had done for autophagy substrates , by plotting fold change in turnover against fold change in abundance for all EV-associated proteins . In Gba1b mutants , 59% of the datapoints representing EV proteins appeared in the quadrant representing faster turnover and increased abundance ( Fig 6A ) ; in Atg7 mutants , only 3% of EV-associated proteins showed the same pattern ( Fig 6B ) . We also looked at the pattern of EV proteostasis in other mutants described in our previous work [21]: the mitophagy mutants parkin and PINK1 , and the oxidative stress mutant Sod2 . Because abundance data for these mutants lacked enough significant changes for analysis , we analyzed turnover only . None of these mutants showed faster turnover of EV proteins ( S2 Fig ) . We also investigated whether the EV proteostasis alterations in Gba1b mutants represented a distinctive pathological process or simply an acceleration of normal aging , given that ubiquitinated protein aggregates accumulate with age even in wild-type flies [51–53] . To do this , we measured protein turnover and abundance in old flies ( 55 to 60 days at the start of labeling ) and young flies ( 5 days ) . Old flies had dramatically slower turnover of most proteins ( mean fold change in half-life for all proteins 2 . 46 ± 4 . 31 ) and milder changes in protein abundance ( both increases and decreases; S4 Data ) . In old flies , only 4% of EV-associated proteins were represented by datapoints in the faster turnover/increased abundance quadrant ( Fig 6C ) , indicating that the altered EV proteostasis observed in Gba1b mutants does not represent an acceleration of normal aging . Together , our findings indicate that altered proteostasis of EV-associated proteins is a specific and novel feature of Gba1b mutants . As mentioned above , all of our proteomic analyses were performed using protein extracts from fly heads . To test whether the observed alterations in EV protein abundance were also evident in EVs themselves , we performed western blotting for known EV markers on EV fractions from hemolymph , the Drosophila equivalent of blood . To do this , we collected cell-free hemolymph extracts containing the full range of circulating EVs , which we designated total EVs ( tEVs ) . We also prepared extracts containing only EVs under 220 nm in size , which we designated small EVs ( sEVs ) . We then performed western blot analysis on tEVs or sEVs compared to whole-fly homogenate to measure the abundance of two EV marker proteins: Rab11 and an HA-tagged form of ALiX ( PDCD6IP ) [48 , 54] . We also used western blotting to verify EV isolation by the absence of microsomal markers Calnexin ( Cnx99A ) and Golgin ( Golgin84; Fig 7A and 7E ) according to International Society for Extracellular Vesicles standards [55] . Rab11 and ALiX-HA were significantly increased in abundance in Gba1b mutants vs . controls in both tEVs and sEVs , but not in whole-fly homogenate ( Fig 7A–7E ) . Although the Rab11 detected in sEVs was 3–5 kDa smaller than in the whole-fly homogenate , this finding is consistent with previous work demonstrating altered molecular weights for several proteins when detected in EVs [56] . A GFP-tagged form of Rab11 also showed increased abundance in sEVs from Gba1b mutants ( S3 Fig ) . The findings using tagged forms of EV proteins are particularly informative because these exogenous proteins were expressed at equivalent overall levels in controls and Gba1b mutants ( Fig 7G , S3 Fig ) . The increased abundance of these markers in EVs from Gba1b mutants indicates that either more of each marker protein is loaded into each EV , or that Gba1b mutants produce more EVs . One of the most striking abnormalities in Gba1b mutants is their accumulation of Ref ( 2 ) P [9] , the Drosophila p62 ortholog , which was markedly elevated by proteomic measurement ( S1 Data ) . This is especially noteworthy given that accumulation of Ref ( 2 ) P is usually interpreted as an indication of impaired autophagic flux [57–59] , and yet we find no evidence of impaired autophagic degradation in Gba1b mutants . Ref ( 2 ) P/p62 has multiple functions , however , and mammalian p62 has been detected in EVs [47 , 60] . We therefore performed western blotting for Ref ( 2 ) P on sEVs from Gba1b mutants and controls to test whether Ref ( 2 ) P accumulates in EVs . The sEVs contained very little monomeric Ref ( 2 ) P , but did reveal a marked increase in higher molecular weight Ref ( 2 ) P oligomers ( Fig 8A–8C ) , which were approximately three times as abundant in Gba1b mutants as in controls ( Fig 8C ) . We confirmed that these high molecular weight bands represented Ref ( 2 ) P by performing RNAi knockdown of Ref ( 2 ) P in Gba1b mutants ( S4 Fig ) . The increased Ref ( 2 ) P abundance in Gba1b mutant EVs suggests that changes in EVs may contribute to the markedly increased Ref ( 2 ) P seen in Gba1b mutant heads . As mentioned above , the increased abundance of multiple EV-associated proteins in Gba1b mutants suggests either that more of each protein is loaded into each EV , or that more EVs are produced . To distinguish these possibilities , we performed nanoparticle tracking analysis on EVs from the hemolymph of Gba1b mutants and controls . For these experiments , we chose to use a 0 . 65 μm rather than a 0 . 22 μm filter to retain EVs of as many sizes as possible while still ensuring removal of all cell debris . While the mean size of EVs was comparable in Gba1b mutants and controls ( Fig 9A ) , the concentration of EVs was approximately six times higher in the mutants ( Fig 9B ) . The mean concentrations were 4 . 55 x 1011 particles/mL ( ± 1 . 87 x 1011 ) for Gba1b mutants and 7 . 28 x 1010 particles/mL ( ±3 . 36 x 1010 ) for controls ( p = 0 . 013 by Student t test ) . Thus , the increased abundance of EV proteins in Gba1b mutants is best explained by the increased production of EVs . Together , our findings give clear evidence of altered EV biology in Gba1b mutants . As previously noted , EVs have been repeatedly described as possible vehicles for the spread of brain protein aggregation in neurodegenerative disease [41] . Our finding that Gba1b mutants had more EVs led us to hypothesize that increased EV release promotes protein aggregation by increasing cell-to-cell transmission of aggregation-prone proteins . As a first step toward testing this model , we determined whether the accumulation of protein aggregates in Gba1b mutants could be suppressed by knocking down components of the ESCRT ( endosomal sorting complexes required for transport ) pathway , which are required for production of many types of EVs [41 , 54] . Using a pan-neuronal driver , we expressed RNAi against proteins from three of the four ESCRT complexes: Mvb12 ( Multivesicular body subunit 12; ESCRT-I ) , lsn ( larsen/Vps22; ESCRT-II ) , and CHMP2B ( Charged multivesicular body protein 2b; ESCRT-III ) . We found that knockdown of each of the three ESCRT proteins significantly reduced accumulation of Ref ( 2 ) P in Gba1b mutants , and that knockdown of Mvb12 and lsn also reduced the accumulation of insoluble ubiquitinated protein ( Fig 10A–10F ) . These findings support the model that excessive production of EVs is responsible for the accumulation of protein aggregates caused by GCase deficiency .
Impairment of autolysosomal degradation is widely thought to explain the increased risk of neurodegeneration associated with mutations in GBA , which encodes the lysosomal enzyme glucocerebrosidase ( GCase ) [1 , 16] , and multiple studies have found hallmarks of impaired autophagy associated with GCase loss of function . These hallmarks have included accumulation of ubiquitinated protein aggregates , increased abundance of autophagic flux markers such as p62/SQSTM1 and LC3-II , impairment of autophagosome-lysosome fusion , and changes in the size and number of autophagosomes and lysosomes [12 , 13 , 61–66] . These indications that GCase deficiency leads to autophagy impairment have been found in diverse experimental systems , including multiple animal models , cultured cells , iPSC-derived human neuronal models , and postmortem patient samples [8 , 11–13 , 31 , 66–70] . Our own initial characterization of Drosophila Gba1b mutants , which revealed extensive ubiquitinated protein aggregates and markedly elevated levels of the p62 ortholog Ref ( 2 ) P , also appeared to support the model that GCase deficiency impairs autophagic degradation [9] . In our current work , however , proteomic measurement of protein turnover and abundance showed no evidence that degradation of autophagy substrates was globally impaired in Gba1b mutants . The mutants also showed no evidence of failure in other protein degradation pathways . Instead , we found faster turnover and increased abundance of proteins associated with extracellular vesicles ( EVs ) . Followup experiments on isolated EVs confirmed increased abundance of EV marker proteins and revealed a strikingly increased number of EVs . Furthermore , genetic manipulations that reduced EV formation suppressed both the increased protein aggregation and the increased Ref ( 2 ) P abundance observed in Gba1b mutants . Our findings suggest that dysregulation of extracellular vesicles , rather than failure of autophagic degradation , may be the primary mechanism by which GCase deficiency leads to protein aggregation and neurodegeneration . Although the many previous reports of autophagy impairment in GCase-deficient organisms appear incompatible with our current protein turnover findings , we do not believe that our findings contradict previous work . When we measure common markers of autolysosomal function such as Ref ( 2 ) P/p62 and insoluble ubiquitinated protein , Drosophila Gba1b mutants show results comparable to those seen in vertebrate models of GCase deficiency [10 , 15 , 68 , 69 , 71] . Our proteomic measurements of protein abundance are also consistent with previous reports of increased lysosomal mass in GCase deficiency [1 , 8 , 66] . The abundance of the lysosomal marker Lamp1 was nearly tripled in Gba1b mutants , and 41% of lysosomal proteins were significantly increased in abundance ( S1 Data ) . Nevertheless , our protein turnover measurements reveal that the overall rates of degradation through lysosomal processes are not grossly altered . Thus , one possible explanation of our findings is that the efficiency of autolysosomal degradation is decreased , with lower throughput per unit of autolysosomal mass , but that the organism has compensated by increasing the amount of autolysosomal machinery available . Because this compensation is sufficient to maintain degradation rates , we would describe Gba1b mutants as being under autolysosomal stress rather than in autolysosomal failure . Over time , the degree of stress may exceed the capacity to compensate , and aged Gba1b mutants may show overt failure of lysosomal degradation . Even if this is the case , late failure of autolysosomal degradation cannot explain the behavioral and biochemical abnormalities that begin in early adulthood [8 , 9] . Another explanation for the apparent discrepancy between our findings of normal autophagic substrate turnover and previous reports of impaired autophagy is that commonly used autophagy markers are not solely representative of autophagic flux [57 , 72] . This is especially true of Ref ( 2 ) P , or p62 , which has multiple nonautophagic functions and is transcriptionally upregulated by stress [57 , 73] . In addition , p62 and LC3 have recently been detected in mammalian EVs [47 , 74 , 75] , and we found increased levels of oligomeric Ref ( 2 ) P in EVs from Gba1b mutants ( Fig 8 ) . It is therefore possible that the increased Ref ( 2 ) P levels detected in Gba1b mutants result from a combination of stress response and EV dysregulation . Our work leaves unanswered the question of how GCase deficiency results in increased EV abundance , but does suggest two possible explanations . Increased production of EVs could be caused either by lysosomal stress or by changes in membrane lipid composition . Lysosomal stress has been shown in cultured cells to promote the release of exosomes , a major type of EV [75 , 76] . Exosomes are generated when a multivesicular endosome ( MVE ) fuses with the plasma membrane rather than the lysosome , releasing its intraluminal vesicles into extracellular space [41 , 54] . Lysosomal blockade increases the probability that an MVE will fuse with the plasma membrane [75 , 76] . If lysosomal stress rather than outright failure is sufficient to trigger increased exosome release , it could account for the overabundance of EVs in Gba1b mutants . A second explanation for increased EVs in GCase-deficient animals is that abnormal membrane lipid composition may directly alter EV biogenesis . Lipid composition determines membrane fluidity and curvature , and thus controls the size , shape , and fusion kinetics of EVs [77–79] . In fact , lipid rafts , particularly those enriched in ceramide , are required for formation at least one type of EV [78] . Membrane changes such as those caused by GCase deficiency , including accumulation of glucosylceramide and altered ceramide levels [80 , 81] , could alter EV functioning at any stage from formation to internalization by a recipient cell . Either increased or decreased probability of ceramide-dependent EV formation could lead to increased overall EV production , as suppression of one type of EV has been shown to cause overproduction of another type [82] . While understanding the mechanism by which GCase deficiency causes increased EV release is an important goal of future work , an equally important question is how increased EV abundance in Gba1b mutants promotes the accumulation of protein aggregates . EVs have been increasingly implicated in the pathogenesis of neurodegenerative disease . Many disease-associated proteins , including prion protein , α-synuclein , β-amyloid , and tau , are detected in EVs [41 , 83 , 84] , which have been proposed as vehicles for the well-documented progressive spread of protein aggregates from one brain region to another [83 , 85 , 86] . In support of this model , toxic forms of these disease-associated proteins are more abundant in EVs from humans with neurodegenerative diseases such as Alzheimer disease , dementia with Lewy bodies , and Parkinson disease ( PD ) [84 , 87 , 88] , and EVs from these patients can induce protein aggregation in recipient cells under experimental conditions [89 , 90] . However , progression of these diseases has not yet been conclusively demonstrated to be mediated by EVs . Perhaps the strongest evidence that EVs promote the spread of protein aggregates has been found for prion protein . Stimulating the release of EVs increased the cell-to-cell spread of misfolded prion protein , and decreasing EV release reduced the spread [91] . Our findings appear to follow the same pattern: genetic interference with EV production suppressed protein aggregation in Gba1b mutants . If the same holds true for other aggregation-prone proteins , conditions that increase EV release could promote the spread of protein aggregates and thus be risk factors for neurodegenerative disease . Fig 11 illustrates this model . When GCase activity is normal ( Fig 11A ) , EVs travel between cells , carrying both factors that promote protein aggregation ( e . g . , disease-associated proteins such as α-synuclein ) [88 , 92] and factors that oppose it ( e . g . , chaperones ) [93] . Some cells likely generate more aggregates than others , and may therefore release more aggregate-promoting factors , including small aggregate “seeds . ” Quality control mechanisms in recipient cells successfully combat protein aggregation , and aggregates accumulate only slowly with age . If GCase activity is absent or reduced , however ( Fig 11B ) , more EVs are generated; this results in greater cell-to-cell transfer of aggregate-prone proteins , perhaps simply because these proteins are normally part of EV cargo . In particular , they may be normal cargo of ESCRT-dependent EVs , given our finding that knockdown of ESCRTs in Gba1b mutants ameliorated the mutants’ protein aggregation phenotype . Alternatively , GCase deficiency may alter cargo selection so that more aggregation-prone proteins are loaded into EVs . The net effect of the EV changes is transfer of aggregation-producing factors in quantities that overwhelm quality control mechanisms , leading to excessive accumulation of ubiquitin-protein aggregates in recipient cells . GBA mutations are the strongest single risk factor for PD and dementia with Lewy bodies , affecting up to 10% of PD patients worldwide [2 , 5] . Our finding that GCase deficiency causes increased EV release offers new insight into these prevalent disorders . For example , increased transmission of protein aggregates via EVs could explain the earlier onset and faster disease progression in PD patients with GBA mutations [6 , 94–97] . Future investigations should determine how glucocerebrosidase deficiency increases EV abundance , and how manipulations of EV production might prevent or delay the progression of neurodegenerative disease .
Fly stocks were maintained on standard cornmeal-molasses food at 25°C . The Gba1b null ( Gba1bΔTT ) , Gba1b control ( Gba1brv ) , Atg7d4 , Atg7d77 , Sod2n283 , Sod2wk , park25 , PINK1B9 , and PINK1rv alleles , as well as the UAS-PINK1#2 strain , have been previously described [9 , 19 , 20 , 98 , 99] . The UAS-ALiX-HA strain was obtained from the former Bangalore Fly center ( National Centre for Biological Sciences , Bangalore , India ) . The UAS-Ref ( 2 ) P-RNAi strain ( v108193 ) was obtained from the Vienna Drosophila Resource Center . Other strains and alleles were obtained from the Bloomington Stock Center: elav-GAL4 ( 458 ) , Act5C-GAL4 ( 3953 ) , UAS-Hsc70-4 ( 5846 ) , w1118 ( 3605 ) , UAS-Rab11-GFP ( 8506 ) , Gba1bMB03039 ( 23602 ) [100] , UAS-Mvb12-RNAi ( 43152 ) , UAS-larsen-RNAi ( 38289 ) [101] , and UAS-CHMP2B-RNAi ( 38375 ) [102] . Atg7 null mutants were Atg7d4/Atg7d77 transheterozygotes . Sod2 mutants were null/hypomorph compound heterozygotes ( Sod2n283/Sod2wk ) . The full genotype of parkin mutants was If/CyO; park25/park25 . The WT controls for Atg7 and parkin mutants were a composite dataset derived from four groups of healthy flies with intentionally diverse genetic backgrounds ( see protein turnover rate calculations section ) . The control for PINK1B9 was its revertant ( precise excision ) strain , PINK1rv , and the control for Sod2 was CyO/+ . The control strain for Gba1b was the revertant Gba1brv . In Fig 7 we used the following genotypes for the experiments involving the ALiX-HA transgene: control = Gba1brv/Gba1bMB03039; Gba1b = Gba1bΔTT/Gba1bMB03039 . This combination of Gba1b mutant alleles , which we used for ease of recombination with the ALiX-HA transgene , produced the same biochemical abnormalities found in Gba1bΔTT homozygotes ( S5 Fig ) . Lipidomic analysis was performed at the Northwest Metabolomics Research Center at the University of Washington , Heads were isolated from 10-day-old control and Gba1b flies flash-frozen in liquid nitrogen , and lipids were then extracted from the frozen head tissue . Levels of glucosylceramide and ceramide were measured by a high-performance liquid chromatography/mass spectrometry ( LC-MS/MS ) method , using a sphingolipids mix as internal standard ( Avanti Sphingolipids Mix II LM-6005 ) . Results were expressed as lipid levels per mass of starting tissue . For each lipid species , three independent samples were analyzed . [5 , 5 , 5 – 2H3] leucine ( D3-leucine; 99 atom % deuterium ) was obtained from Isotec/Sigma-Aldrich . Synthetic complete medium without leucine ( C-Leu ) was supplemented with glucose and 60 mg/L D3-leucine . A strain of Saccharomyces cerevisiae auxotrophic for leucine ( BB14-3A , Brewer Lab , University of Washington [103] ) was grown to saturation at 30°C , then spun down , flash-frozen in liquid nitrogen , lyophilized , and stored at −80°C . Because brewing in-house produced limited quantities of labeled yeast , we made labeled fly food in batches of ~40 mL using a microwave . We did this by substituting cornstarch for cornmeal in the lab’s standard recipe ( 2 . 35% yeast w/v ) and dispensing the cooked food in small amounts into vials lined with wet Whatman paper to maintain moisture . Unlabeled transition food for the first 24 hours after eclosion was made and dispensed in the same way , substituting Red Star yeast . Atg7 , parkin , PINK1 , and Sod2 mutant samples were processed as previously described [21] . GBA1b and old/young samples were processed as follows: Fused silica microcapillary columns of 75 μm inner diameter ( Polymicro Technologies , Phoenix , AZ ) were packed in-house by pressure loading 30 cm of Jupiter 90 Å C12 material ( Phenomenex ) . Kasil ( PQ Corporation ) frit microcapillary column traps of 100 μm inner diameter with a 2-mm Kasil frit were packed with 4 cm of Jupiter 90 Å C12 . A retention time calibration mixture ( Pierce ) was used to assess quality of the column before and during analysis . Three of these quality control runs were analyzed prior to any sample analysis , and another quality control run was performed after every six sample runs . One microgram of each sample digest and 150 femtomoles of the quality control sample were loaded onto the trap and column by the NanoACQUITY UPLC system ( Waters Corporation ) . Buffer solutions used were water , 0 . 1% formic acid ( buffer A ) , and acetonitrile , 0 . 1% formic acid ( buffer B ) . The 60-minute gradient of the quality control consisted of 30 minutes of 98% buffer A and 2% buffer B , 5 minutes of 65% buffer A and 35% buffer B , 10 minutes of 40% buffer A and 60% buffer B , 5 minutes of 95% buffer A and 5% buffer B , and 18 minutes of 98% buffer A and 2% buffer B at a flow rate of 0 . 3 μL/min . The 240-minute gradient for the sample digest consisted of 120 minutes of 98% buffer A and 2% buffer B , 80 minutes of 65% buffer A and 35% buffer B , 20 minutes of 20% buffer A and 80% buffer B , and 20 minutes of 98% buffer A and 2% buffer B at a flow rate of 0 . 25 μL/min . Peptides were eluted from the column and electrosprayed directly into an Q-Exactive HF mass spectrometer ( Thermo Fisher ) with the application of a distal 3 kV spray voltage . For the quality control analysis , a cycle of one 60 , 000 resolution full-scan mass spectrum ( 400–1600 m/z ) was followed by 17 data-independent MS/MS spectra using an inclusion list at 15 , 000 resolution , 27% normalized collision energy with a 2 m/z isolation window . For the sample digests , a cycle of one 120 , 000 resolution full-scan mass spectrum ( 400–1600 m/z ) followed by 20 data-dependent MS/MS spectra on the top 20 most intense precursor ions at 15 , 000 resolution , 27% normalized collision energy with a 1 . 5 m/z isolation window . Application of the mass spectrometer and UPLC solvent gradients was controlled by the Thermo Fisher XCalibur data system . The quality control sample data were analyzed using Skyline [23] . High-resolution MS data were processed by BullsEye to optimize precursor mass information [22] . The MS/MS output was searched using COMET [104] with differential modification search of 3 . 0188325 Da for leucine and 15 . 994915 methionine and a static modification of 57 . 021461 Da for cysteine , against a FASTA database containing all the protein sequences from FlyBase plus contaminant proteins . Peptide-spectrum match false discovery rates were determined using Percolator [105] at a threshold of 0 . 01 , and peptides were assembled into protein identifications using an in-house implementation of IDPicker [106] . Turnover rates were calculated using Topograph software [22] . For a full description of Topograph settings , see Vincow et al . [21] . A protein’s turnover rate was computed based on data from all peptides detected , and values from all biological replicates were pooled for turnover calculations . A protein’s turnover rate was calculated based on at least 6 measurements per genotype of percent turnover for GBA1b mutants and old/young flies , and at least 15 measurements per genotype for Atg7 , parkin , PINK1 , or Sod2 mutants . Peptides that could be the product of more than one gene were excluded from analysis . For a small percentage of genes ( 2%-5% ) , Topograph clustered peptides corresponding to a single gene into 2-3 nonoverlapping “isoform groups . ” For example , isoform group 1 might include peptides mapping only to the COX6B-PA isoform , while isoform group 2 peptides could have come from COX6B-PA , -PB , or -PC . While in most cases the isoform groups for a single protein had essentially identical turnover rates , occasionally they displayed significant differences in turnover behavior . Each isoform group was therefore analyzed as a separate protein . We excluded proteins with excessive inter-replicate variability of turnover rates , defined as coefficient of variation ≥ 0 . 25 . We calculated the turnover rate separately for each biological replicate and determined the coefficient of variation across replicates . Proteins were analyzed only if they met inclusion criteria in both mutants and controls . In previous work , we had compared Atg7 and parkin null mutants to their respective heterozygotes [21] . However , we later found that both Atg7 and parkin heterozygotes had mild but significant slowing of mitochondrial protein turnover compared to WT flies , and we selected the WT dataset as a more appropriate control . For turnover analyses , Atg7 and parkin nulls were both compared to a composite WT dataset derived from four separate groups of healthy flies ( w1118 , PINK1rv , CyOActGFP/+ , and a mixture of`CyO/Hsp70-GAL4 and CyO/UAS-PINK1#2 ) . Turnover rates are the mean values for all genotypes in which the protein was detected; the rates are highly consistent across genotypes , as previously reported [21] . Each mean value for a genotype was treated as one replicate for statistical purposes . Statistical significance of fold change in turnover was calculated for groups of proteins using nested ANOVA [107] , and significance of change for individual proteins was calculated using t tests . The following subgroups of proteins had enough replicates for t tests: 148 mitochondrial , 36 ribosomal , 15 ER/peroxisomal , and 275 nonorganellar proteins . We measured protein abundance from the same raw mass spectrometry data used in the turnover study , using Skyline [23] and MSstats [24] . Prior to MSstats analysis , we obtained total abundance ( labeled plus unlabeled ) for each peptide using a custom R script . The statistical significance of intergroup differences was calculated using a linear mixed model , then adjusted for multiple comparisons by the Benjamini-Hochberg procedure with a false discovery rate of 0 . 05 . All abundance comparisons were made at the second time point , when differences between genotypes were most marked . In abundance analyses , parkin and Atg7 mutants were compared to their original heterozygote controls rather than WT flies ( see calculations above ) . While the composite control group approach was appropriate for measurement of turnover , which is more consistent and less noisy than abundance [21] , measurement of relative protein abundance required mutant and control samples that had been run at the same time . General: Drosophila protein localization was determined from a variety of resources including gene and protein information databases ( FlyBase [108] , MitoDrome [109] , NCBI [110] , UniProt [111] ) , protein localization prediction algorithms ( WoLF PSORT [112] , MitoProt [113] , Predotar [114] , SignalP [115] , NucPred [116] , and PTS1 Predictor [117 , 118] ) , BLAST [119] , and primary literature . Proteasome substrates: We identified proteins as proteasome substrates ( Fig 4 ) if their mammalian orthologs had one or more regulated ubiquitinated sites according to Wagner et al . [35] . These sites showed altered abundance of ubiquitinated peptides after proteasome inhibitor treatment . We identified Drosophila orthologs of proteins from the Wagner et al . data with the DRSC Integrative Ortholog Prediction Tool ( DIOPT ) v6 [50] , minimum score 5 . Microautophagy substrates: We identified microautophagy substrates by searching for targeting sequences ( Fig 4 ) , also called KFERQ-like motifs . These motifs were defined as sequences of five amino acids ( AAs ) that fit criteria established by Dice [120 , 121]: We wrote an algorithm using Python 2 . 7 to search protein sequences for these motifs and applied it to the fly proteome ( FASTA sequences downloaded from FlyBase ) . We then identified cytosolic proteins by annotation as described above , and compared the effects of GCase deficiency on cytosolic proteins with and without KFERQ-like sequences . Endocytic turnover substrates: Proteins designated endocytic turnover substrates in Fig 5 were identified using FlyBase annotation and search terms such as receptor , transmembrane , extracellular matrix , integral component of plasma membrane , and channel . Endosomal machinery proteins ( see below ) were excluded . Endosomal machinery: Proteins designated “endosomal machinery” in Fig 5 were identified by a FlyBase search for the string “endosom*” in at least one of the following fields: GO Molecular Function , GO Biological Process , GO Cellular Component , Gene Snapshot , or UniProt Function . Extracellular vesicle proteins: To identify extracellular vesicle proteins , we compiled a list of proteins detected in EVs in mass spectrometry studies of Drosophila cultured cells [46–49] . The list contained 544 unique proteins , 329 of which were found in Gba1b mutant protein turnover data and 499 in abundance data . In addition , we obtained from ExoCarta [47] the list of “top 100 [mammalian] proteins that are often identified in exosomes , ” and identified 97 Drosophila orthologs of these proteins using DIOPT v6 as previously described [50] . Fifty-nine proteins from this list were found in Gba1b mutant turnover data and 86 in abundance data . The significance of intergroup differences was evaluated using the Fisher exact test except when the total number of proteins was too large , in which case we performed a χ2 test of homogeneity . Proteasome activity was measured in heads from male and female flies 10 to 11 days old ( 50 per sample ) according to the method of Tsakiri et al . [122] , with the following modifications: We used 26S lysis buffer only . We obtained substrate buffer and fluorescent substrates from the UBPBio Proteasome Activity Fluorometric Assay Kit II ( J4120 ) , and we used epoxomicin 20 μM for proteasome inhibition . Specifically , we divided the lysate in half and added DMSO to one half and epoxomicin to the other . We measured the protein concentration of lysates using the Pierce BCA Protein Assay Kit ( 23227 ) , and measured sample fluorescence with a Synergy H1 BioTek plate reader ( excitation 350 nm , emission 450 nm ) . We subtracted the activity measured in the epoxomicin-treated homogenate from the activity in the DMSO-treated homogenate . The experiment was repeated three times . For total EVs ( tEVs ) , hemolymph was obtained from 20 flies ( 10 males and 10 females , 10 to 11 days old ) per sample . In order to obtain whole-fly homogenate from the same animals used for collection of hemolymph , hemolymph was extracted manually from the first four flies and their bodies were reserved for later use . These flies were decapitated , following which their hemolymph was collected by pressing on the thorax with the head of a butterfly pin . The hemolymph from these four flies was collected by capillary action into 1 μL PBS , and the total sample was transferred to a 1 . 7-mL microfuge tube containing 9 μL PBS . The heads and bodies of the four flies were then homogenized in RIPA buffer for whole-fly protein homogenates . Two holes were made with a 25-g needle in the bottom of a 0 . 5-mL tube , and 16 more flies were decapitated and placed in this tube . The 0 . 5-mL tube was then seated in the PBS-containing 1 . 7-mL tube for centrifugation . The tubes were centrifuged at 5000 x g for 5 min at 4°C , after which the extracted hemolymph was centrifuged for 30 min at 10 , 000 x g at 4°C to remove cell debris and the cell-free supernatant was collected . An equal volume of 2x Laemmli buffer ( 4% SDS , 20% glycerol , 120 mM Tris-Cl pH 6 . 8 , 0 . 02% bromophenol blue , 2% β-mercaptoethanol ) was added to the cell-free supernatant and also to the whole-fly protein homogenates , and all samples were boiled for 10 min and then stored at −80°C . The experiment was repeated at least three times . Small EVs ( sEVs ) were prepared as for tEVs , with the following changes: 50–60 adult flies were used per sample . The hemolymph was collected into a volume of PBS scaled to the number of flies used ( 1 μL/fly ) to minimize sample loss during filtration . After the 10 , 000 x g spin , Total Exosome Isolation Reagent for Cell Culture ( Thermo Fisher/Invitrogen , 4478359 ) was used as in Tassetto et al . [123] except that we used Ultrafree 0 . 22 μm spin filters ( Fisher , UFC30GV0S ) . The resulting filtrate was boiled and stored as for the tEVs . Heads from 10-day-old flies ( 6 females and 6 males per sample ) were homogenized in Triton lysis buffer ( 50 mM Tris-HCl pH 7 . 4 , 1% Triton X-100 , 150 mM NaCl , 1 mM EDTA ) , and then spun at 15 , 000 x g for 20 min . The detergent-soluble supernatant was collected and mixed with an equal volume of 2x Laemmli buffer , and the same buffer was used to resuspend the Triton-insoluble pellet . All samples were boiled for 10 minutes . The Triton-insoluble protein extracts were then cleared of debris by centrifugation at 15 , 000 x g for 10 minutes , followed by collection of the supernatant . At least three independent experiments were performed . Proteins were separated by SDS-PAGE on 4%-20% MOPS-acrylamide gels ( GenScript Express Plus , M42012 ) and electrophoretically transferred onto Immobilon PVDF membranes ( Fisher , IPVH00010 ) . Immunodetection was performed using the iBind Flex Western Device ( Thermo Fisher , SLF2000 ) . Antibodies were used at the following concentrations: 1:25 , 000 mouse anti-Actin ( Chemicon/Bioscience Research Reagents , MAB1501 ) , 1:250 mouse anti-Rab11 ( BD Transduction Laboratories , 610657 ) , 1:200 rabbit anti-Ref ( 2 ) P ( Abcam , ab178440 ) , 1:500 mouse anti-ubiquitin ( Santa Cruz , sc-8017 ) , 1:800 mouse anti-Cnx99A ( DHSB , Cnx99A 6-2-1 ) , 1:100 mouse anti-Golgin-84 ( DHSB , Golgin84 12–1 ) , and 1:500 rat anti-HA ( Sigma-Aldrich , 11867423001 ) . HRP secondary antibodies were used as follows: 1:500 to 1:1000 anti-mouse ( BioRad , 170–6516 ) , 1:100 anti-rat ( Sigma-Aldrich , A9037 ) , and 1:500 to 1:1000 anti-rabbit ( BioRad , 172–1019 ) . Signal was detected using Pierce ECL Western Blotting Substrate ( Fisher , 32106 ) . Densitometry measurements of the western blot images were measured blind to genotype and condition using Fiji software [49] . For homogenates , signal was normalized either to Actin or to Ponceau-S [124 , 125] . For EVs , signal was normalized to loading volume . Normalized western blot data were log-transformed when necessary to stabilize variance before means were compared using Student t test . Each experiment was performed at least three times . EVs were prepared for nanoparticle tracking analysis ( NTA ) as described for western blotting through the 10 , 000 x g step , after which they were passed through a 0 . 65 μm Ultrafree-MC filter ( Fisher , UFC30DV0S ) to ensure removal of any remaining cellular debris and stored at −80°C . Hemolymph was obtained from 60 flies per sample , and four biological replicates per genotype were collected . EV size and concentration were measured using NTA by Alpha Nano Tech LLC ( Research Triangle Park , NC ) . NTA was performed using a ZetaView instrument equipped with an sCMOS camera and 532 nm laser . Instrument parameters were as follows: temperature setting 23°C , Max Area 500 , Min Area 20 , Min Brightness 20 . Two cycles of analysis at 11 positions were performed for each sample . Data were analyzed using ZetaView software version 8 . 04 . 02 . Standard laboratory protection equipment was used during all steps of sample preparation and analysis to prevent sample contamination with dust particles . The 1x PBS solution ( Amresco ) used to dilute samples was filtered on the day of analysis through a 0 . 22 μm Millex-GV syringe filter ( Millipore ) , and its purity was confirmed by NTA analysis prior to the study . Instrument qualification was performed by analyzing a polystyrene bead standard ( 100 nm , Particle Metrix ) in 1x PBS prior to each study . Instrument accuracy and precision were confirmed to ± 5% of the target value . | Mutations in the GBA gene , which encodes the enzyme glucocerebrosidase , are common and increase the risk of Parkinson disease . A widely accepted explanation for the increased risk is that the fatty substance normally broken down by glucocerebrosidase builds up in the lysosome , which is the cell’s recycling center , until the cell can no longer get rid of damaged parts . At that point , proteins that should be destroyed in the lysosome form large clumps ( aggregates ) throughout the cell . We used mutant fruit flies without glucocerebrosidase to test this theory , and we were surprised to see no evidence that the lysosome was failing . The destruction of proteins usually recycled by the lysosome was not slowed down in the mutant flies . Instead , we saw evidence that the mutants’ cells might be producing too many extracellular vesicles , tiny spheres that transport cargo and messages from cell to cell . Some researchers have also suggested that extracellular vesicles carry the protein aggregates that spread between cells as Parkinson disease get worse . Our study supports this idea . It suggests that increased spread of aggregates through extracellular vesicles , rather than failure of the lysosome , might explain why GBA mutations increase the risk of neurodegenerative disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cell",
"death",
"invertebrates",
"autophagic",
"cell",
"death",
"lysosomes",
"vesicles",
"protein",
"aggregation",
"cell",
"processes",
"animals",
"animal",
"models",
"membrane",
"proteins",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
... | 2018 | Glucocerebrosidase deficiency promotes protein aggregation through dysregulation of extracellular vesicles |
What does it take to convert a heap of sequencing data into a publishable result ? First , common tools are employed to reduce primary data ( sequencing reads ) to a form suitable for further analyses ( i . e . , the list of variable sites ) . The subsequent exploratory stage is much more ad hoc and requires the development of custom scripts and pipelines , making it problematic for biomedical researchers . Here , we describe a hybrid platform combining common analysis pathways with the ability to explore data interactively . It aims to fully encompass and simplify the "raw data-to-publication" pathway and make it reproducible .
Trees , rivers , and the analysis of next generation sequencing ( NGS ) data are examples of branching systems so ubiquitous in nature [1] . Indeed , numerous types of NGS applications ( i . e . , variation detection , analyses of DNA/Protein interactions [ChIP-seq] or transcriptome [RNA-seq] ) share the same initial processing steps ( quality control , read manipulation and filtering , mapping , post-mapping thresholding , etc . ) , making up the trunk and main branches of this tree . Each of these main branches subsequently gives off smaller offshoots ( variant calling , RNA-seq , ChIP-seq , and other "seqs" ) that , in turn , split further as analyses become focused towards the specific goals of an experiment . As we traverse the tree , the set of established analysis tools becomes increasingly sparse , and it is up to an individual researcher to come up with statistical and visualization approaches necessary to reach the leaves ( or fruits ) that represent conclusive , publishable results . Consider transcriptome analysis as an example . Initial steps of RNA-seq analysis ( in our tree analogy , these are trunk and main branches ) , such as quality control , read mapping , and transcript assembly and quantification are reasonably well established . Yet completion of these steps does not produce a publishable result . Instead , there is still the need for additional analyses ( progressively smaller branches of our tree ) , ranging from simple format conversion to statistical tests and visualizations . Thus , every NGS analysis can , in principle , be divided into two stages . The first stage involves processing of raw data using a small set of common , generic tools . This stage can be scripted and automated and also lends itself to building graphical user interfaces ( GUIs ) . The second stage involves a much greater variety of tools that need to be customized for every given experiment ( in many cases , there are no tools at all , and custom scripts need to be developed ) . As a result , it is not readily coerced into a handful of automated routines or generic GUIs . The main motivation for this work was the development of a system wherein biomedical researchers can perform both stages of data analysis: initial steps using established tools and exploratory and data interpretation steps with ad hoc approaches . Merging both steps into a unifying platform will lower entry barriers for individuals interested in data analysis , significantly improve reproducibility of published results , ease collaborations , and enable straightforward dissemination of best analysis practices .
Jupyter integration into Galaxy takes advantage of the recently developed and increasingly popular Docker containerization platform ( https://www . docker . com ) . It uses the Interactive Environment ( IE ) plug-in functionality written for Galaxy that also allows integration of other similar tools such as RStudio . It consists of an Interactive Environment Entry Point ( IEEP ) and an associated configuration file . The IE configuration allows administrators to set it so that all data transfer is done via Secure Socket Layer ( SSL ) , which is useful for production instances . Additionally , individual sites can specify custom Docker images instead of the default provided Jupyter notebook , allowing administrators to craft Docker images more specific to their users . The default image will be downloaded and installed from Docker Hub or quay . io—popular Docker image hosting services . The default Docker image is specifically crafted for use in conjunction with the Jupyter Interactive Environment ( see below ) . The IEEP launches a Docker container on a random port for communication and configures it to access Galaxy through environment variables passed to the container . This container , by the very nature of Docker itself , is isolated from the filesystem and processes on the Galaxy server . In addition , administrators can configure Docker containers to run on remote computing resources using Docker's built-in client/server architecture . Doing so also provides an additional layer of security by fully resource-separating the IE container from the Galaxy server . For greater scalability , Docker Swarm , the distributed Docker container engine provided with the Docker software , is supported . The host and port on which the container is running are stored in a database on the Galaxy server so that Galaxy and the Dockerized Jupyter web service can communicate securely while isolated from the rest of the Galaxy instance for security reasons . The container is built on top of the official “jupyter/minimal-notebook” image ( which is maintained by Project Jupyter ) and provides a Jupyter server , along with its dependencies , such as NumPy [2] , SciPy [3] , and Matplotlib [4] . Additionally , the image contains several Jupyter kernels ( different programming language environments ) , such as R , Ruby , Haskell , Julia , and Octave . By utilizing a Docker image with a full suite of scientific analysis tools and libraries , users are able to immediately perform their analysis and calculations . In the Python kernel , additional packages can be installed with the python package manager called “pip . ” The same is true for the other kernels and their associated package managers . Moreover , tools that can be installed and run in a nonprivileged user account can be added to the container on demand . Once the container has launched on the backend , it is embedded inside the Galaxy interface , at which point it can be used to interactively program , develop , and analyze data in any of the aforementioned programming languages . Each invocation of the IE by a Galaxy user results in the launch of a new Docker container , meaning that users are isolated from each other . If the page with the Interactive Environment is closed by the user , Galaxy instructs Docker to terminate the process . Additionally , during Docker container startup , a service is launched that monitors whether the IE is still being used by checking the network traffic so that it can automatically terminate itself when the IE is no longer in use . Within the Jupyter notebook , two important custom functions are defined that enable the user to load data from the history or store data to the Galaxy history using the Galaxy API [5] . The “get” function expects one parameter: the numerical identifier of the dataset as shown in the history . The retrieved dataset is stored as a file inside the container , which can be accessed via the usual means for the language kernel in use ( e . g . , the “open” function ) . The “put” function automatically builds a connection to the host Galaxy instance and transfers a specified file from inside the Docker container to the user's history . Thus , any dataset the user has access to in Galaxy can be loaded into the notebook , datasets can be combined or modified programmatically , and the results can be written back to the history . The entirety of Galaxy–IE communication occurs between the Galaxy host and the Docker container , without the need for the user to upload or download data to their personal workstation . This is not only faster in most cases , but it also has positive implications on data security , as the data did not leave the compute center . In addition to the SSL-secured dataset transfer already mentioned , all of Docker's security and resource control features are available to the administrator . These include CPU and memory limits and SSL-secured client/server communication . Additionally , every container can be password protected if desired—a password is randomly generated and presented to the user during startup of the container in his/her web browser . Notebooks can be saved to the Galaxy history at any time; once in Galaxy's history , they can be inspected like any other Galaxy dataset , allowing for a read-only view of the analysis steps that we run . Additionally , notebooks can be reused . A new Jupyter instance is created that retains the stored work . This functionality ensures the reproducibility of data analysis and is therefore an essential feature of the Jupyter Interactive Environment .
HIV-1 was resequenced from the blood of a single individual across three time points with the ultimate goal of tracking nucleotide substitutions of the viral genome through time ( simulated reads were generated for this example ) . After assessing the quality of the reads and mapping against the HIV-1 genome with bwa [15] within Galaxy , we wanted to visualize read coverage across each sample to decide if further analyses are warranted . However , the main public Galaxy server did not have a dedicated tool for this purpose . Normally , the analysis will stop at this point , and only by downloading data and analyzing them offline can one produce the coverage distribution graph needed in this case . Integration of Jupyter to Galaxy changes this . Fig 1 highlights each step of this analysis , resulting in the coverage distribution graph . The entire analysis can be seen in the Galaxy history , accessible at http://bit . ly/ie-hiv ( see S1 File ) . In this example , we use a subset of RNA-seq data from a dataset published by Schurch et al . [16] ( SRA accession ERP004763 ) consisting of 48 replicates of two Saccharomyces cerevisiae populations: wildtype and snf2 knock-out mutants . For simplicity , we selected only two replicates for each wildtype and snf2 knock-outs . Here , we first use Galaxy's existing RNA-seq tools to map reads against the yeast genome using HiSat [17] and to compute the number of reads per gene region using HTseq-count [18] ( Fig 1B; see Galaxy history at http://bit . ly/rnaseq-jupyter and S1 Fig ) . Datasets are then imported into Jupyter's environment ( cells 4–9; see S2 File ) , where we first merge datasets into a single table by joining them on gene names using Python's Pandas library ( cells 10–12 ) . We then proceeded to normalize the counts with DESeq2 [19] ( cells 13–30 ) and assessed the effects of normalization and variance shrinkage on the data ( cells 31–35; also see center pane of Fig 1B ) . In the third example , we replicate the key analyses reported in a study of human mitochondrial heteroplasmy transmission dynamics , previously published by our group [14] . The goal of this study was to detect heteroplasmies ( variants within mitochondrial DNA ) and to trace their frequency changes across mother–child transmission events using primary sequencing data generated by [14] ( mitochondria is transmitted maternally , and heteroplasmy frequencies may change dramatically and unpredictably during the transmission due to a germ-line bottleneck [20] ) . The first part of the analysis is performed using Galaxy's mapping and variant calling workflow outlined in Fig 2A . The goal of this part is to generate a preliminary list of sequence variants . The input data consist of over 118 GB of sequencing reads corresponding to 312 fastq datasets ( SRA accession SRP047378 ) derived from 156 samples ( 39 mothers and 39 children , with two tissues analyzed per individual , each tissue generating two fastq datasets for the forward and reverse read sets , together resulted in the 312 original datasets; Fig 2B , dataset 313 ) . Using Galaxy , we combine all 312 datasets into a single entity , a dataset collection , in order to avoid repetitive tasks ( see Galaxy history at http://bit . ly/jupyter-mt , S1 Fig and Fig 2B ) . The workflow maps the reads and performs de-duplication and extensive filtering of resulting BAM datasets , as well as identifies variable sites . The workflow reduces sequencing reads to a 160 MB data matrix with over 2 . 6 million rows containing variants for all 156 samples . Despite the fact that we have reduced the primary sequence data to a set of variable sites , this dataset hardly resembles an interpretable result . At this point , exploratory analyses must begin . Unfortunately , it is also the point at which users are forced to leave Galaxy , confounding efforts for reproducibility of the analysis . With Galaxy/Jupyter integration , this deficiency can be avoided . The second part of the analysis begins with starting a Jupyter notebook from inside the Galaxy interface . It proceeds through numerous custom data processing steps and statistical analyses outlined in S3 File . Two main conclusions of this analysis are the positive correlation between the age of the mother and the number of heteroplasmic sites , which has potential implications for the higher rate in mitochondrial DNA ( mtDNA ) diseases in children born to older mothers ( Fig 2C ) , and the very small size of mitochondrial bottleneck ( Fig 2D ) at only approximately 40 segregating units . The above three examples highlight the power of combining ad hoc programmatic analyses with a collection of robust tools already provided by Galaxy . In our opinion , this has the potential to streamline the ways in which biomedical data analysis is performed . In particular , we see the following implications: | Galaxy users can utilize a large number of tools and workflows . What they could not previously do is run ad hoc scripts and arbitrary tools within their Galaxy instance . This was very limiting , as initial analyses of data often involve interactive exploration with tools like Jupyter or RStudio—powerful platforms that are becoming increasingly popular in life sciences . Here , we showcase Galaxy Interactive Environment framework , designed to combine Galaxy's tools and workflows with environments such as Jupyter . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results/Discussion"
] | [
"sequencing",
"techniques",
"education",
"mitochondrial",
"dna",
"astronomical",
"sciences",
"invertebrate",
"genomics",
"next-generation",
"sequencing",
"genome",
"analysis",
"forms",
"of",
"dna",
"energy-producing",
"organelles",
"molecular",
"biology",
"techniques",
"bio... | 2017 | Jupyter and Galaxy: Easing entry barriers into complex data analyses for biomedical researchers |
In order to propagate a solid tumor , cancer cells must adapt to and survive under various tumor microenvironment ( TME ) stresses , such as hypoxia or lactic acidosis . To systematically identify genes that modulate cancer cell survival under stresses , we performed genome-wide shRNA screens under hypoxia or lactic acidosis . We discovered that genetic depletion of acetyl-CoA carboxylase ( ACACA or ACC1 ) or ATP citrate lyase ( ACLY ) protected cancer cells from hypoxia-induced apoptosis . Additionally , the loss of ACLY or ACC1 reduced levels and activities of the oncogenic transcription factor ETV4 . Silencing ETV4 also protected cells from hypoxia-induced apoptosis and led to remarkably similar transcriptional responses as with silenced ACLY or ACC1 , including an anti-apoptotic program . Metabolomic analysis found that while α-ketoglutarate levels decrease under hypoxia in control cells , α-ketoglutarate is paradoxically increased under hypoxia when ACC1 or ACLY are depleted . Supplementation with α-ketoglutarate rescued the hypoxia-induced apoptosis and recapitulated the decreased expression and activity of ETV4 , likely via an epigenetic mechanism . Therefore , ACC1 and ACLY regulate the levels of ETV4 under hypoxia via increased α-ketoglutarate . These results reveal that the ACC1/ACLY-α-ketoglutarate-ETV4 axis is a novel means by which metabolic states regulate transcriptional output for life vs . death decisions under hypoxia . Since many lipogenic inhibitors are under investigation as cancer therapeutics , our findings suggest that the use of these inhibitors will need to be carefully considered with respect to oncogenic drivers , tumor hypoxia , progression and dormancy . More broadly , our screen provides a framework for studying additional tumor cell stress-adaption mechanisms in the future .
Most solid tumors have substantial physiological deviations from normal tissue , which manifest as tumor microenvironment ( TME ) stresses [1 , 2] . These TME “stresses” include , among others , the limited availability of oxygen ( hypoxia ) , glucose or amino acids , and an accumulation of lactic acid ( lactic acidosis ) . In order to grow and propagate a solid tumor , tumor cells must adapt to and survive under these TME stresses . Additionally , tumor cells in regions of LA or hypoxia are more radio- and chemo-resistant and are more likely to metastasize [3] . Since TME stresses are found in most solid tumors , targeting stress-adaptation mechanisms of tumor cells may offer a significant therapeutic window to selectively eradicate tumor cells and improve patient outcomes [4] . Yet , current therapies targeting cells specifically under stress have significant limitations . For example , angiogenesis is a well-established , valuable therapeutic target with agents developed to block it at various stages of tumor development . However , many anti-angiogenic therapies fail over time , through acquired or inherited resistance that may involve the presence of tumor hypoxia [5] . Therefore , a number of other strategies are being developed to directly target hypoxic cells , such as blocking lactate transporters [6 , 7] , or pro-drugs that are activated only in the presence of low oxygen [4] . There remains significant room for improvement to target cells under stress , and thus there remains a need to better understand the genes that impact cellular survival under TME stresses . Tumor cells employ at least two kinds of adaptive strategies to cope with TME stresses , transcriptional and metabolic , and these are often interconnected . Transcriptional changes are mediated by stress-activated transcription factors . For example , the most significant factors known to regulate a cell’s hypoxia response are the hypoxia-inducible factors ( HIFs ) , which are stabilized under low oxygen and mediate complex transcriptional programs that increase glucose uptake and enhance glycolysis [8 , 9] . HIF–1α also regulates glutamine metabolism by affecting the ubiquitination of its oxidizing enzyme , AKGDH , to promote reductive carboxylation of glutamine under hypoxic conditions [10–12] . In addition to the HIFs , many other transcription factors regulate cellular responses to stresses in the TME . For example , the MondoA:Mlx complex senses and initiates transcriptional changes under both glucose deprivation and lactic acidosis to induce TXNIP and restrict glucose uptake [13–15] . Importantly , multiple co-factors and modulators add to the complexity of stress mediated transcriptional responses [16] . While these adaptive mechanisms can be successful to sustain cell growth under stress , in multiple contexts hypoxia induces apoptosis [17–19] . Transcriptional responses mediate many metabolic reprogramming events , but , recently , it is also becoming evident that metabolic events can regulate gene expression . For example , the metabolic enzyme ATP citrate lyase ( ACLY ) generates glucose-derived acetyl-CoA from citrate to alter histone acetylation and , therefore , transcriptional activation [20] . Importantly , the metabolically sensitive mTOR signaling cascade can activate HIF , even under normoxia [21 , 22] . Mutations in succinate dehydrogenase and fumarate hydratase both lead to increased levels of their substrates ( succinate and fumarate , respectively ) , causing increased HIF–1α stability and alterations of genome-wide histone and DNA methylation [23 , 24] . Metabolites such as NAD+ , NOS and α-ketoglutarate ( α-KG ) can also affect HIF function and histone and DNA modifications [21 , 22 , 25–28] . While cellular metabolism and transcriptional changes can provide flexibility for adaptation , cancer cells can also become reliant on , and thus vulnerable to the inhibition of , specific metabolic pathways or gene products [29–31] . Therefore , a better understanding of the genes necessary for modulating cancer cell survival under TME stresses will improve the development of targeted therapies that selectively eradicate cancer cells under stresses [4 , 32] . Functional genetic screens provide an unbiased and powerful means of identifying genes responsible for any phenotype that can be measured experimentally . Unbiased RNAi screens have identified genes that influence the survival of organisms and cells under various stresses [33–36] . These studies provide a foundation for using RNAi screens to uncover genes involved in TME-relevant stress responses , but so far have not been applied genome-wide to identify genes that modulate the survival of cancer cells under hypoxia or lactic acidosis . To better understand the genes involved in the adaptations of cancer cells under TME stresses , we performed genome-wide pooled shRNA screens of lung cancer cells under hypoxia and lactic acidosis . Completing these screens revealed that the inhibition of ACC1 or ACLY , two key enzymes of de novo lipogenesis , protected cancer cells from hypoxia-induced apoptosis . ACC1 or ACLY inhibition protected cells by elevating levels of α-ketoglutarate under hypoxia to reduce the activity of the oncogenic transcription factor ETV4 . Together , these data provide evidence to support a molecular connection between cellular metabolic and transcriptional hypoxia adaptation via the ACLY-ACC1-ETV4 axis through α-ketoglutarate .
To identify genes that modulate cell survival under lactic acidosis and hypoxia , we conducted genome-wide , shRNA-based , contextual pooled screens in the lung cancer cell line H1975 under hypoxia or lactic acidosis ( Fig 1a ) . To preferentially discover genes important for survival rather than proliferation , the screen was done in low proliferative conditions ( see Methods ) . Cells were transduced with a genome-wide MSCV-based shRNA library [37] , selected for successful transduction with puromycin and then grown under control ( 21% O2 , pH 7 . 4 ) , hypoxia ( 2% O2 , pH 7 . 4 ) , or lactic acidosis ( 21% O2 , 25mM lactic acid , pH 6 . 7 ) for 4 days , as performed previously [15 , 38] . At these stress treatments , there was a ~50% reduction in cell number , which allowed us to uncover both genes whose suppression reduced or improved survival under stresses . Genomic DNA was isolated from cells under each condition and the genome-incorporated shRNA sequences were amplified by PCR . The amplified PCR products were labeled and competitively hybridized to a custom microarray to identify those shRNA sequences that were either enriched or depleted relative to the control treatment ( Fig 1a ) . The custom array was modified slightly from similar arrays used in other shRNA screens [39 , 40] . When different ratios of differentially labeled PCR products were hybridized on the arrays , we noted distinguishable differences in the signals , demonstrating the specificity and sensitivity of the array ( S1a Fig ) . Biological triplicates of each condition had highly reproducible signals ( S1b Fig ) . The abundance of each shRNA sequence reflected the effect of its target gene on cell survival under stresses: if the shRNA was depleted in the stress treatment , the gene it targeted had a “synthetic sick/lethal” phenotype; if the shRNA was enriched in the stress treatment , the gene it targeted had a “synthetic survival/protective” phenotype under stress . In order to analyze the effect of each shRNA in stress , we calculated an “R/G” ratio ( see Methods ) . R/G ratios were distributed on a scale of +/- 4 . 0 that was highly consistent between replicates and stresses ( S1c Fig ) . To minimize false positives due to off-target effects of individual shRNAs , we focused only on the genes that had at least two distinct shRNA sequences that were enriched or depleted ( S1 and S2 Tables , see Methods ) . Importantly , this “multiple hairpin analysis” identified EPAS1 ( hypoxia-inducible factor 2α , HIF–2α ) as a synthetic lethal gene under hypoxia . We further validated this by showing that silencing EPAS1 by shRNA reduced cell survival under hypoxia ( Fig 1b and 1c ) . This result was consistent with the critical role of EPAS1 in cellular adaptation to hypoxia [8] . The “re-discovery” of EPAS1 provided confidence in our screen and analysis methods . However , no pathways or gene ontology groups were significantly enriched within the different categories of “multiple hairpin hits” . We then performed a RIGER analysis using a log-fold change and the second best shRNA for each gene criteria ( Fig 1d , S3 Table ) [41] . This RIGER analysis revealed an enrichment for genes affecting mRNA regulation and binding , as well as membrane dynamics and nuclear localization [42] . Additionally , there was little overlap between the genes targeted by multiple hairpins , either enriched or depleted , in the two stresses ( Fig 1e ) . This was consistent with past reports of distinct responses and adaptations to hypoxia and lactic acidosis [15 , 43] . Next , we identified high confidence “hits” in both the multiple hairpin and RIGER analyses for further investigation . From these considerations , we chose ACC1 ( acetyl-CoA carboxylase 1 or ACACA ) as the top candidate . ACC1 encodes the cytosolic isoform of acetyl-CoA carboxylase , which converts acetyl-CoA to malonyl-CoA in the rate-limiting step of de novo fatty acid synthesis . There was an enrichment of shRNAs targeting ACC1 in the hypoxic versus the control condition , suggesting that ACC1 knockdown allowed for improved survival under hypoxia . ACC1 had 4 hairpins enriched under hypoxia ( Fig 1f ) and scored as the 13th best gene in the RIGER analysis using the second best shRNA metric ( Fig 1d ) . Additionally , the down-regulation of ACC1 was previously shown to protect cancer cells from glucose deprivation and matrix detachment stresses [44] . Together , these data prompted us to validate and investigate the role of ACC1 under hypoxia . To validate the shRNA screen result , we silenced ACC1 expression through lentiviral infection of multiple shRNAs that targeted different sequences from those shRNAs used in the screen ( see Methods ) . We confirmed the successful reduction of ACC1 protein by these shRNAs ( Fig 2a ) . In the control cells transduced with a scramble shRNA , hypoxia significantly decreased cell viability and induced apoptosis ( Fig 2 ) . However , silencing ACC1 by multiple shRNAs inhibited the hypoxia-induced apoptosis as shown by crystal violet staining ( Fig 2b ) , cell counting ( Fig 2c ) , propidium-iodide staining ( flow cytometry ) ( Fig 2d ) and PARP cleavage ( Fig 2e ) . This hypoxic protection associated with ACC1 silencing was also reproduced in additional hypoxia-sensitive cell lines , including MDA-MB–231 ( breast cancer; S2a and S2b Fig ) , and PANC–1 ( pancreatic cancer; S2e and S2g Fig ) . Furthermore , chemical inhibition of ACC1 through the AMPK agonist metformin also protected H1975 cells from hypoxia-induced apoptosis ( S2h and S2i Fig ) . Collectively , these data successfully validated the screen results and showed that the depletion of ACC1 enhanced cell survival under hypoxia in multiple cancer cells from different tissues of origin . We next investigated the specificity of the hypoxia protection by ACC1 depletion . Besides ACACA ( ACC1 ) , ACACB ( ACC2 ) encodes another isoform of acetyl-CoA carboxylase , located in the outer mitochondrial membrane [45] . ACC2 was not a hit in our screen and its silencing by shRNA did not offer a similar hypoxia protection as seen with ACC1 depletion ( S2j Fig ) . Next , we determined whether depletion of ACC1 protected against other TME stresses ( lactic acidosis , glutamine deprivation and glucose deprivation ) . We found that the protective effect of shACC1 was seen only under hypoxia ( S2k and S2l Fig ) . Since the HIFs are the major transcriptional responders to hypoxia , we examined how loss of ACC1 affected HIF–1α levels . Interestingly , we found decreased levels of HIF–1α under hypoxia with ACC1 depletion across multiple cell lines ( S3a–S3c Fig ) . These data suggested that the protective phenotype was not due to upregulation of the HIF response . Overall , these results indicated that only the cytosolic isoform of acetyl-CoA carboxylase ( encoded by ACC1 ) was essential for apoptosis , specifically under a hypoxic stress . When we examined the effect of blocking enzymes up- or downstream of ACC1 , we found that silencing ATP citrate lyase ( ACLY ) also enhanced survival under hypoxia ( Fig 2f and 2g ) . ACLY encodes the enzyme immediately upstream of ACC1 in lipogenesis , catalyzing the formation of acetyl-CoA and oxaloacetate from citrate . Similar to ACC1 , this protection results from the inhibition of hypoxia-induced apoptosis ( Fig 2h ) . The protective effect of ACLY depletion was also reproduced in MDA-MB–231 cells ( S2c and S2d Fig ) and PANC–1 cells ( S2f and S2g Fig ) . In H1975 and MDA-MB–231 shACLY cells there was also decreased HIF–1α expression under hypoxia as was seen in the shACC1 cells ( S3d and S3e Fig ) . These data showed that blocking lipogenesis at the points of either ACLY or ACC1 inhibited apoptosis and permitted cell survival under hypoxia in cells of multiple tissue types . Lipogenesis is a highly anabolic process that requires significant amounts of NADPH and ATP . Previously , silencing ACC1 protected cells from death caused by glucose deprivation and matrix detachment by preserving NADPH and ATP to counteract the ensuing oxidative stresses [44 , 46] . We tested the relevance of these factors in our system . In H1975 cells , silencing ACC1 trended toward increasing the NADP+/NADPH ratio , suggesting a decrease in available NADPH ( S4a Fig ) . We reasoned that if the NADPH were being used to combat elevated reactive oxygen species under hypoxia , then supplementation with antioxidants should protect control cells from hypoxia-induced death similar to [44] . However , neither the addition of N-acetyl cysteine nor glutathione antioxidants rescued hypoxia-induced death in control cells ( S4b and S4c Fig ) . While ATP levels were higher with ACC1 silenced , the change in ATP levels from normoxia to hypoxia was consistent in control and knockdown cells and thus could not readily explain the hypoxia protection ( S4d Fig ) . Therefore , in these cells with ACC1 or ACLY depleted , changes in NADPH and ATP levels may not be the primary mechanism for cell survival under hypoxia . Therefore , we sought to identify another mechanistic explanation for this hypoxia protection phenotype . In our “multiple hairpin analysis” of the hypoxia genome-wide screen , there was an enrichment of shRNAs targeting a PEA3 transcription factor family member , ETV4 . These results suggested that silencing ETV4 may be protective under hypoxia . A link between lipogenesis and ETV4 was previously established when levels of malonyl-CoA were associated with ETV4 activity [47] . This prompted us to investigate a potential regulatory relationship between ACLY , ACC1 and ETV4 . Real-time PCR analysis showed that hypoxia led to a reduction of ETV4 mRNA in the ACC1-depleted , but not control cells ( Fig 3a ) . Additionally , we noted correspondingly reduced ETV4 protein in the shACC1 cells as compared to the scramble cells ( Fig 3b ) . Reduced ETV4 mRNA and protein levels were also noted in the ACLY-depleted cells ( S4e and S4f Fig ) . While ETV4 protein levels were somewhat decreased under normoxia , the down-regulation was stronger under hypoxia . This regulation was mostly specific to ETV4; the other PEA3 subfamily members , ETV1 and ETV5 , were not consistently altered by ACC1 depletion ( S4g Fig ) . Additionally , neither ETV1 nor ETV5 were identified as “multiple hairpin hits” in the shRNA screen . We validated the hypoxia-protective phenotype of ETV4 loss with two different shRNAs targeting ETV4 ( Fig 3c and 3d ) . Similar to ACC1/ACLY depletion , the depletion of ETV4 also reduced the percentage of cells in the sub-G1 phase ( Fig 3e ) and decreased PARP cleavage ( Fig 3f ) under hypoxia . These data showed that the loss of ETV4 decreased hypoxia-induced apoptosis , similar to the phenotype of reduced ACC1 or ACLY . Since ETV4 is a transcription factor , we investigated the contribution of reduced ETV4 activity to the transcriptional response of ACLY or ACC1 depletion . We used microarrays to analyze the global transcriptional response to the silencing of each ACC1 , ACLY or ETV4 by two independent shRNAs under hypoxia ( each shRNA was done in triplicate ) . The transcriptional responses were determined by zero-transformation against the shScramble cells [48] . Next , the data were filtered with a 1 . 7-fold change in at least six arrays and the selected 641 probesets were grouped by hierarchical clustering ( Fig 4a ) . This analysis revealed a remarkable similarity between the transcriptional responses to the depletion of ETV4 , ACC1 or ACLY with the induction and repression of common sets of genes ( Fig 4a ) . Using the GATHER [49] algorithm , we noticed an “anti-apoptotic expression program” that included both the induction of negative regulators of apoptosis signaling such as NQO1 , CYP1B1 and SERPINE1 [50 , 51] and the repression of the apoptosis-promoting genes BIK , TNFRSF9 , TNFAIP3 , GLIPR1 , DDIT and TRIB3 ( Fig 4a ) . The induction and repression of multiple genes in the shACC1 , shACLY and shETV4 cells were confirmed by real-time qPCR ( Fig 4b and 4c ) . Using GSEA , these gene expression changes were highly overlapping in all pairwise comparisons with both up- and down-regulated genes ( S4h and S4i Fig ) . Next , we evaluated whether the ACC1-affected genes were transcriptional targets of ETV4 by two different approaches . First , we compared publicly available ETV4 ChIP-seq data ( Cistrome Finder [52] ) from PC3 cells [53] with the genes that were changed in our microarray analysis of H1975 cells with loss of ETV4 , ACC1 or ACLY ( Fig 4a ) . While performed in a different cell ( PC3 ) , these analyses still identified at least two potential direct ETV4 target genes , PLEC ( S5a Fig ) and DUSP6 ( S5b Fig ) . For both genes , there were peaks indicating direct ETV4 binding that overlapped with histone H3 lysine 27 acetylation ( a mark of actively transcribed gene bodies ) and DNase hypersensitive regions of open chromatin ( S5a and S5b Fig ) . DUSP6 has been previously described as an ETS transcription factor family target [54 , 55] . While PLEC was reported to interact with vimentin , an ETV4 direct target [56] , this analysis suggested that PLEC itself may represent a novel ETV4 target . In the second approach , we used qPCR to determine if the ACC1-affected genes could be “rescued” by ETV4 over-expression . We found that CTSS , COL13A1 , DUSP6 and SERPINE1 could be reversed by ETV4 re-expression ( S6a and S6b Fig ) . In contrast , the ACC1-altered expression of other genes was either partially or not restored upon ETV4 overexpression ( S6c Fig ) . This analysis suggested that some of the gene expression changes discovered by our microarray analysis may represent direct effects of changed ETV4 transcription , while others likely represent more indirect changes with ETV4 loss . Together , these data indicated that the repression of ETV4 played an important role in a subset of the transcriptional response to ACC1 depletion . In order to better understand if these changes reflected an in vivo biological regulation between these genes , we developed “gene signatures” associated with the silencing of ETV4 , ACC1 or ACLY using the CreateSignature algorithm [57] . These gene expression signatures represent “quantitative phenotypes” that reflect the loss of these genes . Comparing their similarity in different expression datasets allowed us to recognize similar quantitative changes in these genes in both in vitro experimental perturbation and human tumors . Similar “gene signature” approaches have been used to define the influences of oncogenic signaling and TME stresses in multiple cancer types [58–60] . Gene expression patterns from human tumor samples [61] were then separated by their similarity to our developed gene signatures associated with loss of ETV4 ( shETV4 ) , ACC1 ( shACC1 ) and ACLY ( shACLY ) . Binary regression from this analysis in human tumors showed highly statistically significant correlations between the shACC1 or shACLY and shETV4 signatures ( Fig 4d ) . In other words , patient tumors with gene expression patterns more similar to the ACC1-depletion ( shACC1 ) signature had expression patterns that were also more similar to the ETV4-depletion ( shETV4 ) signature; likewise , patient tumors with gene expression patterns similar to the ACLY-depletion ( shACLY ) signature also had similar gene expression patterns with the ETV4-depletion ( shETV4 ) signature . Importantly , this showed that the regulation between ACC1/ACLY and ETV4 was relevant in tumor expression datasets . Overall , these analyses demonstrated the similarity of the transcriptional responses to the depletion of ACC1 , ACLY or ETV4 and suggested that ETV4 mediated a portion of the transcriptional effect downstream of ACLY or ACC1 both in vitro and in vivo . Considering that ACC1 and ACLY are critical lipogenic enzymes , we performed a metabolomics experiment to interrogate the metabolic effects of ACC1 or ACLY depletion under normoxia or hypoxia . Five cell lines were evaluated in triplicate: 1 control “hypoxia-sensitive” cell line ( shScramble line ) and four “hypoxia-survival” cell lines ( 2 shACC1 lines , 2 shACLY lines ) . After 36 hours of treatment , cells were lysed on ice and collected to measure the intracellular levels of 15 amino acids and 45 acyl-carnitines by tandem mass spectrometry ( MS/MS ) and levels of 7 organic acids by gas chromatography and mass spectrometry ( GC/MS ) . All measurements were normalized by total protein content per sample . We observed several expected metabolic changes to validate our approach . Silencing ACC1 depleted basal and hypoxia-induced palmitate levels , reflecting reduced de novo lipogenesis in these cells ( S7a Fig ) . Consistent with hypoxia-induced inhibition of pyruvate utilization in the TCA cycle in favor of anaerobic glycolysis , hypoxia modestly increased the levels of pyruvate and lactate in control cells ( S7b and S7c Fig ) . In addition , as previously noticed [62] , hypoxia generally reduced the levels of TCA metabolites succinate , fumarate , malate , and citrate in most cells ( Fig 5a ) . These results indicated that our metabolomics assay accurately detected the expected metabolic changes associated with inhibited lipogenesis and hypoxia exposure . The pattern of α-ketoglutarate ( α-KG ) levels in this experiment suggested that it may be an interesting candidate for offering protection under hypoxia . In the control cells , hypoxia reduced the levels of α-KG . However , in the hypoxia-resistant cells with depleted ACC1 and ACLY , hypoxia increased the α-KG levels ( Fig 5b ) , as has been seen before in hypoxia-resistant cells [62] . We could test the possibility that levels of α-KG contributed to survival by adding cell-permeable dimethyl-α-KG to H1975 cells . We determined the level of α-KG achieved intracellularly after extracellular supplementation to choose a supplementation treatment that would be relevant to the levels of α-KG seen with ACC1 or ACLY depletion . Mass spectrometry analysis showed increasing amounts of intracellular α-KG after supplementation in a dose-dependent manner ( Fig 5c ) and that dimethyl-α-KG supplementation at 1mM achieved levels of α-KG comparable to the hypoxia-induced increase found in the ACC1 and ACLY depleted cells under hypoxia ( Fig 5b ) . Supplementation of relevant levels α-KG also inhibited PARP cleavage under hypoxia ( Fig 5d ) . These data indicated that the increased α-KG under hypoxia in the ACC1 and ACLY depleted cells recapitulated the hypoxia-protective phenotype of these cells . With both α-KG and ETV4 acting downstream of ACC1 or ACLY , we next determined if α-KG was mediating the effects of ACC1 or ACLY on ETV4 expression . α-KG supplementation reduced mRNA and protein levels of ETV4 at both lower ( Fig 6a and 6b ) and higher concentrations ( S7d and S7e Fig ) . Additionally , α-KG supplementation in control cells caused similar changes in repressed and induced genes as was caused by ACC1 , ACLY or ETV4 silencing ( Fig 6c and 6d ) , and some of these mRNA effects were dose-dependent with α-KG supplementation ( S7f and S7g Fig ) . Unexpectedly , diM-α-KG supplementation increased HIF–1α protein levels ( S7h Fig ) . Since the regulation of HIF–1α was different with depletion of ACC1/ACLY or α-KG supplementation , the changes in HIF–1α protein did not likely explain the improved hypoxic cell survival of the ACC1/ACLY depleted cells . Combined , these data indicated that , in the ACLY or ACC1 depleted cells , the α-KG increase was a hypoxic trigger that reduced ETV4 levels and activity to mediate an anti-apoptotic gene expression response . In addition to being a TCA cycle intermediate , α-KG is a substrate for the abundant 2-oxoglutarate/Fe ( II ) -dependent dioxygenases ( 2-OGDDs ) [26] . 2-OGDDs use α-KG and molecular oxygen as substrates to perform a number of different protein modification reactions . These enzymes include families of histone demethylases that recognize and remove methylation marks from histones , as well as the TET family of proteins that facilitate DNA demethylation [26] . Thus , α-KG levels can affect gene expression through the activities of these 2-OGDDs [28] . An elevated ratio of α-KG/succinate ( substrate/product ratio ) has been proposed as a potential indicator of increased 2-OGDD activity [25 , 28] . We found that the α-KG/succinate ratio was significantly elevated in all of the shACC1 and shACLY cells under hypoxia ( Fig 6e ) . To better understand if the level of α-KG or the α-KG/succinate ratio determined our hypoxia-survival phenotypes , we supplemented ACLY or ACC1 depleted cells with cell-permeable dimethyl-succinate to theoretically drive the α-KG/succinate ratio in the opposite direction from when α-KG was added . Succinate supplementation did not affect the survival of either shACC1 or shACLY cells under hypoxia ( S8a and S8b Fig ) and also did not affect the regulation of ETV4 by ACC1 or ACLY ( S8c and S8d Fig ) . Therefore , in our experimental system , we concluded that the levels of α-KG , rather than the α-KG/succinate ratio , were driving the hypoxia survival phenotypes . While the succinate supplementation did not affect our phenotype , the elevated α-KG in the ACC1 or ACLY depleted cells under hypoxia still suggested that 2-OGDDs may be relevant in our system . Therefore , we hypothesized that α-KG affected ETV4 mRNA abundance by altering the activity of 2-OGDDs and subsequent histone methylations . As a control , we tested the DNA methylation status of the two shore regions and the center region of the ETV4 promoter CpG islands by bisulfite pyro-sequencing and saw no significant change upon α-KG supplementation ( S8e Fig ) . However , α-KG caused a global reduction in two ( H3K4me2 and H3K4me3 ) “active” histone methylation marks and also globally reduced two “repressive” marks , H3K27me3 and , to a lesser extent , H3K9me3 ( Fig 6f ) . When we compared the epigenetic changes associated with α-KG supplementation with Carey et al . [28] , we found that H3K27me3 was consistently reduced with the addition of α-KG in both studies . However , there were also differences in the histone methylation changes caused by α-KG across the studies: 1 ) H4K20me3 was reduced with α-KG previously and we saw an increase in this mark ( S8f Fig ) ; 2 ) whereas we saw a decrease in the levels of H3K4me3 , no changes were seen previously . These differences could be due to different cellular contexts ( embryonic stem cells vs . cancer cells ) or the presence of glutamine deprivation during the previous examination of α-KG effects [28] . Similar changes were observed using either water or DMSO as control ( S8g Fig ) . Collectively , the reduced levels of multiple methylation marks were consistent with our hypothesis that predicted more active histone demethylases as a result of increased levels of α-KG . To extend evidence in support of our model , we also examined various histone methylation modifications associated with the depletion of ACC1 or ACLY . Overall , we saw similar global histone methylation changes in both the shACC1 ( Fig 7a ) and shACLY ( Fig 7b ) cells as compared to α-KG supplementation . Among all the tested epigenetic markers , the H3K4me3 mark was the most pronouncedly decreased across both gene depletions and the α-KG treatment . The H3K4me2 mark was decreased in all three conditions ( shACC1 , shACLY , α-KG supplementation ) to a modest degree . Similarly , H3K9me3 was decreased somewhat with the 8 hour α-KG treatment and in the shACC1 cells , while it was more strongly decreased in the shACLY cells . H3K27me3 was lowered by both α-KG and ACC1 depletion . Levels of H4K20me3 were unchanged in the shACC1 and shACLY cells while they were increased with α-KG treatment , and so this suggested that the changes in this methyl mark were likely not due to the changes we explain in our model . Besides global epigenetic changes , we also determined if histone methylation at the ETV4 locus was changed by chromatin immunoprecipitation of the “active” histone H3 lysine 4 tri-methylation ( H3K4me3 ) mark . This mark was chosen because it showed the most consistent and strongest changes across either genetic depletion or with α-KG supplementation . In addition , a loss of this “active” mark H3K4me3 would be consistent with decreased ETV4 expression in these conditions . ChIP experiments showed that α-KG treatment decreased the abundance of the H3K4me3 modification at the ETV4 locus by ~30% ( Fig 7c ) . Additionally , there was a decreased abundance of H3K4me3 in both ACC1 and ACLY depleted cells under hypoxia as compared to control cells ( Fig 7d and 7e ) . Together , these data showed that elevated levels of α-KG affected the histone , but not DNA , methylation status of the ETV4 locus and this pattern was similar to the histone methylation changes seen under hypoxia in the shACC1 and shACLY cells . Collectively , our data was consistent with a model in which , under hypoxia , the inhibition of ACC1 or ACLY increases levels of α-ketoglutarate to block hypoxia-induced apoptosis by reducing the levels and activity of ETV4 , possibly through altered histone methylation patterns ( Fig 8 ) . These data offer a novel molecular connection showing that the transcriptional output of altered lipogenic metabolism can modulate the cellular response to and survival under hypoxia .
Here we have described a pooled shRNA screen that successfully identified genes that influence cancer cell survival under hypoxia and lactic acidosis . Specifically , we showed that blocking de novo lipogenesis through the genetic depletion of ACLY or ACC1 protected multiple cancer cells from hypoxia-induced apoptosis through increased levels of α-ketoglutarate and the inhibition of ETV4 and its transcriptional activities . Therefore , inhibition of ACLY or ACC1 affected both metabolism and transcription to protect cells from hypoxia-induced apoptosis . These results also suggest that one important mechanism of hypoxia-induced apoptosis is through the reduction of α-KG , which potentially elevates the levels and activity of ETV4 through histone modifications to promote oncogenesis and trigger apoptosis . Inhibitors that target ACLY and ACC1 are proposed cancer therapeutics [63 , 64] . ACC1 has been targeted as it is the rate-limiting enzyme in lipogenesis , a process critical for cancer cells’ rapid proliferation [63] . For ACLY , not only is it a critical enzyme for active lipogenesis , but it also regulates epigenetic states by generating acetyl-CoA , and so it offers two mechanisms to target clinically [20] . Our data provides several insights on the biological effects of ACC1 and ACLY in hypoxic cancer cells that should be considered when targeting these enzymes . First , while we expect ACLY and ACC1 inhibition to have opposite effects on levels of acetyl-CoA , their effects on gene expression were highly similar . This suggests that levels of acetyl-CoA ( and subsequent histone acetylation or changes in epigenetic states ) may not readily explain the majority of gene expression responses to the inhibition of ACC1 or ACLY under hypoxia . Instead , the ACC1/ACLY-induced reduction in ETV4 levels and activity seemed to account for a significant portion of the hypoxic transcriptional changes . These results suggest that lipogenic inhibitors that block ACC1 or ACLY may be particularly effective for tumors driven by ETV4 . Likewise , this reduction of the oncogenic driver ETV4 may account for portions of the therapeutic potential of lipogenic inhibitors . However , our data also showed that these treatments may allow for the survival of slowly proliferating cancer cells in regions of hypoxic tumors . Therefore , targeting lipogenesis in cancer may need to be combined with other therapeutic approaches that target hypoxic regions ( such as hypoxia-activated pro-drugs ) , to eliminate the cancer cells that may be dormant and protected in the hypoxia-niche . Elevated α-KG levels in the shACLY and shACC1 cells under hypoxia could be due to increased generation or decreased consumption of this metabolite . A portion of α-KG is consumed during lipogenesis , which is further promoted under hypoxia by reductive carboxylation [12 , 65] . Therefore , blocking lipogenesis under hypoxia may lead to a “build-up” of upstream metabolites , including α-KG . It is well appreciated that α-KG plays a critical role in supporting cell survival by replenishing the metabolic intermediates of the TCA cycle . Here , our data indicated than an additional manner by which α-KG can affect cellular survival under hypoxia was by regulating ETV4 expression , possibly through epigenetic mechanisms . Transcriptional adaptation to hypoxia is most often orchestrated by the HIFs; however , here we showed that cancer cells’ hypoxic survival can be mediated by a different transcription factor , ETV4 . ETV4 was proposed as an essential co-activator of HIF–1α and to have a hypoxic transcriptional program [66] . Our data revealed that ETV4 was critical for hypoxia-induced apoptosis . Interestingly , the levels of ACC1 or ACLY influenced α-KG and ETV4 levels under hypoxia , thereby providing a link between lipogenesis , a TCA cycle intermediate and transcription . Therefore , ETV4 mediated the transcriptional output of varying degrees of active lipogenesis caused by changing ACC1 and ACLY levels . Our gene signature analyses also suggested that this regulation was preserved in human tumors in vivo . While the global and local ETV4 epigenetic changes we describe herein were consistent with decreased activity of ETV4 to protect cells from hypoxia-induced apoptosis , our evidence provides only correlative support with reduced promoter activity and does not provide a causal explanation for the ETV4 repression . Additionally , although this regulation seemed mostly specific to ETV4 , we do not fully exclude the possibility that other ETS transcription factor family members contribute to an apoptotic , hypoxic transcriptional program . Overall , our data establishes ETV4 as one critical factor that influences hypoxic cell survival and transcriptional responses downstream from ACC1 and ACLY , and reveals the metabolite α-ketoglutarate as a molecular link between metabolic and transcriptional adaptation to hypoxia . In addition to its importance in the TCA cycle , our experiments showed that elevated α-KG could alter histone methylation patterns , likely via α-ketoglutarate-dependent dioxygenases , to potentially regulate ETV4 . While the causal relationship of several regulatory steps of this hypothesis was not rigorously tested by genetic manipulations here , our data was consistent with such a model . Previous reports suggested that α-KG/succinate ratios determined the direction and activities of dioxygenases [28] , yet our data indicated that it was increased α-KG , not succinate or α-KG/succinate , that drove the epigenetic changes , ETV4 repression and hypoxia survival phenotypes . As a substrate , α-KG levels likely affect the enzymatic activities of of the 2-OGDDs histone demethylases , each of which have different sets of lysine residues upon which they act [26] . The number and combination of global histone methylation events that were changed with depletion of ACC1 , ACLY or α-KG treatment suggested that many different demethylases could be affected by α-KG or the depletion of ACC1/ACLY in these cellular states . While there are no histone methyltransferases currently known to use α-KG as a substrate , they are regulated by hypoxia [67 , 68] , and so we do not exclude the potential for changes in histone methylation patterns to be due to changes in both methyltransferase and demethylase enzyme activities . Additionally , the changes in global histone methylation patterns are very likely affecting a number of other genes’ expression patterns , in addition to the changes we see at the ETV4 locus . Since α-ketoglutarate supplementation and loss of ACC1 or ACLY reduced the abundance of H3K4me3 at the ETV4 promoter , we speculate that the JARID1 ( KDM5 ) family of histone demethylases , which specifically demethylate H3K4me2/3 residues , could be affecting the abundance of this mark at the ETV4 locus [69 , 70] . These enzymes are known to range in their expression and activity by cell type and are differentially influenced by oxygen levels [71–73] . A full investigation in to the mRNA , protein and activity levels of many of these family members would be necessary to determine the extent to which each plays a role in regulating ETV4 under hypoxia . Effort has been made to investigate the JARID1 family of H3K4 demethylases as potential cancer drug targets [70 , 74 , 75] as the importance of a mis-regulated “histone code” for tumorigenesis is well recognized . While some reports suggest a tumor suppressive role of these enzymes , more suggest an oncogenic function [69] . This information and the data we presented here suggest that various 2-OGDDs in distinctive contexts differentially affect tumorigenesis and tumor cell survival . It will be important to understand the proper context of treatment if these drugs continue into the clinic . Multiple previous reports show that the activity of ACLY [64 , 76 , 77] , ACC1 [78–81] or ETV4 [82–85] is associated with increased tumorigenicity and/or poor patient outcome , or that inhibiting these genes’ activities reduces tumorigenicity and improves patient outcome . We show that the inhibition of ACLY , ACC1 or ETV4 paradoxically allows tumor cells to survive better under hypoxia . To address this apparent conundrum , we propose a conceptual model in which there are two cellular states: one of activating oncogenesis and the other of stress survival . This model includes a trade-off between the two states , such that the promotion of one comes at the expense of the other . Stated another way , the activation of oncogenic programs may also render cells susceptible to apoptosis , especially under stress; likewise , reduced oncogenic programs slow cellular proliferation to a “dormant” state that could allow for better stress survival . A similar model has been proposed for several oncogenes . For example , oncogenesis driven by MYC rendered non-transformed cells vulnerable to hypoxia-induced apoptosis [86 , 87]; the degradation or cleavage of c-MYC under hypoxia allowed tumor cells to evade hypoxia-induced apoptosis [88 , 89] . E2F can promote proliferation ( oncogenesis ) or apoptosis in different contexts , such as with differing PI3K activity [90] . A recent paper also indicated that HIF–1α repressed the ATF4 stress response pathway to allow for the expansion of fetal cardiomyocytes [91] . According to this proposed model and our data , cancer cells treated with ACLY or ACC1 inhibitors ( including metformin ) , may die due to blocked lipogenesis , but may also survive in hypoxic regions . As these cells resist death under stress , they may become the “dormant” cells that recur after treatment regimens end , but they also may become more targetable as they persist under hypoxia . Consistent with past literature and clinical attempts to target these genes , our model advises that treatment regimens be carefully considered . While data to prove such a model remains necessary , this study suggests that the ACLY-ACC1-ETV4 axis might mediate the balance between oncogenesis and stress survival under hypoxia . It is very likely that similar mechanisms of balance between proliferation and stress survival exist in a wide variety of biological contexts .
H1975 cells were spin-infected with the pMSCV-based retroviral genome-wide library , at an MOI of 0 . 3 , divided into six sub-pools , achieving a final library representation of 1000 cells per shRNA after selection with 1 ug/ml puromycin [37] . After three days of puromycin selection , cells were split into control and stress conditions , maintaining 1000-fold representation of each shRNA per triplicate . Cells were serum starved to 0 . 1% FBS 24 hours after plating . 24 hours after serum starvation , media was changed to treatment media ( 0 . 1% FBS , 25mM Hepes ) ; control and hypoxia media pH = 7 . 4; lactic acidosis treatment had 25mM lactic acid ( Sigma cat . no L6402 ) adjusted to pH = 6 . 7 and filter sterilized . After 4 days of treatment , cells were harvested , centrifuged , and frozen at -80°C . Genomic DNA was extracted with the QIAamp DNA Blood Maxi kit ( QIAGEN , cat . no 51194 ) then shRNA sequences were PCR amplified . The amplified products from the control and each stress were labeled ( Cy3 and Cy5 , respectively ) and then interrogated by a custom Agilent microarray , which contained probes against the library’s shRNA sequences [37] . We validated the sensitivity and specificity of the array to different ratios of labeled PCR product ( S1a Fig ) . Therefore , relative hybridization of the Cy5/Cy3 labeled shRNA populations determined the abundance of each shRNA under control , hypoxia or LA . The Cy3 and Cy5 signals across the three biological replicates were highly reproducible ( S1b Fig ) . Probes with signal intensities of less than 2-fold above background were discarded . Cy5/Cy3 ratios , also called “R/G” ratios , for remaining probes were calculated , log2 transformed and quantile normalized across pools . The R/G ratios ranged from +/- 4 . 0 , although many fell in the “unchanged” range of +/-0 . 5 ( S1c Fig ) . For the “multiple hairpin analysis , ” genes were considered a hit when they had 1 ) at least 2 different shRNAs with ( absolute value R/G ) > 0 . 7 in at least 2 of the three biological replicates ( 2 ) the ( stdev/ave ) of the biological replicates was <0 . 5 ( S1 Table ) . H1975 cells were cultured in RPMI media ( GIBCO cat . no 11875 ) supplemented with 10% Fetal bovine serum ( heat-inactivated ) , 1% glucose , 10mM HEPES , 1mM sodium pyruvate , and 1x antibiotics ( penicillin , 10 , 000 UI/ml; streptomycin , 10 , 000 UI/ml ) , as directed by the Duke Cell Culture Facility . MDA-MB–231 and PANC–1 cells were cultured in DMEM ( GIBCO cat . no . 11995 ) supplemented with 10% Fetal bovine serum ( heat-inactivated ) and 1x antibiotics ( penicillin , 10 , 000 UI/ml; streptomycin , 10 , 000 UI/ml ) . Cell lines , obtained from and initially validated by the Duke Cell Culture Facility ( Durham , NC , USA ) , were maintained for fewer than 6 months and validated by microscopy every 1 to 2 days . Lactic acidosis was generated via addition of lactic acid ( Sigma-Aldrich , St . Louis , MO , USA , cat . no L6402 ) and media pH adjustment to pH 6 . 7 by HCl immediately before use . Hypoxia was generated with a cell culture incubator with 93–94% N2 , 5% CO and 1–2% O2 . For the α-KG rescue experiments , media was supplemented with 0 . 875-4mM dimethyl α-KG as indicated in figure legends ( Sigma , cat . no . 349631 ) . For all stress experiments , cells were serum starved ( 0 . 5% FBS ) for 24 hours before treated with stress under 0 . 5% FBS . All survival/viability measurements were made after 4 days of stress treatment . Stable cell lines were created with the pLKO . 1 shRNA constructs purchased from the Duke RNAi Core Facility . Virus was generated by transfecting HEK-293T cells with a 1: 0 . 1: 1 ratio of pMDG2: pVSVG: pLKO . 1 with Lipofectamine 2000 in the evening . Media was changed the following morning and virus collected 48 hours after transfection . Stable cell lines were generated by adding 200ul virus to a 60mm dish of parental cells with polybrene ( final concentration 8ug/ml ) . Complete death in blank infection dishes was used to determine success of infection and puromycin selection . The efficiency of silencing or overexpression was determined by western blots . Concentrations of puromycin needed for selection: H1975 cells = 1ug/ml , MDA-MB–231 cells = 1ug/ml , PANC–1 cells = 2ug/ml . For stable overexpression , concentration of blasticidin used was 2 . 5ug/ml in H1975 cells . Cells were fixed either in 4% paraformaldehyde ( PFA ) overnight at 4°C or at room temperature for 30 min . PFA was removed and crystal violet staining solution ( 0 . 2% crystal violet , 25% methanol , 75% water ) gently shaken on cells for 30+ minutes at room temperature . Staining solution was removed and plates rinsed with tap water 2–3 times . For quantitation , completely dried stain was dissolved by adding 10% acetic acid and shaking gently at room temperature for 30+ min before reading absorbance at 570 nm . Cell number was evaluated by either direct cell counting ( trypan blue exclusion ) or high-throughput microscopic counting ( HTC ) of fixed and stained nuclei . For direct cell counting , at designated time after treatment , media was removed , cells were not rinsed for fear of losing loosely-attached cells , trypsinized , diluted 1:1 with trypan blue and immediate counted on a hemocytometer . For HTC experiments , after designated time period , cells were fixed in 4% PFA either overnight at 4°C or for 30 min at RT . Cells were washed 2x , permeabilized with 0 . 1% Triton-X in PBS , wash 2x , stained with 50ug/ml Hoescht dye ( Sigma cat . no B2261 ) for 30 min , RT in the dark , then washed 2x and PBST added to each well and scanned by the Cellomics high-throughput microscope at the Duke RNAi Core Facility . For cell cycle analysis , after 4 days of stress treatment , media was collected , cells trypsinized and pooled with the media . Cells were centrifuged then fixed by resuspension in ice cold 70% ethanol while gently vortexing . Fixed cells were placed at -20°C until prepared for FACS analysis . Immediately before FACS analysis , cells were centrifuged for 5 min at room temperature , washed twice in PBS then resuspended in 25ug/ml Propidium iodide ( Sigma cat . no P4864 ) and 10ug/ml RNAse A in PBS . Cells were stained for 30+ min in the dark then 8000 events measured on a Canto II Flow cytometer . Cell lysis: Cells were washed once with ice cold PBS , lysed by RIPA buffer with protease and phosphatase inhibitors added fresh , scraped into a microcentrifuge tube , allowed to swell on ice for 15–20 min , vortexed briefly , then spun down at top speed for 15 min at 4°C . Supernatant was transferred to pre-cooled new tube and protein concentration assayed with the Pierce BCA kit ( ThermoScientific , cat . no . 23225 ) . Effort was made to immediately rinse and lyse cells coming from a hypoxia condition as cells very quickly re-equilibrate to normoxic conditions . Western blots: Between 15-30ug of lysate was loaded on SDS-PAGE gels , wet-transferred to PDVF membrane , blocked with 5% milk in 1xTBST ( 0 . 1% Tween–20 ) , then primary antibodies were incubated overnight at 4°C . For analysis of histones , protein was extracted with the EpiQuick Total Histone Extraction Kit ( Epigentek , cat . no . OP–0006 ) and 2ug of protein were resolved on 15% SDS-PAGE gels or nuclear fractions were collected by the REAP fractionation method [92] and 7 . 5–30 ul of lysate were run in each lane . Please contact for details on antibody usage . RNA was extracted using the RNeasy Kit ( QIAGEN ) . A total of 1 μg of total RNA was reverse transcribed by SuperScript II ( Invitrogen ) for real-time PCR with Power SYBRGreen Mix ( Applied Biosystems/Life Technologies ( Grand Island , NY , USA ) ) . Primers were designed across exons whenever possible , verified for specificity by regular PCR prior to use in real-time PCR . Please contact for the sequences of primers used . Samples were collected on ice and RNA was isolated with QIAGEN’s RNeasy Mini Kit ( cat . no 74104 ) according to manufacturer’s instructions . After quality control assessment with the Agilent BioAnalyzer , cDNA was amplified from 200ng RNA with the Ambion MessageAmp Premier RNA Amplification ( Life Technologies , Grand Island NY , USA ) . The gene expression pattern of the RNA samples were interrogated with Affymetrix U133A genechips and normalized by the RMA ( Robust Multi-Array ) algorithm . cDNA synthesis and microarray interrogation was performed by the Duke Microarray Core . The influence of the silencing of ACC1 or ACLY on gene expression was derived by a zero transformation process , in which we compared transcript level for each gene in cells with stably integrated shRNAs targeting ACC1 or ACLY to the average transcript levels in control scramble shRNA cell line samples . Data was then filtered as described with Cluster 3 . 0 software and heat maps were generated with TreeView . To generate gene signatures of knockdown of ACLY , ACC1 or ETV4 , the CreateSignature module in GenePattern ( https://genepattern . uth . tmc . edu/gp/pages/login . jsf ) was used with scramble cells expression pattern as the train0 set , knockdown cells’ gene expression pattern as train1 set and the Gray dataset [61] used as the test set . Default parameters were used for the analysis similar to [58] . The resulting probabilities of gene signature expression in each patient for each knockdown signature were analyzed by simple linear regression in the JMP Pro 11 software . Data , unless otherwise noted , represent the mean +/- the standard error of the mean and n indicates number of replicates used to generate the SEM . P-values were determined either by a two-tailed Student’s t-test in Excel or by a two-way ANOVA with StatView . The measurement of amino acids and acyl carnitines was performed using stable isotope dilution techniques and flow injection tandem mass spectrometry and mass spec sample preparation methods described previously [93–95] . Derivatized organic acids were analyzed by capillary gas chromatography/mass spectrometry ( GC/MS ) using a TRACE DSQ instrument ( Thermo Electron Corporation; Austin , TX ) [93–95] . All MS analyses employed stable-isotope-dilution . The standards serve both to help identify each of the analyte peaks and provide the reference for quantifying their levels . Quantification was facilitated by addition of mixtures of known quantities of stable-isotope internal standards from Isotec ( St . Louis , MO ) , Cambridge Isotope Laboratories ( Andover , MA ) , and CDN Isotopes ( Pointe-Claire , Quebec , CN ) to samples . Sample Preparation: Biological triplicates of 15cm plates under each treatment condition were placed on ice and washed twice with ice-cold PBS before as much PBS as possible was removed . Cells were lysed in 620ul ml of 0 . 78% Formic Acid in water and scraped to collect . 30ul were removed for protein quantification . 1x volume of the collected pellet ( ~800ul ) of acetonitrile was added and sample was vortexed vigorously . Aliquots were separated for mass spectrometry measurements ( 300 ul for organic acids , 100 ul for amino acids/acyl-carnitines ) and were immediately frozen on dry ice and transferred to -80°C . Analysis: Protein concentration per replicate was determined by the Pierce BCA Kit ( ThermoScientific , Waltham , MA , USA ) and used to normalize all metabolite levels . A ratio of NADP+/NADPH was calculated after measuring each molecule separately with the Amplite Fluorimetric NADP/NADPH Ratio Assay Kit from AAT Bioquest , Inc ( Sunnyvale , CA ) . Protocol was conducted as the manufacturer suggested and all values were normalized to protein content , as measured by the Pierce BCA kit , on similarly plated and treated samples done in parallel . 3 million H1975 cells were plated in 15 cm dishes , after 24 hours they were serum starved to 0 . 5% FBS , 24 hours later they were treated with either a 4-hour treatment of α-KG or an 8 hour treatment of hypoxia before collection for a native ChIP . Protocol was carried out as manufacturer suggested with the SimpleChIP Plus Enzymatic Chromatin IP Kit ( Agarose Beads ) ( Cell Signaling Tech . , cat . no . 9004 ) . Sonication and digestion were performed to obtain chromatin 1–4 nucleosomes in size , which was verified by gel electrophoresis . Each IP used chromatin prepared from 4–5 million cells and was performed overnight . Genomic DNA was extracted with the DNeasy Blood & Tissue Kit according to the protocol provided by the manufacturer ( Qiagen ) . The genomic DNAs ( 800 ng ) were modified by treatment with sodium bisulfite using the Zymo EZ DNA Methylation kit ( Zymo Research , Irvine , CA ) . Bisulfite treatment of denatured DNA converts all unmethylated cytosines to uracils , leaving methylated cytosines unchanged , allowing for quantitative measurement of cytosine methylation status . Pyrosequencing was performed using a Pyromark Q96 MD pyrosequencer ( Qiagen ) . The bisulfite pyrosequencing assays were used to quantitatively measure the level of methylation at CpG sites contained . Assays were designed to query CpG islands using the Pyromark Assay Design Software ( Qiagen ) . Pyrosequencing was performed using the sequencing primer . PCR conditions were optimized to produce a single , robust amplification product . Defined mixtures of fully methylated and unmethylated control DNAs were used to show a linear increase in detection of methylation values as the level of input DNA methylation increased ( Pearson r > 0 . 98 for all regions ) . Once optimal conditions were defined , each assay was analyzed using the same amount of input DNA from each specimen ( 40 ng , assuming complete recovery after bisulfite modification ) . Percent methylation for each CpG cytosine was determined using Pyro Q-CpG Software ( Qiagen ) . | During the development of most solid tumors , there are characteristic physiological differences in the tumor that result from tumor cells outgrowing their local blood supply . Two of these physiological differences , or “stresses , ” that occur in the tumor are low oxygen levels ( hypoxia ) and an accumulation of lactic acidic ( lactic acidosis ) . Cancer cells experiencing hypoxia and lactic acidosis tend to be more resistant to chemo- and radio-therapy and metastasize more readily . Therefore , it is important to understand how tumor cells adapt to and survive these stresses . We used a large scale screening experiment in order to find which genes and proteins are involved in tumor cell adaptation and survival under hypoxia or lactic acidosis . We found that inhibiting either of two genes involved in lipid synthesis allowed tumor cells to survive hypoxia . This occurred because silencing these genes led to an increase in the metabolite α-ketoglutarate , which repressed a transcription factor that contributed to cell death under hypoxia . This research specifically advances our understanding of how tumor cells survive hypoxia and lactic acidosis and more broadly enhances our understanding of the cellular biology of solid tumors . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | ACLY and ACC1 Regulate Hypoxia-Induced Apoptosis by Modulating ETV4 via α-ketoglutarate |
Heterozygous Twirler ( Tw ) mice develop obesity and circling behavior associated with malformations of the inner ear , whereas homozygous Tw mice have cleft palate and die shortly after birth . Zeb1 is a zinc finger protein that contributes to mesenchymal cell fate by repression of genes whose expression defines epithelial cell identity . This developmental pathway is disrupted in inner ears of Tw/Tw mice . The purpose of our study was to comprehensively characterize the Twirler phenotype and to identify the causative mutation . The Tw/+ inner ear phenotype includes irregularities of the semicircular canals , abnormal utricular otoconia , a shortened cochlear duct , and hearing loss , whereas Tw/Tw ears are severely malformed with barely recognizable anatomy . Tw/+ mice have obesity associated with insulin-resistance and have lymphoid organ hypoplasia . We identified a noncoding nucleotide substitution , c . 58+181G>A , in the first intron of the Tw allele of Zeb1 ( Zeb1Tw ) . A knockin mouse model of c . 58+181G>A recapitulated the Tw phenotype , whereas a wild-type knockin control did not , confirming the mutation as pathogenic . c . 58+181G>A does not affect splicing but disrupts a predicted site for Myb protein binding , which we confirmed in vitro . In comparison , homozygosity for a targeted deletion of exon 1 of mouse Zeb1 , Zeb1ΔEx1 , is associated with a subtle abnormality of the lateral semicircular canal that is different than those in Tw mice . Expression analyses of E13 . 5 Twirler and Zeb1ΔEx1 ears confirm that Zeb1ΔEx1 is a null allele , whereas Zeb1Tw RNA is expressed at increased levels in comparison to wild-type Zeb1 . We conclude that a noncoding point mutation of Zeb1 acts via a gain-of-function to disrupt regulation of Zeb1Tw expression , epithelial-mesenchymal cell fate or interactions , and structural development of the inner ear in Twirler mice . This is a novel mechanism underlying disorders of hearing or balance .
Twirler ( Tw ) spontaneously arose in a crossbred stock of mice segregating multiple recessive mutant alleles [1] . Heterozygous Tw mice develop obesity after three months of age , and exhibit stereotypic behavior that includes waltzing , spinning , and horizontal head-shaking [1] . This behavior is thought to result from malformed vestibular labyrinths that include hypomorphic or absent lateral semicircular canals , irregular contours of the anterior and posterior semicircular canals , and absent otoconia in the utricle and saccule [1] . In contrast , all homozygous Tw mice are born with cleft palate and die soon after birth [1] . Tw is located on proximal chromosome 18 but the causative mutation has not been identified [1] , [2] . A transgene insertional mutant , Tg9257 , exhibits a similar inner ear phenotype and is also located on proximal chromosome 18 , raising the possibility that these phenotypes are allelic [3] . However , complementation testing is inconclusive [3] . Similarly , the Irxl1 gene , located within a broad critical map interval for Tw and expressed in developing palate , has also been ruled out as a candidate for Tw [4] . Zeb1 is also located on proximal chromosome 18 and encodes a transcription factor , Zeb1 , that binds E-box-like elements to either repress [5] , [6] , or activate transcription [7]–[9] . Mice that are homozygous for a targeted deletion of exon 1 of Zeb1 ( Zeb1ΔEx1 ) die soon after birth with cleft palate , limb defects and other skeletal abnormalities , and T-cell deficiency [10] , whereas heterozygous Zeb1ΔEx1/+ mice are viable and adult females show increased adiposity [11] . This partial phenotypic overlap with Twirler does not include stereotypic vestibular behavior or inner ear malformations , although these were likely not examined in Zeb1ΔEx1 mice . Ectopic expression of Zeb1 in neoplastic epithelium has been implicated in the epithelial-to-mesenchymal transition ( EMT ) leading to local tumor invasion and metastasis [12] . In normally developing mesenchymal tissue , Zeb1 is thought to repress epithelial-specific genes such as E-cadherin and activate mesenchyme-specific genes such as collagen , smooth muscle actin and myosin [9] . Genome-wide expression profiling reveals a probable similar role for Zeb1 in the regulation of gene expression in developing mouse inner ear mesenchyme [13] . In humans , heterozygous mutations of ZEB1 cause posterior polymorphous corneal dystrophy , characterized by an epithelial transition and abnormal proliferation of corneal endothelium [14] . Zeb1ΔEx1/+ mice also show corneal abnormalities and further implicate Zeb1 in the suppression of an epithelial phenotype [15] . In the current study we show that Twirler is caused by a noncoding point substitution in the first intron of Zeb1 . The mutation does not affect splicing , but does disrupt a consensus binding site sequence for Myb proteins [16] . The maintenance of inner ear mesenchyme- and epithelium-specific gene expression is disrupted in Twirler inner ears [13] , demonstrating a novel mutation and developmental mechanism for the pathogenesis of hearing or balance disorders .
Heterozygous Tw/+ adult mice had smaller spleens ( 38±2 mg vs . 68±7 mg , P<0 . 013 ) in comparison to wild type littermates . Tw/+ thymi were also smaller although the difference was not significant ( 13±2 mg vs . 31±6 mg , P<0 . 06 ) . Tw/+ mice had lower counts of white blood cells ( 1×103/µl vs . 7 . 2×103/µl , P<0 . 0001 ) , lymphocytes ( 0 . 5×103/µl vs . 5 . 9×103/µl , P<0 . 0004 ) and polymorphonuclear neutrophils ( 0 . 4×103/µl vs . 1 . 3×103/µl , P<0 . 04 ) . No abnormalities were found in other adult Tw/+ tissues . Histopathological examination of P0 animals revealed no abnormalities in the thymus or spleen of wild type , Tw/+ or homozygous Tw/Tw mice . Tw/Tw mice had cleft palates . There was no significant difference in average body weight between Tw/+ and wild type littermates of either sex until 12 weeks of age ( Figure 1A and 1B ) . Beginning at seven weeks of age , Tw/+ mice consumed approximately 15 to 20% more food than wild type littermates ( Figure 1C and 1D ) . There was a significant increase in the percentage of body fat and slightly reduced lean body mass in Tw/+ mice of both sexes ( Table 1 ) , indicating that fat accounts for the increased body mass . Body weight-adjusted energy expenditure , estimated from oxygen consumption , revealed a reduced metabolic rate in Tw/+ mice that did not reach statistical significance ( Table 1 ) . Tw/+ mice had normal serum glucose levels but elevated levels of serum free fatty acids , triglycerides , insulin , leptin , corticosterone and adiponectin ( Table 1 ) . Insulin and glucose tolerance tests of 15-week-old females showed insulin resistance and slight glucose intolerance in Tw/+ mice ( Figure 1E and 1F ) , consistent with data for other obese mice with hyperinsulinemia [17] . We evaluated the morphology of mutant inner ears using the paint-filling technique ( Figure 2 ) . The Tw/+ inner ears had grossly intact semicircular canals and neurosensory cristae ampullaris , but the contours of the canals were irregular due to small bulges and projections ( Figure 2B ) . The most anatomically consistent malformation was found at the non-ampullated end of the lateral canal where it normally narrows to join the vestibule in wild type ears ( Figure 2D ) . In contrast , the non-ampullated ends of Tw/+ lateral canals were irregular or constricted ( Figure 2E ) . Tw/Tw inner ears have more severe malformations that include absence of the lateral semicircular canal , truncation of the posterior semicircular canal , and shortening of the cochlear duct ( Figure 2C , 2F and 2I ) . The average length of Tw/+ cochlear ducts ( Figure 2H ) was 91% ( ±5% ) that of wild type ears ( P<0 . 00002; Figure 2G ) . Binaural average ABR thresholds were elevated for Tw/+ mice in comparison to wild type controls at one month of age ( 33±1 . 6 dBSPL vs . 55±5 . 3 dBSPL at 8 kHz , p<0 . 0006; 33±1 . 8 dBSPL vs . 46±4 dBSPL at 16 kHz , p<0 . 01; 29±1 . 9 dBSPL vs . 39±3 . 6 dBSPL at 32 kHz , p<0 . 023; Figure 2J ) . Tw/+ mice showed no significant change in ABR thresholds measured at three months of age in comparison to thresholds measured at one month of age ( not shown ) . Tw/+ utricles had giant otoconia that were transparent by light microscopic examination but visible by scanning electron microscopy ( Figure 2N ) . In contrast , Tw/+ saccular otoconia appeared normal ( Figure 2L ) . We screened 1679 [ ( C57BL/6J-Tw/+ x CAST/Ei ) F1-Tw/+ x C57BL/6J]N2 progeny for recombinations . Recombination locations were refined with additional markers to narrow the Tw interval to 814 kb between D18Nih6 and D18Nih42 ( Figure 3A ) . This interval was five Mb proximal to the Tg9257 transgene insertion site [3] . The Tw interval contained three genes: Zeb1 , Zeb1os ( Zeb1 opposite strand transcript , annotated in MGI as predicted gene Gm10125 ) and Zfp438 ( Figure 3A ) . Zeb1 encodes a transcription factor with two zinc finger motifs and one homeobox motif . Zeb1os is predicted to encode a long noncoding RNA of unknown function . It is located on the opposite strand of Zeb1 where the two overlapping genes share parts of their first introns . Finally , Zfp438 is predicted to encode a zinc finger protein whose biological function is unknown [18] . Zeb1 was a good candidate for the gene mutated in Tw based upon the phenotype associated with a targeted deletion allele , Zeb1ΔEx1 . Homozygous Zeb1ΔEx1 mice are born with cleft palate , skeletal and thymus abnormalities , and die shortly after birth [10] . We observed that Zeb1ΔEx1/+ heterozygotes have inner ear morphology and hearing thresholds that are indistinguishable from those of wild type littermates , whereas Zeb1ΔEx1/ΔEx1 homozygotes have a subtle constriction of the midportion of the lateral semicircular canal that differs in location and severity from that observed in Tw/+ mice ( Figure S1 ) . This difference is probably not due to genetic background since both lines were congenic on a C57BL/6J background . To determine if Tw and Zeb1ΔEx1 can complement to form a normal palate or inner ear , we crossed heterozygous Tw and heterozygous Zeb1ΔEx1 mice . We observed an approximate Mendelian ratio of genotypes: five +/+ , five Tw/+ , seven Zeb1ΔEx1/+ and eight Tw/Zeb1ΔEx1 . All Tw/Zeb1ΔEx1 mice were born with normal palates and developed into adults with circling behavior typical of Tw/+ mice . The lateral semicircular canals resembled those of Tw/+ mice ( Figure S1 ) . These results suggest these mutations exert their effects via different genes or mechanisms . While the Zeb1 pathway may be altered in Twirler mice , it is unlikely to be due to a loss-of-function allele of Zeb1 . To identify the Tw mutation , we first used 5′-RACE and 3′-RACE to identify novel exons of Zeb1 , Zeb1os and Zfp438 . 5′-RACE revealed Zeb1 transcripts with each of five additional alternative first exons ( designated 1b , 1c , 1d , 1e and 1f ) between exon 1 ( heretofore termed exon 1a ) and exon 2 ( Figure S2 ) . We amplified and sequenced all novel and annotated exons of Zeb1 , Zeb1os and Zfp438 from genomic DNA of Tw/Tw , Tw/+ and wild type mice . We also amplified and sequenced cDNA transcripts of these genes from embryonic mRNA . All major transcripts of these genes were amplified from mice with each genotype . We found no sequence differences in the cDNAs or genomic exons . Sequence analysis of the 192-bp region of overlap of Zeb1 and Zeb1os revealed a single nucleotide substitution ( G>A ) 181 bp downstream of Zeb1 exon 1 and 12 bp downstream of Zeb1os exon 1 in Tw ( Figure 3B ) . We designated this Tw variant as c . 58+181G>A , which was the only sequence variation we detected . The wild type variant c . 58+181G was conserved among 13 normal control inbred mouse strains as well as other vertebrate species ( Figure 3B ) . In silico analyses ( NNsplice , GeneSplicer , Net2Gene ) predict that c . 58+181G>A does not affect splicing of the adjacent splice donor site for exon 1 of Zeb1os . Sequence analysis of Zeb1 and Zeb1os cDNA transcripts confirmed no effect of c . 58+181G>A on splicing . c . 58+181G>A disrupts a predicted site for Myb protein binding ( Figure 3B ) [16] . To test if this change can alter the binding of a Myb protein , recombinant mouse C-Myb was expressed and purified for an electrophoretic mobility shift assay ( EMSA ) of its binding to oligonucleotide probes containing either c . 58+181G or c . 58+181A and the flanking genomic sequences . There was a shift of the mobility of the wild type DNA probe in the presence of C-Myb , while the Tw DNA probe mobility was unchanged ( Figure 4A ) . The binding of C-Myb to wild type DNA was inhibited by both the wild type probe and a mim-1 control probe which has been shown to interact with C-Myb [19] , but not by the Tw probe ( Figure 4B ) . These data provide in vitro evidence that the Tw mutation can disrupt binding of a Myb protein ( C-Myb ) to the mutated first intronic sequence of Zeb1 . We analyzed mRNA expression levels of Zeb1 , Zeb1os and Zfp438 from inner ears of Tw/Tw , Tw/+ or wild type mice at E13 . 5 . We performed the same analysis with Zeb1ΔEx1 heterozygotes , homozygotes , and wild type littermates . We designed primer pairs to specifically amplify Zeb1 transcripts starting from each of exons 1a , 1b , 1c , 1d , 1e or 1f . One primer pair for constitutively spliced exons 2 and 3 was designed to amplify all Zeb1 transcripts . The levels of Zeb1 transcripts containing exon 1b , 1c , 1d , 1e , or 1f , as well as the Zeb1os and Zfp438 transcripts , were too low to be reliably quantified by RT-PCR . The levels of transcripts containing exons 1a and 2 , as well as exons 2 and 3 , were significantly increased from the Tw allele of Zeb1 ( Zeb1Tw ) in comparison to wild type Zeb1 ( Figure 5A ) . In contrast , Zeb1ΔEx1 expressed no Zeb1 transcripts containing exons 1a and 2 , and nearly non-detectable levels of any other Zeb1 transcripts containing other exons ( Figure 5B ) . Transcripts levels for the closely related Zeb2 gene were unchanged among all three Zeb1 genotypes ( Figure 5B ) . These results indicate that Zeb1ΔEx1 is a loss-of-function allele whereas Zeb1Tw is likely to act via gain-of-function . To confirm the pathogenic effect of c . 58+181G>A , we generated two knockin mouse lines: KIA segregates the Tw variant c . 58+181A and KIG segregates the wild type variant c . 58+181G ( Figure 6 ) . Compound heterozygous KIG/KIA mice consumed more food and grew heavier with increased adiposity in comparison to KIG/KIG control males and females ( Figure 7A–7D , Table 2 ) . The energy expenditure and circulating hormone levels in KIG/KIA mice recapitulated the Tw/+ phenotype ( Table 2 ) . The reduction in body weight-adjusted energy expenditure reached statistical significance in KIG/KIA female mice , whereas it did not in Tw/+ females ( Table 1 ) . Insulin and glucose tolerance tests showed insulin resistance and slight glucose intolerance in KIG/KIA mice ( Figure 7E and 7F ) . Although KIG/KIA mice showed neither circling behavior nor constricted semicircular canals , the semicircular canals were irregular ( Figure 8B ) and the utricles contained giant otoconia ( Figure 8N ) . Average ABR thresholds for KIG/KIA and KIG/KIG mice were not significantly different ( Figure 8J ) . KIA/KIA and KIA/Tw inner ears displayed the same malformations as Tw/Tw ears ( Figure 8C , 8F , 8I , and Figure 9B , 9D , 9F ) . KIG/KIA average spleen weight was decreased by 15% ( P<0 . 05 ) but average thymus weight did not differ relative to KIG/KIG littermates ( Table 2 ) . We observed cleft palate with or without cleft lip in KIA/KIA and KIA/Tw mice with 50% and 90% penetrance , respectively ( not shown ) . We did not observe cleft palate or cleft lip in KIG/KIA , KIG/KIG or KIG/Tw mice , indicating that the recapitulation of the Tw phenotype is specific . The different phenotypic severity and penetrance of KIA in comparison to Tw could result from genetic background differences , since Tw arose on a different undefined stock . However , we serially backcrossed Tw to wild type C57BL/6J for over 30 generations , and KIA was generated from C57BL/6-derived Bruce4 ES cells and maintained on an isogenic C57BL/6J background . Therefore the differences in severity and penetrance could result from closely linked sequence variation , the residual loxP site in KIA , or a combination of these effects . To determine if Zeb1 protein is expressed from the Tw allele , we stained inner ears of Tw/Tw mice with anti-Zeb1 antibodies ( Figure 10A and 10B ) . We observed Zeb1 expression in non-epithelial ( mesenchymal ) cells surrounding Tw/Tw inner ears in which epithelial and mesenchymal tissue compartments could be microanatomically differentiated ( Figure 10B ) . Other Tw/Tw inner ears had poorly preserved microarchitecture , precluding a differentiation of epithelium versus mesenchyme ( Figure 10C ) . We conclude that Zeb1 protein is expressed in Tw/Tw ears , consistent with the result of real-time RT-PCR . To determine if Zeb1 protein levels are altered by Tw , we performed a western blot analysis of inner-ear or whole-head protein extracts from E13 . 5 mice . We compared Tw/Tw , Tw/+ and wild type littermates , as well as KIG/KIG , KIG/KIA and KIA/KIA littermates . We were unable to detect Zeb1 in inner-ear protein extracts , but were able to reliably detect it in samples from whole heads . Total Zeb1 protein levels appeared to be slightly increased by Tw in comparison to wild type littermates ( Figure 10D ) . This difference was not significant ( ANOVA , P>0 . 05 ) , possibly due to small numbers of animals and the degree of variation of Zeb1 band intensities within genotypes ( Figure 10F ) . In contrast , Zeb1 protein levels in KIG/KIA and KIA/KIA mice were 2- to 3-fold higher than in KIG/KIG littermates ( Figure 10E ) . The variation within knockin genotype groups was smaller , resulting in differences between knockin genotype groups that were significant ( P<0 . 05 ) ( Figure 10G ) .
This study demonstrates that the phenotype of Twirler is caused by a noncoding nucleotide substitution within a shared first intron of the Zeb1 and Zeb1os genes on mouse chromosome 18 . This is a rare example of a Mendelian noncoding point mutation that does not affect a splice site or promoter . Our results demonstrate the potential for complex phenotypic effects of noncoding point variants , which are increasingly implicated in association studies of genetically complex traits . Our recombinant knockin mouse model and wild type knockin control for testing the pathogenic potential of the Tw mutation may be a useful paradigm to explore the effects of other noncoding variants of unknown pathogenic potential . The altered penetrance potentially associated with a residual loxP site in the Tw knockin line serves a cautionary note to include a wild type knockin control . Although the initial study by Lyon [1] described abnormal development of the sensory neuroepithelium in the cristae ampullaris of some semicircular canals of Tw/+ mice , we have not observed the same alteration . Instead we observed a highly penetrant constriction of the non-ampullated end of the lateral semicircular canal that could impede or prevent the flow of endolymph and disrupt neurosensory detection of angular acceleration . A difference in strain background [1] may account for the different result . Moreover , Lyon reported that utricular otoconia were absent in Tw/+ ears whereas we observed giant utricular otoconia . This difference could also result from the strain background difference , loss of giant otoconia during the dissection process , or our use of scanning electron microscopy in addition to light microscopy . Nevertheless , either of the described utricular phenotypes could impair the detection of linear acceleration by Tw/+ utricles . We conclude that our observed semicircular canal and utricular anomalies underlie the vestibular behaviors of Tw/+ mice , although we cannot estimate their relative contributions to the observed vestibular functional phenotype . Correlating mouse vestibular structural or functional abnormalities with behavior is difficult due to a complex interrelationship between vestibular behavior and anxiety that is also dependent upon strain background [20] . The cause of hearing loss observed in some Tw/+ mice also remains obscure . Postmortem examination of middle ears did not reveal otitis media or developmental malformations of the external or middle ears that could account for the hearing loss . Although severe hearing loss has been observed in other mouse mutants with much shorter cochlear ducts [21] , the severity of hearing loss in Tw/+ mice was highly variable but the degree of shortening of the cochlear duct was nearly constant . This lack of correlation leads us to conclude that associated physiologic defects or undetected structural anomalies underlie hearing loss in Tw/+ mice . The Tw/+ phenotype includes hyperphagia with elevated levels of circulating corticosterone and adiponectin that are similar to those in a corticotropin-releasing factor ( CRF ) transgenic mouse model of Cushing syndrome [22] , [23] . Other phenotypic similarities of that model to Tw/+ include increased body weight and adiposity , alopecia , atrophy of the thymus and spleen , and muscle wasting . This may suggest that Tw disrupts , at least in part , the hypothalamus-pituitary-adrenal axis . This is consistent with expression of Zeb1 in the pituitary gland [24] . However , Zeb1 protein is also expressed in adipose tissue and increases during adipogenesis in cell culture [11] . Moreover , Zeb1ΔEx1/+ mice develop obesity that is not associated with hyperphagia [11] , unlike Tw/+ mice ( Figure 1C and 1D ) . Therefore different mechanisms or tissues may underlie obesity phenotypes associated with Zeb1ΔEx1 and Zeb1Tw . The pathogenetic mechanism for one or both of these Zeb1 alleles may also underlie a locus for susceptibility to obesity on human chromosome 10p11 [25]-[28] , which includes the human ZEB1 gene . The results presented here and in Hertzano et al . [13] , in combination with the body of published data on Zeb1 in cancer and normal development , show that Zeb1 is a master regulator of mesenchyme-specific gene expression in the developing mouse ear . Twirler is a novel example of a disorder of hearing or balance caused by a disruption of mesenchymal-epithelial identities or interactions . A similar lateral semicircular canal phenotype is seen in other hyperactive circling mice , including epistatic circler mice [29] and mice segregating a gene-trap allele of Chd7 [30] . Chd7 encodes a chromodomain protein required for the development of multipotent migratory neural crest cells [31] , which includes an epithelial-to-mesenchymal transition . An auditory-vestibular phenotype approximating that of Twirler and Chd7 mutant mice is also observed in human patients with CHARGE syndrome and mutations of the human CHD7 gene [32] , [33] . Semicircular canal formation is also known to require Bmp4 [34] and heterozygosity for a knockout allele of mouse Bmp4 primarily affects the lateral semicircular canal [35] . Bmp4 is a member of the transforming growth factor-β ( TGF-β ) super-family [36] , providing another link to Zeb1 since Zeb1 and Zeb2 have been implicated in TGF-beta/BMP signaling [8] . Why do Twirler mice have a different inner ear phenotype than Zeb1ΔEx1 mice ? Genetic background differences seem unlikely to account for this difference since Zeb1ΔEx1 and Twirler were both maintained on a congenic C57BL/6J background . It is possible that other Zeb1 transcripts could compensate for the loss of exon 1 in Zeb1ΔEx1 ears , but our quantitative RT-PCR and expression profiling results [13] render this hypothesis unlikely . Alternatively , the closely related Zeb2 gene may be able to compensate for the loss of Zeb1 expression in the inner ear , but not other affected tissues such as the palate or lymphoid system . However , our quantitative RT-PCR results revealed no compensatory change in Zeb2 transcript levels in the mutants . It is also possible that disruption of Zeb1os may contribute to the Tw phenotype . Ectopic expression of an analogous long noncoding antisense RNA in epithelial cells leads to altered Zeb2 RNA splicing , increased Zeb2 protein levels , and epithelial-to-mesenchymal transition [37] . However , Zeb1os RNA levels were too low for us to reliably detect and monitor by either qRT-PCR or Northern blot analyses to confidently address this possibility ( data not shown ) . Finally , perhaps Twirler does not exert its pathogenic effect via Zeb1 . This also seems unlikely since there is significant phenotypic overlap of Twirler with Zeb1ΔEx1 , including abnormalities of the semicircular canals associated with both mutant alleles . Furthermore , the phenotypic effects of compound heterozygosity for Zeb1ΔEx1 and Tw are consistent with the conclusion that Zeb1ΔEx1 is an amorphic or hypomorphic allele whereas Twirler acts as a hypermorphic or neomorphic allele to misregulate Zeb1 expression . Our electrophoretic mobility shift experiment ( Figure 4 ) suggests that Tw exerts its pathogenic effect by disruption of binding of C-Myb or other Myb proteins to the first intron of Zeb1 . There are also published observations supporting the general hypothesis that loss of Myb protein binding leads to de-repression of Zeb1Tw and inner ear malformations: First , C-Myb can function as either an activator or repressor of gene transcription [38] and is thought to function in regulation of epithelial-mesenchymal cell identity [39] . Second , a pathogenic effect of up-regulation of developmental transcription factors has been demonstrated for Pax6 in the eye [40] and Tbx1 in the inner ear [41] . In the inner ear , increased expression of Tbx1 can cause malformations that include incomplete coiling and reduced extension of the cochlear duct [41] . Furthermore , Tbx1 expression in the periotic mesenchyme is required for cochlear duct outgrowth [42] , suggesting a potential link to the observed inner ear phenotype of Twirler . Taken together , these observations and our results support the hypothesis that Twirler disrupts inner ear development via mis-regulation of Zeb1 . The cell type-specific gene expression profiles of Twirler ears [13] suggest that a pathologic disruption of epithelial and mesenchymal cell identities underlies the inner ear malformations . This could arise from a loss of mesenchymal cell identity leading to mesenchymal-epithelial transition ( MET ) , a loss of epithelial cell identity leading to epithelial-mesenchymal transition ( EMT ) , or a combination of these mechanisms . Although the gene expression profiles [13] seem consistent with MET , it is difficult to conceive a simple MET pathway that does not invoke a loss-of-function mechanism in Tw mesenchyme . In contrast , EMT would involve a gain-of-function with ectopic expression of Zeb1 in Tw inner ear epithelium . Indeed , ectopic expression of Zeb proteins in other epithelial tissues has been shown to lead to EMT in other neoplastic and developmental processes [43] . Distinguishing among EMT and MET mechanisms may be difficult if they involve complex regulatory pathways mediated by Zeb1os , Zeb2 , microRNAs or other genes . In summary , we have identified the pathogenic mutation of Twirler as a noncoding point mutation that leads to over- or mis-expression of Zeb1 , pathologic alterations of gene expression [13] , cell fate and interactions in the developing inner ear . The ultimate result is a gross alteration of the structure and function of the vestibular and auditory organs . Disruption of epithelial-mesenchymal identity or interactions may be a shared pathogenetic mechanism underlying phenotypes that primarily affect development of the lateral semicircular canals , extension of the cochlear duct , or both .
Mice were maintained on a 12∶12-h light-dark cycle . All experiments and procedures were approved by the Animal Care and Use Committees of the National Institute of Diabetes and Digestive and Kidney Diseases , National Institute of Neurological Disorders and Stroke and National Institute on Deafness and Other Communication Disorders . Twirler mice were a kind gift from Drs . Miriam Meisler and Siew-Ging Gong at the University of Michigan and were maintained on a C57BL/6J background by backcrossing heterozygous Tw males to C57BL/6J females for at least 30 generations . Zeb1ΔEx1 mice [10] were a generous gift from Dr . Douglas Darling and were serially backcrossed to C57BL/6J to maintain the line . Bacterial artificial chromosome ( BAC ) clone RP23-135A18 containing mouse genomic DNA encoding exon 1 of Zeb1 was digested with PacI/SphI and SphI to yield 7 . 6-kb and 2 . 6-kb homology arms , respectively , for targeting constructs ( Figure 6 ) . Each targeting construct included loxP sites flanking a splice acceptor site and internal ribosomal entry site ( IRES ) ( pGT1 . 8IresBgeo , provided by Austin Smith at University of Edinburgh ) [44] , E . coli lacZ , and a reverse-oriented pPGK-neomycin resistance cassette cloned into the pPNT plasmid [45] ( Figure 6 ) . The wild type ( KIG ) and Twirler ( KIA ) 7 . 6-kb PacI/SphI homology arms contained G and A at position c . 58+181 , respectively . Bruce4 embryonic stem ( ES ) cells [46] were electroporated with the KIG or KIA targeting constructs and grown in the presence of G418 and ganciclovir , using standard protocols at the University of Michigan Transgenic Animal Model Core [47] . G418-resistant ES clones were screened for homologous recombination by PCR and Southern blot analyses . At least three recombinant ES cell lines for each targeting construct were injected into C57BL/6 blastocysts . Chimeric males were mated with C57BL/6 females and offspring were analyzed by Southern blot and PCR analyses for germline transmission of KIG or KIA ( Figure S3 ) . Table S1 shows PCR primer pairs used to genotype KI alleles before and after neomycin cassette removal ( Figure 6 ) . Mice transmitting KIG or KIA in the germline were crossed to Cre recombinase-expressing mice ( C57BL/6-TgN ( Zp3-Cre ) 93Knw , Jackson laboratory , ME ) to delete the IRES-lacZ-neomycin resistance cassette , leaving a single loxP site 606 bp downstream from Zeb1 exon 1 . A comprehensive gross anatomical , histological , and serological analysis of three 15-week-old Tw/+ and three wild type littermate males was performed as described [48] . Tissue sections from two Tw/Tw , two Tw/+ and two wild type mice at postnatal day 0 ( P0 ) were analyzed . Heterozygous Tw males and females were mated . Pregnant females were identified by the presence of a vaginal plug and gestational stage was estimated by defining that morning as 0 . 5 days post-conception ( dpc ) . Embryos at 14 . 5 dpc were harvested and processed for paint-filling as described [49] . The length of the cochlear duct was measured along its outer contour from a ventral view [50] . For scanning electron microscopy ( SEM ) , whole-mounted inner ears were fixed in 2 . 5% glutaraldehyde in 0 . 1 M sodium cacodylate with 2 mM CaCl2 for 90 min . The organ of Corti , saccule , utricle , and crista ampullaris were dissected free in water and dehydrated with a serial dilution series of acetone . Samples were critical point-dried and sputter-coated followed by visualization with a field-emission scanning electron microscope ( S-4800 , Hitachi ) . Auditory brainstem response ( ABR ) thresholds were measured in response to click or pure-tone stimuli of 8 , 16 , or 32 kHz as described [51] . Six Tw/+ male , six Tw/+ female , six wild type male and six wild type female mice were housed individually with regular mouse chow and water provided ad libitum . Body weights were measured weekly from 5 weeks of age . Weekly food intake was measured from weeks 6 through 22 to calculate average daily food intake . At 23 weeks of age , mice were transferred to the NIDDK Mouse Metabolism Core Laboratory for measurement of oxygen consumption , carbon dioxide production and motor activity as described [52] . Body composition was measured using Echo3-in-1 NMR analyzer ( Echo Medical Systems , Houston , TX ) . Tail vein blood was used for serologic analyses . Fifteen-week-old female mice ( eight Tw/+ , six wild type ) were tested for glucose and insulin tolerance as described [52] . All data are expressed as a mean ± SEM . Student's t-test was used to identify statistically significant differences between genotype groups . Twirler males ( C57BL/6J-Tw/+ ) were crossed with DBA/2J or Castaneus ( CAST/Ei ) females since Twirler females are poor caretakers of offspring . Male ( C57BL/6J-Tw/+ x DBA/2J ) F1-Tw/+ or ( C57BL/6J-Tw/+ x CAST/Ei ) F1-Tw/+ progeny were backcrossed with DBA/2J or C57BL/6J females , respectively , to generate 337 and 1679 N2 backcross progeny , respectively . Progeny were scored for circling behavior or obesity by visual inspection . We genotyped short tandem repeat ( STR ) markers on 337 DBA/2J N2 backcross progeny to identify two STR markers ( D18Mit65 , D18Mit64 , D18Mit19 and D18Umi1 ) flanking each side of Tw . These markers were genotyped in the 1679 CAST/Ei N2 backcross progeny to identify recombinations in the Tw region . The Tw map interval was defined by genotypes of additional markers in the recombinants . We genotyped MIT markers between D18Mit65 and D18Umi1 , as well as 40 novel STR markers ( denoted D18Nih1 through D18Nih44; PCR primer sequences listed in Table S1 ) located between D18Mit19 and D18Mit219 . Genomic DNA of Tw/Tw , Tw/+ and wild type mice were isolated for PCR amplification as described [53] . The primers were designed to amplify and sequence all of the annotated exons of the Zeb1 , Zeb1os ( MGI predicted gene Gm10125 ) and Zfp438 genes in the Tw critical interval . Additional novel exons were identified by 5′ and 3′- RACE ( 5′ and 3′ rapid amplification of cDNA ends ) of the Zeb1 , Zeb1os and Zfp438 genes . This revealed multiple alternative first exons for Zeb1 that were also sequenced . Reverse transcription ( RT ) -PCR was performed to amplify and sequence full-length cDNA clones of the three genes using whole body mRNA collected from embryonic Tw/Tw , Tw/+ and wild type littermates . PCR reaction conditions were modified to amplify and sequence the overlapping genomic region of Zeb1 and Zeb1os . Fifty-µl PCR reactions contained 50 to 100 ng of genomic DNA , 5 pmol each of forward and reverse primers , 200 mM each dNTP , 0 . 5 M betaine , 10% dimethyl sulfoxide ( DMSO ) , 2 . 5 mM MgCl2 , and 0 . 5 U of thermostable polymerase . Thermal cycling conditions were: 95°C for 1 min; 33 cycles of 20 s at 95°C , 20 s at 57°C , and 45 s at 72°C; and a final 2-min extension at 72°C . For sequencing , 50 µl PCR reaction products were purified with a QIAquick PCR purification kit ( Qiagen , Hilden , Germany ) and eluted with 30 µl elution buffer . Three µl of purified products were added to a 10-µl sequencing reaction containing 3 . 2 pmol primer , 0 . 25 µl Big Dye Terminator Ready Reaction mix ( PE Biosystems ) , sequencing buffer and 10% DMSO . Cycling conditions were 96°C for 2 min and 33 cycles of 96°C for 10 s , 55°C for 10 s , and 60°C for 4 min . We also amplified and sequenced the overlapping genomic region of Zeb1 and Zeb1os from normal mouse control strains 129/J , AKR/J , BALB/cJ , C3H/HeJ , C57BL/6J , C58/J , CBA/J , CE/J , DBA/2J , P/J , RF/J , SEA/GnJ and SWR/J DNA . Double-stranded oligodeoxyribonucleotide probes were synthesized to encode genomic sequences containing c . 58+181G ( 5′-TGCTGGACTGGACCGTTATGTCTTACCTGC and 5′-GCAGGTAAGACATAACGGTCCAGTCCAGCA ) , c . 58+181A ( 5′-TGCTGGACTGGACCATTATGTCTTACCTGC and 5′-GCAGGTAAGACATAATGGTCCAGTCCAGCA ) , or a C-Myb binding site control from the mim-1 gene [19] ( 5′-GCTCTAAAAAACCGTTATAATGTACAGATATCTT and 5′-AAGATATCTGTACATTATAACGGTTTTTTAGAG ) . Probes were end-labeled with [γ-32P]ATP by T4 Polynucleotide Kinase ( New England Biolabs ) . Mouse C-Myb cDNA was cloned in pET-41a ( + ) ( Novagen ) , and the protein was expressed in E . coli strain BL21 ( DE3 ) pLys ( Invitrogen ) and purified with Ni-NTA columns ( Qiagen ) . Twenty-μl reactions were performed with the EMSA Accessory Kit ( Novagen ) . Unlabeled oligonucleotide competitors were added at 25- or 50-fold molar excess . Binding reaction products were separated by 6% DNA retardation gel electrophoresis ( Invitrogen ) and visualized with a Typhoon Trio+ ( GE Healthcare ) . Inner ears with adjacent mesenchyme were microdissected from E13 . 5 offspring of Tw/+ x Tw/+ matings . Total RNA was isolated from inner ears using PicoPure ( Applied Biosystems , Foster City , CA ) . Total RNA from 10 to 14 ears of the same genotype was pooled and purified with the RNAeasy MinElute Cleanup kit ( Qiagen ) . RNA integrity was measured with an Agilent 2100 Bioanalyzer ( Applied Biosystems ) . One µg of total RNA was reverse-transcribed with oligo ( dT ) primers and SuperScriptIII ( Invitrogen , Carlsbad , CA , USA ) . For TaqMan real-time PCR , PCR primers were designed to amplify Zeb1 exons 1a to 2 , 1b to 2 , 1c to 2 , 1d to 2 , 1e to 2 , 1f to 2 , and 2 to 3 , Zeb1os exons 1 to 2 , and Zfp438 exons 3 to 4 with ZEN double-quenched probes containing a 5′ FAM fluorophore , 3′ IBFQ quencher , and an internal ZEN quencher ( IDT , Coralville , IA ) . Sequences for the primers and probes are listed in Table S1 . Comparative TaqMan assays were performed in triplicate on an ABI 7500 real-time PCR system ( Applied Biosystems ) . PCR reactions were performed in a 50-µl volume containing 5 µl cDNA , 5 µl primer mix ( IDT ) , and 25 µl of Universal PCR Master Mix ( Applied Biosystems ) . Cycling conditions were 50°C for 2 min , 95°C for 10 min , followed by 40 cycles of 15 s at 95°C and 1 min at 60°C . Relative expression was normalized as the percentage of β-actin expression , and calculated using the comparative threshold cycle method of 2−ΔΔCT . Data are presented as mean values ± S . D . from six technical replicates . ANOVA was used to identify statistically significant differences between genotype groups ( P<0 . 05 ) . Proteins were extracted from E13 . 5 mouse inner ears or whole heads with NE-PER Nuclear and Cytoplasmic Extraction Reagents ( Pierce Biotechnology ) in the presence of Halt Protease Inhibitor Cocktail ( Thermo Fisher Scientific Inc . ) . Proteins were separated by SDS-PAGE in 4–20% NuPage Bis-Tris gels followed by transfer to PVDF membranes ( Millipore Corp . , Billerica , MA ) . Proteins were detected with primary antibodies for Zeb1 ( ab64098 , Abcam , 1∶200 ) and β-actin ( A2228 , Sigma-Aldrich , 1∶1000 ) . Secondary antibodies were conjugated with Cy 3 or Cy 5 ( GE Healthcare ) and detected with a Typhoon Trio+ ( GE Healthcare ) . Band density was measured using ImageQuant TL software . β-actin levels were used for normalization . ANOVA analysis of two to six biological replicates from each genotype group was used to identify statistically significant differences ( P<0 . 05 ) . Mouse inner ear sections were harvested , processed and immunostained with anti-Zeb1 or anti-CD326 antibodies as described in Hertzano et al . [13] . CD326 is also known as epithelial cell adhesion/activating molecule ( EpCAM ) that serves as a specific antigenic marker for epithelial cells [13] . | Twirler ( Tw ) mice have a combination of abnormalities that includes cleft palate , malformations of the inner ear , hearing loss , vestibular dysfunction , obesity , and lymphoid hypoplasia . In this study , we show that the underlying mutation affects the Zeb1 gene . Zeb1 was already known to encode a protein normally expressed in mesenchymal cells , where it represses expression of genes that are uniquely expressed in epithelial cells . The Tw mutation is a rare example of a single-nucleotide substitution in a region of a gene that does not encode protein , promoter , or splice sites , so we engineered a mouse model with the mutation that confirmed its causative role . The Tw mutation disrupts a consensus DNA binding site sequence for the Myb family of regulatory proteins . We conclude that this mutation leads to abnormal expression of Zeb1 , structural malformations of the inner ear , and a loss of hearing and balance function . A similar mechanism may underlie other features of Twirler , such as obesity and cleft palate . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"biology",
"veterinary",
"science"
] | 2011 | A Noncoding Point Mutation of Zeb1 Causes Multiple Developmental Malformations and Obesity in Twirler Mice |
Kaposi’s sarcoma-associated herpesvirus ( KSHV ) is a human oncogenic virus associated with Kaposi’s sarcoma and two B-cell malignancies . The rhesus monkey rhadinovirus ( RRV ) is a virus of nonhuman primates that is closely related to KSHV . Eph family receptor tyrosine kinases ( Ephs ) are cellular receptors for the gH/gL glycoprotein complexes of both KSHV and RRV . Through sequence analysis and mutational screens , we identified conserved residues in the N-terminal domain of KSHV and RRV glycoprotein H that are critical for Eph-binding in vitro . Homology-based structural predictions of the KSHV and RRV gH/gL complexes based on the Epstein-Barr-Virus gH/gL crystal structure located these amino acids in a beta-hairpin on gH , which is likely stabilized by gL and is optimally positioned for protein-protein interactions . Guided by these predictions , we generated recombinant RRV and KSHV strains mutated in the conserved motif as well as an RRV gL null mutant . Inhibition experiments using these mutants confirmed that disruption of the identified Eph-interaction motif or of gL expression resulted in complete detargeting from Ephs . However , all mutants were infectious on all cell types tested , exhibiting normal attachment but a reduction in infectivity of up to one log order of magnitude . While Eph-binding-negative RRV mutants were replication-competent on fibroblasts , their infectivity was comparatively more reduced on endothelial cells with a substantial subpopulation of endothelial cells remaining resistant to infection . Together , this provides evidence for a cell type-specific use of Ephs by RRV . Furthermore , our results demonstrate that gL is dispensable for infection by RRV . Its deletion caused a reduction in infectivity similar to that observed after mutation of Eph-binding residues in gH . Our findings would be compatible with an ability of KSHV and RRV to use other , less efficient entry mediators in lieu of Ephs , although these host factors may not be uniformly expressed by all cells .
Kaposi’s sarcoma-associated herpesvirus ( KSHV ) , the etiological agent of Kaposi’s Sarcoma [1] , is also closely associated with two B-cell malignancies , namely the primary effusion lymphoma [2] and the plasmablastic variant of multicentric Castleman’s disease [3] ( reviewed in [4] ) . Together with the rhesus monkey rhadinovirus ( RRV ) , a closely related herpesvirus of rhesus macaques , KSHV belongs to the rhadinovirus , or γ2 , genus of herpesvirus [5] . Two RRV isolates , RRV isolate 26–95 [6] and RRV isolate 17577 [7] , representing different subtypes have been characterized . RRV shares many biological features with KSHV and is therefore regarded as an in vivo model system for many aspects of γ2-herpesvirus biology [8–10] ( reviewed in [11] ) . With regard to entry into target cells , some differences but also strong similarities exist between KSHV and RRV . A prominent difference is the interaction with integrins , which is shared by most herpesviruses ( reviewed in [12] ) . In the case of KSHV , interaction with integrins is mediated through glycoprotein B ( gB ) [13] , the conserved herpesviral fusion executor . Detectable interaction of RRV with integrins has not been observed , at least not via the same glycoprotein or mechanism [14] . On the other hand , the interaction of the respective gH/gL glycoprotein complex of KSHV and RRV with members of the Ephrin receptor tyrosine kinase ( RTK ) family of proteins ( Ephs ) is a conserved feature in the entry process of both rhadinoviruses . Whether the interaction of rhadinoviral gH/gL with Ephs also promotes fusion is so far unclear . In contrast , contribution to virus endocytosis , trafficking , and establishment of infection has been described by several reports [15–19] . KSHV binds EphA2 with high affinity and only exhibits very weak interactions with other A-type Ephs [15 , 16] . Despite very divergent primary sequences of gH and gL of RRV 26–95 and 17577 , both isolates were found to interact with a broad spectrum of A- and B-type Eph receptors and to bind EphB3 with the highest avidity [16] . In addition , both KSHV and RRV require the presence of gH as well as gL in the gH/gL complex for Eph-interaction . While we recently have shown that the binding site for the KSHV gH/gL complex on EphA2 is similar to that of natural ephrin ligands [20] , the corresponding interaction site on the gH/gL complex has remained elusive until now . Recent structure-function analyses of other herpesviruses suggested different domains of the herpesviral gH as determinants for entry into target cells . For instance , the N-terminal tip of domain ( D ) I of the varicella-zoster virus ( VZV ) homolog was shown to play a role in virus entry and fusion , as well as VZV skin tropism [21] . Similarly , in Epstein-Barr-Virus ( EBV ) infection , DI of gH was described as a determinant of membrane fusion activity and gB interaction [22] . Additionally , the monoclonal anti-gH/gL antibody E1D1 which inhibits EBV membrane fusion with epithelial cells was shown to bind to the tip of the gH/gL DI through interaction with gL residues [23] . Other studies suggested a role of residues in DII of EBV gH in gB-mediated membrane fusion , which is mediated by an integrin-binding ‘KGD’ motif located in the central region of the gH/gL complex [24] . Therefore , the site of the Eph interaction cannot easily be inferred from similar receptor interactions by other viruses . A detailed analysis of the evolutionarily conserved interaction with Eph family receptors and the regions on the gH/gL complex involved in this interaction would further our general understanding of the herpesviral gH/gL glycoprotein complex . The actual contribution of the Eph receptor interaction to infection of different cell types by KSHV and RRV also deserves further analysis . Inhibition of KSHV infection by blocking of the Eph interaction ranged from almost complete to around twofold in previous studies depending on experimental setup and cell type [16 , 18] . These findings raised the question whether the interaction with Eph family receptors by KSHV and RRV is obligatory , obligatory only on certain cell types , or simply has a very strong enhancing effect on infection . Such a strong enhancing effect , depending on the setting , may still make this interaction obligatory to achieve detectable infection . Studying entry of KSHV and RRV is complicated by the fact that the Eph family comprises 14 homologous members in both humans and rhesus monkeys , by the complexity of the Eph-ephrin signaling network , and by the physiological importance of this RTK family . Various members of the Eph family were shown to play prominent roles in a wide range of physiological and pathological events , including the regulation of developmental processes , angiogenesis , cancer , and inflammation ( reviewed in [25–27] ) . Blocking or ablating expression of Ephs may have strong effects on some cell types , exemplified by the dependence of several tumor cell lines on EphA2 expression for ongoing growth in vitro [28 , 29] . Furthermore , RRV binds very promiscuously to many of the 14 A- and B-type Ephs , impeding clean knockout experiments of distinct Ephs by compensatory effects of other members of the family . Even for KSHV , which binds EphA2 very selectively with high affinity , weak interactions with other A-type Ephs were detectable [16] and may confound results . Targeting the viral side of the gH/gL-Eph interaction would overcome this limitation . Therefore , the aim of this study was first to address the question of whether the conserved interaction with Ephs is based on an equally conserved viral binding motif on gH/gL and to map this interaction site on the gH/gL complex of KSHV and RRV . Second , we sought to use this information to generate mutant KSHV and RRV strains that are unable to interact with Eph family receptors and to evaluate the relative importance of this interaction for infection of different cell types .
Based on the known structures of the herpes simplex virus type 2 and EBV gH/gL complex [22 , 30] and the conserved nature of this glycoprotein complex , it can be assumed that gH of KSHV comprises domain I ( D I ) to domain IV ( D IV ) , a transmembrane domain ( TM ) and a short C-terminal intravirion domain ( IVD ) . To identify the region of gH/gL critical for Eph receptor interaction , we constructed chimeras composed of different regions derived from either the KSHV or RRV gH primary sequence . Transfected chimeric gH constructs were co-expressed with KSHV gL to form stable gH/gL complexes . Using co-immunoprecipitation assays , we tested the complexes of KSHV/RRV chimeric gH and KSHV gL for their ability to bind EphA2 , which is the high affinity Eph family receptor for KSHV gH/gL and exhibits only marginal or no interaction with RRV gH/gL ( Fig 1A ) . Out of the seven tested constructs , three were detected in Western blot analysis ( Fig 1A , lane 2 , 4 , 8 ) . Due to the relatively high sequence diversity of rhadinovirus gH proteins , molecular mass and glycosylation vary slightly between KSHV and RRV derived regions leading to visible shifts in apparent migration of gH chimeras when compared to KSHV gH . All of the chimeras that were expressed , even if only comprising the N-terminal and so-called shoulder region of KSHV gH ( Fig 1A , lane 8 ) , were found to complex with KSHV gL and to bind EphA2 . In contrast , while full-length RRV gH did form a complex with KSHV gL , this complex did not interact with EphA2 ( Fig 1A , lane 9 ) . Likewise , N-terminal KSHV/RRV chimeras of gL or full-length RRV gL did not support binding to EphA2 when co-expressed with full-length KSHV gH in the gH/gL complex ( Fig 1B ) . Additionally , KSHV gH/gLΔ135–164 , consisting of full-length KSHV gH and a C-terminal gL truncation mutant , precipitated EphA2 to wild-type levels ( Fig 1C ) . This indicates that the N-terminal domains of KSHV gH and of KSHV gL in the gH/gL complex are essential for the interaction with EphA2 . The natural ligands of Ephs , the eight ephrins , interact with their respective receptors through a structurally conserved G-H loop that also exhibits substantial conservation on the amino acid level [31] . We therefore aimed to identify an equally conserved Eph interaction motif on the rhadinoviral gH . First , we performed comparative sequence analysis of the gH proteins of KSHV and the two RRV isolates 17577 and 26–95 , comparing the three rhadinoviral gH sequences to that of EBV gH ( Fig 2A ) . Interestingly , we found a highly conserved motif present in all three gH sequences listed here as well as in all RRV and KSHV sequences currently listed in the NCBI database that consists of the five amino acids Glu ( E ) -Leu ( L ) -Glu ( E ) -Phe ( F ) -Asn ( N ) ( Fig 2A , black rectangle ) and is not perfectly conserved in EBV . To investigate the relevance of this E-L-E-F-N motif for the interaction with Eph receptors , a mutational scan of the N-terminal regions of KSHV and RRV gH was performed by substituting single amino acids with alanine . First , we tested the influence of these amino acid substitutions on the stability of the gH/gL heterodimer by immunoprecipitation assays ( S1A and S1B Fig ) . The mutant KSHV and RRV gH/gL complexes were immunoprecipitated via the V5-tagged gH in the absence of recombinant Eph receptors and complexation with gL was assayed by Western blot after normalization for differences in expression levels . Mutations L47A and I49A ( KSHV ) and W64A ( RRV ) resulted in a strongly decreased interaction with gL . RRV gH E52A and L53A exhibited a reduced expression and slightly aberrant glycosylation pattern , and L53A also incorporated fully glycosylated gL less efficiently ( S1B Fig ) . The ability of mutant gH/gL complexes to interact with either myc-tagged full-length EphA2 for KSHV gH/gL ( Fig 2B ) , or myc-tagged full-length EphB3 for RRV gH/gL ( Fig 2C ) was analogously analyzed by immunoprecipitation via the V5-tagged gH and Western blot . This approach identified several point mutations that resulted in a loss of EphA2 or EphB3 interaction , respectively . Among those , the described mutations that lead to a reduced gL interaction represent one group . As loss of gL in the gH/gL complex in itself is sufficient to abolish Eph interaction ( Fig 2B and 2C , first lane ) , the reduced interaction of these gH point mutants with Eph receptors can most likely be attributed to a loss of the gL interaction . We therefore aimed to identify gH point mutants that exhibited a normal expression level and glycosylation pattern and incorporated gL to wt levels . Notably , we identified two amino acids in the conserved E-L-E-F-N motif ( Glu52 and Phe53 for KSHV , Glu54 and Phe55 for RRV ) whose side chains are essential for Eph receptor interaction but at the same time dispensable for gL binding ( Fig 2B and 2C , indicated by black lines ) . Single point mutations of other amino acids adjacent to the described E-L-E-F-N motif , specifically V51A , R59A and Y60A ( RRV ) as well as L60A and W62A ( KSHV ) also abrogated binding of Eph receptors while apparently causing only a minor decrease in gL association . This might indicate additional , direct interaction with EphB3 or EphA2 , respectively , at those positions . Combination of mutations E52A and F53A of KSHV gH completely abrogated EphA2 binding without affecting gL association or expression ( Fig 2B , second lane from the right , S1C Fig ) . Likewise , a V51A-E54A-F55A triple mutant of RRV gH was negative for EphB3 interaction , while maintaining the capacity to bind RRV gL and a normal expression level ( Fig 2C , rightmost lane , S1C Fig ) . Finally , introducing a charge reversal by mutating glutamic acid at position 52 of KSHV gH to lysine ( E52K ) also abrogated EphA2 binding ( Fig 2B , rightmost lane ) , implicating a critical role of this negative charge in the interaction . The importance of the E-L-E-F-N motif and its surrounding region for binding of KSHV and RRV gH/gL to their respective high-affinity binding partners from the Eph family was additionally supported by a structural prediction of the KSHV and RRV gH/gL complex based on the crystal structure of the Epstein-Barr-Virus gH/gL complex [22] . In this prediction , the residues of gH crucial for Eph interaction in vitro , E52/F53 ( KSHV ) and E54/F55 ( RRV ) , form the turn region residues of a putative beta-hairpin in an optimal array for protein-protein interaction ( Fig 3 ) . The spatial layout of gH and gL in this region suggests a possible stabilizing effect of the N-terminal domain of gL on the parallel beta-sheet structure . Formation of such a putative receptor-binding sub-domain of gH/gL fits with the observation that the N-terminal domain of gL is crucial for Eph interaction as well , as shown by co-immunoprecipitation of KSHV and RRV chimeras ( Fig 1B ) . Based on the results of our alanine scan and the structural predictions , we constructed mutant KSHV Bac16 and RRV-YFP 26–95 , in which amino acids E52/F53 or E54/F55 , respectively , are mutated to alanine ( termed KSHV gH-ELAAN and RRV gH-AELAAN ) ( Fig 4A and 4B ) . For RRV we additionally mutated the valine at position 51 ( V51 ) to alanine to avoid reversion as production of RRV stocks requires several rounds of lytic replication , which will select for revertants if these have a growth advantage . KSHV , on the other hand , is produced by induction of one lytic cycle after expansion of latently infected producer cells making emergence of revertants less likely . As an additional Eph-binding-deficient control we included a RRV-YFP 26–95 gL deletion mutant ( RRV ΔgL ) ( Fig 4A and 4B ) . Western Blot analysis of KSHV and RRV wt and mutant virus particles verified that KSHV gH-ELAAN ( S1D Fig ) and RRV gH-AELAAN ( S1E Fig ) incorporate gH to wt levels and confirmed the complete deletion of gL on protein level for RRV ΔgL and the efficient incorporation of gL by RRV gH-AELAAN ( S1E Fig ) . All gH mutants that were introduced into KSHV and RRV also reached expression levels comparable to wt when expressed alone or together with gL in transfected cells ( S1C Fig ) . Analysis of KSHV gL in virus particles was precluded by our inability to generate antibodies to KSHV gL despite several attempts . All mutants were viable and infectious in vitro as assayed by the expression of green fluorescent protein ( GFP ) for KSHV or yellow fluorescent protein ( YFP ) for RRV under the control of a constitutively active promoter ( Fig 4C ) . To evaluate the receptor usage of the described KSHV and RRV mutants we conducted blocking experiments by either pre-incubation of viral inocula with soluble Eph decoy receptor fused to the Fc part of IgG ( EphA2-Fc/EphB3-Fc ) ( Fig 5A and 5B ) or ligand competition on target cells to block access to Eph receptors ( Fig 5C and 5D ) . For ligand-dependent blocking experiments of KSHV infection , we used recombinant ephrinA4 ( ephrinA4-Fc ) as a high affinity ligand [16 , 20] , which blocks EphA2 on target cells . As RRV was shown to interact with both A- and B-type Ephs , a mix of all described recombinant ligands of Ephs ( ephrinA1 , ephrinA2 , ephrinA3 , ephrinA4 , ephrinA5 , ephrinB1 , ephrinB2 and ephrinB3 , each fused to Fc at the end of the extracellular part of the protein ) was used in ligand competition experiments . For both viruses , wt and Eph-binding-negative mutants were titrated and normalized to achieve comparable infections in the absence of inhibitor . In blocking experiments targeting viral particles , pre-incubation of the virus with soluble Fc alone as a control did not appreciably influence KSHV or RRV infection while soluble Eph decoy receptors led to a dose-dependent inhibition of KSHV infection of up to 90% ( Fig 5A ) and RRV infection of approximately 80% ( Fig 5B ) on SLK cells , as described before [15 , 16] . Contrarily , we observed no influence of saturating concentrations of EphA2-Fc/EphB3-Fc on the infection with Eph-binding-negative mutants KSHV gH-ELAAN , RRV gH-AELAAN or RRV ΔgL when compared to infection with untreated or soluble Fc treated viral inocula ( Fig 5A and 5B ) . Correspondingly , soluble ephrinA4-Fc significantly reduced KSHV infection of SLK , human umbilical vein endothelial cells ( HUVEC ) , lymphatic endothelial cells ( LEC ) , and human foreskin fibroblasts ( HFF ) in a range from approximately 35% to 70% depending on the cell type ( Fig 5C ) . Ligand competition using the recombinant ephrin-Fc mix resulted in similar , significant blocking of RRV infection of SLK , HUVEC , LEC , and rhesus monkey fibroblasts ( RF ) ( Fig 5D ) . Pre-treatment of target cells with soluble ephrins did not influence infection with Eph receptor-detargeted virus mutants ( Fig 5C and 5D ) , comparable to soluble decoy receptor pre-incubation . In summary , using either soluble Eph decoy receptors or recombinant ephrins as blocking agents we observed a robust inhibition of infection with wt KSHV and RRV , while infection with Eph-detargeted mutants remained unaffected . To analyze the importance of the Eph interaction for cellular attachment and infectivity of KSHV and RRV particles , we normalized infectious dose to genome copies per cell . First the capacity of wt and mutant virus to bind target cells was analyzed . A comparison of the ratio of input genome copy numbers and bound genome copy numbers at 4°C revealed no differences in the attachment of KSHV wt and KSHV gH-ELAAN ( Fig 6A ) or RRV wt and RRV gH-AELAAN and RRV ΔgL ( Fig 6B ) . We observed attachment comparable to wt of both Eph-binding-negative KSHV and RRV mutants over a range of 3 logs of input virus per cell ( Fig 6A and 6B ) . In contrast , KSHV gH-ELAAN and RRV gH-AELAAN/ΔgL exhibited a reduced specific infectivity on cells of epithelial ( S2A and S2B Fig ) and endothelial origin , as well as fibroblasts when compared to their corresponding wt virus . For both viruses , a representative experiment and averaged , fitted curves from repeat experiments are shown ( Fig 6C and 6D , upper panels ) . The effect of the introduced mutations on specific infectivity was determined by the ratio of the rate constant K of fitted curves for wt and mutant viruses . Kwt/Kmutant describes the shift of fitted curves of Eph-binding-negative viruses to the right indicating the fold increase in input virus required to achieve wt infection levels . The impairment of specific infectivity of KSHV gH-ELAAN ranged from a factor of 4 . 4 on LEC to a factor of 9 on HUVEC . Eph-binding-negative RRV mutants exhibited a reduction in specific infectivity from 5-fold for RRV gH-AELAAN on RF to approx . 20-fold for RRV gH-AELAAN and RRV ΔgL on HUVEC and LEC , respectively . Similarly , when comparing infection with an identical number of viral input genomes for wt and mutant KSHV and RRV , we observed a robust reduction in the percentage of infected cells with Eph-binding-negative viruses . For example , infection with KSHV was reduced approx . 5-fold to 10-fold on HFF and 5-fold to 25-fold on HUVEC or LEC when identical input genome numbers for KSHV wt and KSHV gH-ELAAN were compared in a range of 500 to 25000 genomes/cell ( Fig 6C , bar graphs ) . Similarly , for RRV the reduction in the percentage of infected cells ranged from approx . 2 . 5-fold to 8 . 5-fold on RF and 2 . 6-fold to 10-fold on LEC to approx . 5-fold to 82-fold on HUVEC when comparing identical input genome numbers of wt and mutant viruses in a range of 500 to 5000 genomes/cell ( Fig 6D , bar graphs ) . Results using the mean fluorescence intensity ( MFI ) of the reporter gene instead of percentage of infected cells as a readout corroborated the same conclusions ( S2C and S2D Fig ) . Notably , RRV gH-AELAAN and RRV ΔgL performed highly similar in these assays .
Eph family receptors , and specifically EphA2 , have been described as host factors for a wide range of pathogens besides KSHV and RRV [15 , 16] , including hepatitis C virus [32] , Chlamydia trachomatis [33] , Cryptococcus neoformans [34] , malaria parasites [35] , and as very recently reported EBV [36 , 37] . While it has been shown that KSHV interacts specifically with the ligand-binding domain of EphA2 in a manner that competes with the natural ephrin ligands [20] , little was known until now about the specific interaction sites or motifs on the surface proteins of the respective pathogens . In this study , we present evidence for a conserved , distinctive binding site on DI of the gH/gL complex of the rhadinoviruses KSHV and RRV that is crucial for interaction with members of the Eph family of receptor tyrosine kinases . Using a combination of comparative sequence and structural analysis together with in vitro mutational screens , we were able to map the Eph interaction site to the central amino acids of a conserved five amino acid motif Glu ( E ) -Leu ( L ) -Glu ( E ) -Phe ( F ) -Asn ( N ) on KSHV and RRV gH . Even though the amino acid sequence of gH is relatively variable within the herpesvirus family , and even so between KSHV and the RRV isolates 26–95 and 17577 ( Fig 2A ) as well as within a large number of RRV gH sequences isolated by Shin et al [38] , the described E-L-E-F-N motif is strictly conserved in all KSHV and RRV gH sequences . Direct evidence for the functional role of the conserved rhadinoviral Eph receptor interaction motif , and in particular of residues E52/F53 ( KSHV ) and E54/F55 ( RRV ) of gH , for Eph targeting was provided by the construction of KSHV gH-ELAAN and RRV gH-AELAAN virus mutants and subsequent inhibition experiments using either soluble decoy receptors ( Fig 5A and 5B ) or soluble ephrins ( Fig 5C and 5D ) . Interestingly , the conserved asparagine of the E-L-E-F-N motif is also a conserved N-glycosylation site , and its mutation leads to a visible shift in molecular weight ( Fig 2B and 2C ) . While the asparagine itself does not seem to contribute to Eph binding , glycosylation could potentially play a role in steric shielding of this region from antibodies , as described for example for human immunodeficiency virus type 1 ( HIV-1 ) gp120 [39] , Ebola Virus ( EBV ) glycoprotein [40] and influenza virus hemagglutinin [41] , as well as from MHC presentation [42] . High conservation on the amino acid sequence level also translates into an equally conserved structural prediction of the gH/gL complex of KSHV and RRV 26–95 ( Fig 3 ) , when modeled using the crystal structure of the EBV gH/gL complex [22] . In these computational models the amino acid residues E52/F53 of KSHV gH and E54/F55 of RRV gH that are crucial for Eph interaction are located in a predicted outward-angled beta-hairpin . Several reports indicate that the interaction of both A- and B-type ephrins with different Eph receptors is structurally conserved and mediated by insertion of the so-called G-H loop of ephrins into a conserved hydrophobic groove on the Eph receptors [31 , 43–45] . KSHV gH/gL interacts with the ephrin binding region of EphA2 , which suggests that binding of rhadinoviral gH/gL complexes to Ephs may occur in a fashion that structurally mimics the binding of ephrins to their receptors . The importance of this structural motif is further supported by the effect of gH point mutants R59A , Y60A ( RRV ) and L60A , W62A ( KSHV ) on binding of Eph receptors but not on gH/gL complexation . In our model , these residues are located in the base region of the anti-parallel beta-sheets that form the beta-hairpin . Disruption of the base region may lead to a destabilization of the beta-hairpin structure and therefore loss of Eph receptor interaction on gH while binding to gL is maintained . Alternatively , these residues might directly contact Eph receptors and help determine the specificity of the interaction for e . g . EphA2 or EphB3 . According to our model , the putative beta-hairpin makes considerable contact with the N-terminal region of gL , suggesting that proper folding of the Eph interacting sub-domain may be gL-dependent . The importance of the N-terminal region of gL is also supported by our co-expression and immunoprecipitation experiments , in which the N-terminal regions of both gH and gL were necessary for Eph binding ( Fig 1A and 1B ) . Furthermore , mutation of several amino acids adjacent to the E-L-E-F-N motif which led to a reduced gL interaction also led to reduced binding to EphA2 or EphB3 , respectively , most likely due to an instability or a lack of gL in the gH/gL complex ( Fig 2B and 2C ) . Our results are in good agreement with several reports on the functional importance of the N-terminal domain of the gH/gL complex for the entry process of EBV and VZV [21–23 , 46 , 47] . Our approach , however , does not exclude the existence of additional Eph interaction sites on gH/gL , which might determine the specificity of KSHV and RRV for individual Eph receptors . While this manuscript was in revision , EBV was reported to also interact with EphA2 not only through gH/gL , but additionally also through gB [36] . Whether the mechanism by which EBV and KSHV interact with EphA2 is conserved remains to be determined . Likely , differences do exist as EBV fuses directly with the membrane of epithelial cells , whereas KSHV enters through endocytosis . To validate our model and to fully elucidate structural aspects of the role of the rhadinoviral gH DI , as well as of potential additional interaction sites , for receptor binding and entry into target cells , crystal structures of rhadinoviral gH/gL complexes are needed , preferably in complex with the respective high affinity receptor from the Eph family . Surprisingly , the deletion of orf47 , which encodes gL , from the RRV genome did not abrogate the production of infectious virus particles . Similar attempts at deleting gL of KSHV were so far unsuccessful as recombinant bacmids harboring the corresponding deletion did not yield infectious virus . One explanation for this might be the function of recently described spliced genes enclosing the orf47-orf46-orf45 locus during reactivation from latency [48] . RRV , however , replicates lytically on RFs and rarely enters latency , which could mask defects in reactivation and make potential homologous RRV transcripts dispensable for RRV growth in culture . Accordingly , in our experiments RRV ΔgL performed practically indistinguishable from RRV gH-AELAAN . Interestingly , an essential role for infectivity was described for gL of herpes simplex virus type 1 [49] and both human and rhesus cytomegalovirus [50] , whereas gL was described as non-essential for infectivity and cell-to-cell spread of the murine gamma-herpesvirus MHV-68 [51] . Similarly , pseudorabiesvirus gL null mutants remain infectious , although exhibiting impaired entry and cell-to-cell spread [52–54] . Thus , whether gL is essential for infectivity in vitro or not seems to vary within the herpesviruses . The question of the quantitative contribution of Eph receptors to rhadinovirus infection could not be answered satisfactorily until now , due to the intrinsic limitations of blocking assays , such as concentration and amount of blocking agent , and possible confounders of knockout experiments , such as potential use of alternative Eph receptors or detrimental effect of the knockout on the cell in general . The dramatic inhibition of KSHV and RRV wt infection demonstrated by specifically targeting the Eph interaction or after EphA2 knockdown or knockout [15 , 16 , 18 , 55] is similar in extent to what we observed using a soluble decoy receptor block ( Fig 5A and 5B ) . This is also in very good agreement with our experiments analyzing the specific infectivity of our mutants where the percentage of infected cells with Eph-binding-deficient mutant virus was reduced 2 . 5-fold to 82-fold when compared to wt virus using equal input genome numbers . Similarly , approx . 4 to 20 times more Eph-binding-deficient mutant virus was needed to achieve infection levels comparable to those of wt virus ( Fig 6C and 6D ) . Our results now quite unequivocally demonstrate that the interaction of gH/gL with Ephs is not essential for KSHV and RRV infection but contributes significantly to infectivity . The interaction of gH with different viral or cellular proteins is thought to be a determinant of cell tropism in a wide range of herpesviruses [56–58 , 21] . In previous studies , we observed differences in the amount of inhibition of KSHV and RRV infection achieved by blocking the Eph interaction [15 , 16] depending on the cell type . While for KSHV , the cell type-specific defect in infectivity of KSHV gH-ELAAN varied significantly between different batches of fibroblasts and endothelial cells and did not allow for a clear conclusion , for RRV the cell type-specific differences were consistently observed . On LEC and HUVEC , both RRV gH-AELAAN and RRV ΔgL exhibited a defect in infectivity that was more pronounced than that observed on primary RF . These results confirm our previous findings that Eph receptors play a minor role in the infection of fibroblasts compared to infection of endothelial cells by RRV [16] , as also exemplified by the ability of RRV gH-AELAAN and RRV ΔgL to replicate to high titers on RF . The observation that ablation of the Eph interaction does not fully abrogate infectivity and has cell type-specific effects for RRV suggests the contribution of yet to be identified additional host factors for RRV and potentially also KSHV; for RRV in particular as integrins do not seem to play a major role for RRV [14] . This becomes most apparent for the infection of lympathic endothelial cells with Eph-binding-deficient RRV mutants . Despite increasing amounts of input virus the infection appeared to be approaching a plateau at around 50% infected cells ( Figs 6D and 7A ) , which is in good agreement with previous observations where RRV failed to infect a sizable subpopulation of endothelial cells despite increasing amounts of input virus when the Eph interaction was blocked [16] . In our opinion , the most likely explanation would be that the cell population that is refractory to Eph-independent infection lacks a host factor that can functionally substitute for the Ephs . The critical sub-domain on the gH/gL glycoprotein complex identified in our study may serve as a target for inhibition by monoclonal antibodies . We identified several critical amino acid residues that most likely mediate direct interaction with Eph receptors . Even if the gH/gL-Eph interaction is not strictly essential , strong inhibition of KSHV infection can be achieved by targeting this region . Blocking this highly conserved region with an antibody might afford inhibition similar to that achieved by soluble EphA2-Fc decoy receptor ( Fig 5A ) , which presumably binds in a manner that should be very similar to antibodies targeting the E-L-E-F-N motif . The Eph-ephrin system is a complicated signaling network that may play a role beyond simple receptor interaction in the viral infection and that interacts with a surprising number of pathogens . With the construction of Eph-binding-negative rhadinovirus mutants described in this paper we not only present direct evidence for a conserved Eph interaction motif of KSHV and RRV , but also provide a useful toolkit for the future analysis of EphA2-specific signaling in the case of KSHV and signaling of a broader range of both A-type and B-type Ephs for RRV .
A549 [59] ( laboratory of Stefan Pöhlmann , German Primate Center—Leibniz Institute for Primate Research , Göttingen , Germany ) , Human embryonic kidney ( HEK ) 293T cells [60 , 61] ( laboratory of Stefan Pöhlmann ) , human foreskin fibroblasts ( HFF ) ( laboratory of Klaus Korn , Universitätsklinikum Erlangen , Institute for Clinical and Molecular Virology , Erlangen , Germany ) , SLK cells [62 , 63] ( NIH AIDS Research and Reference Reagent program ) and rhesus monkey fibroblasts ( RF ) ( laboratory of Prof . Rüdiger Beer , German Primate Center—Leibniz Institute for Primate Research , Göttingen , Germany ) were cultured in Dulbecco’s Modified Eagle Medium ( DMEM ) , high glucose , GlutaMAX , 25mM HEPES ( Thermo Fisher Scientific ) supplemented with 10% fetal calf serum ( FCS ) ( Thermo Fisher Scientific ) , and 50μg/ml gentamycin ( PAN Biotech ) . iSLK cells [64] ( laboratory of Don Ganem , Novartis Institutes for BioMedical Research , Emeryville , CA , USA ) were maintained in DMEM supplemented with 10% FCS , 50μg/ml gentamycin , 2 . 5μg/ml puromycin ( InvivoGen ) and 250μg/ml G418 ( Carl Roth ) . Human vascular endothelial cells ( HUVEC ) ( PromoCell ) were maintained in standard Endothelial Cell Growth Medium 2 ( PromoCell ) . Human lymphatic endothelial cells ( LEC ) from juvenile donors ( a kind gift from Anja Boos , Universitätsklinikum Erlangen , Department of Plastic and Hand Surgery , Erlangen , Germany ) were maintained in Endothelial Cell Growth Medium MV 2 ( PromoCell ) . Eph-interaction-negative KSHV ( KSHV gH-ELAAN ) and RRV ( RRV gH-AELAAN , RRV ΔgL ) recombinants were generated using a two-step , markerless λ-red-mediated BAC recombination strategy as described by Tischer et al . [65] . KSHV gH-ELAAN and RRV gH-AELAAN harbor amino acid substitutions E52A and F53A ( KSHV ) or V51A , E54A and F55A respectively . The deletion in RRV ΔgL encompasses 128 nucleotides from position 78 to position 205 ( amino acids 27 through 68 ) , resulting in a frameshift after amino acid 26 and a stop codon after 37 amino acids . Deletion in this region was chosen in order not to destroy known and potential overlapping genes that may so far not have been charted , and because we identified sequences reminiscent of regulatory elements in the region directly surrounding the start codon . Bacmid clones BAC16 ( KSHV ) [66] and BAC35-8 ( RRV ) were used , respectively . In short , recombination cassettes were generated from the pEPKanS template by polymerase chain reaction ( PCR ) with Phusion High Fidelity DNA polymerase ( Thermo Fisher Scientific ) using long oligonucleotides ( Ultramers; purchased from Integrated DNA Technologies ( IDT ) ) ( see S1 Table for a complete list of primers ) . Recombination cassettes were transformed into BAC16-carrying Escherichia coli strain GS1783 or RRV-YFP-carrying GS1783 respectively , followed by kanamycin selection , and subsequent second recombination under 1% L ( + ) arabinose ( Sigma-Aldrich ) -induced I-SceI expression . Colonies were verified by PCR of the mutated region followed by sequence analysis ( Macrogen ) , pulsed-field gel electrophoresis and restriction fragment length polymorphism . For this purpose , bacmid DNA was isolated by standard alkaline lysis from 5ml liquid cultures . Subsequently , the integrity of bacmid DNA was analyzed by digestion with restriction enzyme XhoI and separation in 0 . 8% PFGE agarose ( Bio-Rad ) gels and 0 . 5×TBE buffer by pulsed-field gel electrophoresis at 6 V/cm , 120-degree field angle , switch time linearly ramped from 1s to 5s over 16 h ( CHEF DR III , Bio-Rad ) . Infectious KSHV recombinants were generated by transfection of purified bacmid DNA ( NucleoBond Xtra Midi ( Macherey-Nagel ) ) into iSLK cells using GenJet Ver . II ( Signagen ) according to manufacturer’s instructions . After visible GFP expression , selection was performed using 200μg/ml hygromycin B ( InvivoGen ) until only GFP positive cells remained . Lytic replication of KSHV-BAC16 was induced in DMEM supplemented with 10% fetal calf serum ( FCS ) and 50μg/ml gentamycin by addition of 2 . 5mM sodium-butyrate and 1μg/ml doxycycline . Supernatant was harvested after the cell monolayer was destroyed . For RRV , infectious recombinants were generated by transfection of purified bacmid DNA ( NucleoBond Xtra Midi ) into 293T cells with GenJet Ver . II ( Signagen ) according to manufacturer’s instructions . After 2 days , BAC-transfected 293T cells were transferred onto a confluent rhesus monkey fibroblasts monolayer and co-cultivated until a visible cytopathic effect ( CPE ) was observed . Virus stocks were prepared by inoculating fresh primary rhesus monkey fibroblasts with virus containing supernatant of 293T/rhesus monkey fibroblast co-cultures at a very low multiplicity of infection ( MOI; about 1 infected cell in 1000 cells ) and letting the virus replicate until the cell monolayer was destroyed . Virus-containing cell supernatant from iSLKs and rhesus monkey fibroblasts was clarified by centrifugation ( 4750g , 10 minutes ) , concentrated by overnight centrifugation ( 4200rpm , 4°C ) and careful aspiration of approximately 95% of the supernatant . The pellet was resuspended overnight in the remaining liquid . Stocks of wt and recombinant viruses were aliquoted and stored at -80°C . The integrity of the L-DNA part of virus recombinants was confirmed by Illumina-based next-generation sequencing . For isolation of viral DNA , concentrated virus stocks were incubated with DNAseI ( 40 units/ml ) for 1h at 37°C . Addition of EDTA to a final concentration of 15mM was followed by a second incubation step ( 70°C , 10min ) . After a third incubation step with ProteinaseK ( 1mg/ml ) and SDS ( 0 . 5% ) ( 60°C , 2h ) standard phenol chloroform extraction was performed . Sample preparation was performed using the Nextera DNA Sample Preparation system , dual indexing , and sequencing using the MiSeq Reagent Kit , 600 Cycles on the Illumina MiSeq system . Demultiplexed paired 300 nt sequence reads were analyzed by Genomics Workbench 10 ( Qiagen Bioinformatics , Aarhus , DK ) . pcDNA4 vectors containing full-length EphA2 ( ref|NM_004431| , pcDNA4-EphA2-myc ) , the soluble ectodomain of EphA2 ( amino acids 1–534 ) ( ref|NM_004431| , pcDNA4-EphA2-HA ) , a soluble EphA2-Fc fusion construct comprising amino acids 25–534 of human EphA2 in the pAB61 Fc-fusion backbone vector ( pEphA2-Fc ) [15] and EphB3 ( ref|BC052968| , pcDNA-EphB3-myc ) [15 , 16] were described elsewhere . KSHV/RRV chimeric gH constructs were generated based on pcDNA6aV5His/pcDNA3 . 1 backbone vectors containing RRV/KSHV gH and gL coding sequences ( ref|GQ994935 . 1| , pcDNA6aV5-KSHV-gH , pcDNA3 . 1-KSHV-gL-Flag; ref|AF210726 . 1| , pcDNA6aV5-RRV-gH , pcDNA3 . 1-RRV-gL-Flag ) [15 , 16] by PCR based restriction enzyme cloning . The KSHV gLΔ135-164-Flag construct ( pcDNA3 . 1-KSHV-gLΔ135-164-Flag ) was generated by a PCR-based mutagenesis using phosphorylated primers followed by blunt end ligation of the PCR product ( see S1 Table for a complete list of primers and constructs ) . Plasmids harboring point mutations in domain I of RRV/KSHV gH used in the alanine scan were purchased from GenScript based on pcDNA6aV5-KSHV-gH and pcDNA6aV5-RRV-gH . Recombinant , soluble EphA2 Fc/Strep-fusion protein was purified under native conditions by Strep-Tactin chromatography from 293T cell culture supernatant . 293T cells were transfected by Polyethylenimine "Max" ( PEI ) ( Polysciences ) [67] transfection with pEphA2-Fc . The protein-containing cell culture supernatant was filtered through 0 . 22μm PES membranes ( Millipore ) and passed over 0 . 5ml of a Strep-Tactin Superflow ( IBA Lifesciences ) matrix in a gravity flow Omniprep column ( BioRad ) . Bound protein was washed with approximately 50ml phosphate buffered saline pH 7 . 4 ( PBS ) and eluted in 1ml fractions with 3mM desthiobiotin ( Sigma-Aldrich ) in PBS . Protein-containing fractions were pooled and buffer exchange to PBS via VivaSpin columns ( Sartorius ) was performed . Protein concentration was determined by absorbance at 280nm . Aliquots were frozen and stored at −80°C . Recombinant , human , soluble EphB3-Fc ( 5667-B3-050 ) and soluble ephrin ligands , as either human ( rh ) or mouse ( rm ) Fc-fusion proteins ( rm-ephrinA1 , rm-ephrinA2 , rh-ephrinA3 , rh-ephrinA4 , rh-ephrinA5 , rm-ephrinB1 , rm-ephrinB2 and rh-ephrinB3 Fc ) ( SMPK3 ) were purchased from R&D Systems . Concentrated virus samples were treated with DNAseI ( 0 . 1 units/μl ) to remove any non-encapsidated DNA ( 37°C , overnight ) . Subsequently , DNAseI was deactivated and viral capsids were disrupted by heating the samples to 95°C for 30 minutes . Realtime-PCR ( qPCR ) was performed on a StepOne Plus cycler ( Thermo Fisher Scientific ) in 20μl reactions using the SensiFAST Probe Hi-ROX Kit ( Bioline ) ( cycling conditions: 3min initial denaturation at 95°C , 40 cycles 95°C for 10s and 60°C for 35s ) . All primer-probe sets were purchased from IDT as complete PrimeTime qPCR Assays ( primer:probe ratio = 4:1 ) . Samples were analyzed in technical triplicates . A series of six 10-fold dilutions of bacmid DNA was used as standard for absolute quantification of viral genome copies based on qPCR of ORF59 for KSHV and ORF73 for RRV ( see S1 Table for a complete list of primers ) . For virus attachment assays LEC were plated at 50 000 cells/cm2 . Target cells were incubated with ice-cold virus dilutions at the indicated concentrations , normalized to genomes per cell , at 4°C for 30min . After three washes with ice-cold PBS genomic DNA was isolated using the ISOLATE II Genomic DNA Kit ( Bioline ) according to manufacturer’s instructions . The ratio of viral DNA to cellular DNA as a measurement of attached virus was determined by qPCR as described above . Relative values of bound viral genomes to cellular DNA were calculated on the basis of ΔCt values for viral genomic loci ( ORF59 for KSHV , ORF73 for RRV ) and a cellular genomic locus ( CCR5 ) . For infection assays cells were plated at 50 000 cells/cm2 ( SLK , HUVEC , LEC ) or 25 000 cells/cm2 ( RF , HFF ) respectively . One day after plating , cells were infected with the indicated amounts of virus . 24h or 48h post infection cells were harvested by brief trypsinization , followed by addition of 5% FCS in PBS to inhibit trypsin activity , spun down ( 1200rpm , 10min ) , washed once with PBS , re-pelleted and fixed in PBS supplemented with 2% formaldehyde ( Carl Roth ) . A minimum of 10 000 cells was analyzed per sample for GFP or YFP expression on a LSRII flow cytometer ( BD Biosciences ) . Data was analyzed using Flowing Software ( Version 2 . 5 ) . For block with soluble ephrins , cells were pre-incubated with the indicated ephrin-Fc fusion proteins at a final concentration of 2μg/ml for 30min at room temperature followed by infection with KSHV or RRV . Block of KSHV/RRV infection with soluble decoy receptor was assayed by infection with virus inocula that were pre-incubated with the indicated concentrations of soluble EphA2-Fc , EphB3-Fc or Fc alone at room temperature for 30min . 293T cells were transfected using PEI [67] or Lipofectamine with Plus reagent ( Thermo Fisher Scientific ) as per the manufacturer’s instructions . Recombinant gH-V5/gL-Flag complexes were precipitated from the lysates of 293T cells transfected with the respective expression constructs . Lysates were prepared in NP40 lysis buffer ( 1% Nonidet P40 Substitute ( Sigma-Aldrich ) , 150mM NaCl ( Sigma-Aldrich ) , 50mM HEPES ( VWR ) , 1mM EDTA ( Amresco ) with freshly added Protease Inhibitor Cocktail , General Use ( Amresco ) ) and subsequently incubated with agitation with 0 . 5μg V5-tag antibody ( Serotec or Bio-Rad ) and ProteinG sepharose ( GenScript or GE Healthcare ) for 2h or overnight at 4°C . Amount of input gH/gL between mutants was normalized by diluting lysates with cell lysate from non-transfected 293T cells prior to immunoprecipitation according to Western blot analysis of lysates and evaluation of the gH/gL content for each construct . After one wash , pre-coupled complexes were incubated overnight at 4°C with agitation with equal amounts of lysate of full-length EphA2-myc or full-length EphB3-myc expression plasmid transfected 293T cells ( for KSHV gH/gL or RRV gH/gL binding analysis , respectively ) or with supernatant of 293T cells transfected with an expression plasmid for HA-tagged EphA2 ectodomain ( for KSHV gH/gL binding analysis ) . ProteinG beads were collected by brief centrifugation and washed 3 times in NP40 lysis buffer . Precipitates were heated in SDS sample buffer and analyzed by polyacrylamide gel electrophoresis ( PAGE ) using 8–16% tris-glycine gradient gels ( Thermo Fisher Scientific ) and Western blot ( 100mA , max 30V , 1h in Towbin buffer ( 25mM Tris , 192mM glycine ) ) . For Western blot analysis of virus particles , a 5% OptiPrep ( Sigma-Aldrich ) in PBS cushion was overlayed with 1ml concentrated virus stock and centrifuged for 2h at 20 000g . Approx . 95% of the supernatant was discarded and the virus pellet was washed once with 1ml PBS and spun down ( 20 000g , 1h ) . The virus pellet was resuspended in 30μl PBS , dissolved overnight and subsequently heated after the addition of 50μl SDS sample buffer ( 99°C , 15min ) . Western Blot analysis was performed as described above . Homology based structure prediction was performed using the Iterative Threading ASSembly Refinement ( I-TASSER ) server on standard settings for structure prediction of KSHV or RRV gH based on the crystal structure of the EBV gH/gL complex ( 3PHF ) . Modeling of the KSHV or RRV gH/gL complexes was additionally performed using both the SPRING and CO-THreader algorithms for protein-protein complex structure and multi-chain protein threading with no differences between determined structures . Resulting CO-THreader and I-TASSER structures were aligned with the VMD 1 . 9 . 3 OpenGL RMSD Trajectory Tool based on the overlapping region of gH domain I predicted in both models ( KSHV amino acids 43 to 87 , RRV amino acids 45 to 88 ) at an RMSD of 0 . 589Å for KSHV and 0 . 616Å for RRV . All further analyses and visualizations were performed using VMD 1 . 9 . 3 OpenGL . Curve fitting of specific infectivity normalized to genome copies/cell was performed using the built-in exponential equation for one phase association of GraphPad Prism version 6 for Windows ( GraphPad Software , La Jolla California USA ) based on the poisson distribution [68] . The span was set from 0 to 100 , representing 0% or 100% infected cells , respectively , resulting in the simplified function f ( x ) = 100* ( 1-e-K*x ) , with x representing input genome number and K representing the specific infectivity per input genome . The ratio between Kwt and Kmutant was used to calculate the differences in infectivity between wt and mutant viruses . Statistical difference between fitted curves was determined by the extra sum-of-squares F test with confidence intervals corrected for multiple comparisons using the Bonferroni correction . Statistical analysis of multiple groups was performed using regular two-way analysis of variance ( ANOVA ) followed by Sidak’s multiple comparison test . Statistical difference between two groups was determined by unpaired Student’s t-tests followed by correction for multiple comparison using the Holm-Sidak method when necessary . All Statistical analyses were performed with GraphPad Prism version 6 . For all statistics , *: p-value < 0 . 05 , **: p-value < 0 . 01 , ***: p-value < 0 . 001 . See S1 Table for a complete list of primers . See S1 Table for a complete list of antibodies . | In immunocompromised individuals in general and in the context of HIV infection in particular , KSHV is a major cause of cancer and B-cell proliferative malignancies . We identified and mutated conserved residues in the N-terminal domain of the gH/gL glycoprotein complex of KSHV and the related monkey virus RRV that are critical for the interaction with cellular receptors from the Eph family . These findings provide important insight into the function of the γ-herpesviral entry machinery . Using recombinant KSHV and RRV carrying these mutations , we demonstrated that while not strictly essential , gH/gL-Eph interactions are important for efficient infection—for RRV also in a cell-specific manner—but not for attachment of KSHV and RRV . The Eph-detargeted virus mutants described in this study can be used to further dissect the requirements for KSHV and RRV entry and to identify potential alternative entry mediators . Domains and residues on the viral glycoproteins with critical roles in receptor recognition , such as the Eph-binding motif described in this paper , can be informative for the design of inhibitory monoclonal antibodies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"crystal",
"structure",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"endothelial",
"cells",
"condensed",
"matter",
"physics",
"microbiology",
"fibroblasts",
"epithelial",
"cells",
"viruses",
"mutation",
"dna",
"vir... | 2018 | A conserved Eph family receptor-binding motif on the gH/gL complex of Kaposi’s sarcoma-associated herpesvirus and rhesus monkey rhadinovirus |
Proteins interact in complex protein–protein interaction ( PPI ) networks whose topological properties—such as scale-free topology , hierarchical modularity , and dissortativity—have suggested models of network evolution . Currently preferred models invoke preferential attachment or gene duplication and divergence to produce networks whose topology matches that observed for real PPIs , thus supporting these as likely models for network evolution . Here , we show that the interaction density and homodimeric frequency are highly protein age–dependent in real PPI networks in a manner which does not agree with these canonical models . In light of these results , we propose an alternative stochastic model , which adds each protein sequentially to a growing network in a manner analogous to protein crystal growth ( CG ) in solution . The key ideas are ( 1 ) interaction probability increases with availability of unoccupied interaction surface , thus following an anti-preferential attachment rule , ( 2 ) as a network grows , highly connected sub-networks emerge into protein modules or complexes , and ( 3 ) once a new protein is committed to a module , further connections tend to be localized within that module . The CG model produces PPI networks consistent in both topology and age distributions with real PPI networks and is well supported by the spatial arrangement of protein complexes of known 3-D structure , suggesting a plausible physical mechanism for network evolution .
Life is highly organized at all levels of molecules , cells , tissues , and organisms , and such relationships among biological entities are often represented as networks , with vertices representing e . g . genes or proteins , and edges representing e . g . physical protein interactions , transcriptional regulation , or metabolic reactions . The topology of biological networks shows many interesting characteristics , such as scale-free topology ( power-law or broad degree distribution ) and hierarchical modularity ( reviewed in [1] ) . These properties are believed to be the basis of functional modularity , error-tolerance , and stability [2]–[5] characteristic of many biological networks . One important question is thus how these important network architectures originate , and what driving forces underlie the observed networks . It has not been clear whether network architecture results from the mosaic sum of each gene or protein's inherent properties , such as stickiness or interactive promiscuity [6] , [7] , or from a stochastic mechanism underlying network evolution , in which the trajectory of network evolution is conditioned on the previous state of the network [8] . This problem has been of wide interest because it raises fundamental questions about design principles of molecular networks and the role of natural selection in the evolution of network structure [9] . Initially , Barabási and Albert proposed a preferential attachment rule as a general mechanism to generate scale-free networks [8] . In this model , a newly introduced node is more likely to be attached to highly connected nodes , resulting in a power-law degree distribution . In a network of protein-protein interactions ( PPI ) , gene duplication and divergence ( DD ) is most popularly thought of as the origin of the scale-free topology of protein interaction networks [10]–[15] . In the DD model , the degree of a node increases mainly by having duplicate genes as its neighbors . Therefore , the preferential attachment rule is achieved implicitly , with highly connected nodes having more chance to have duplicate genes as their neighbors [1] . The DD model is also shown to generate hierarchically modular networks under certain conditions [16] . Although the DD model generates scale-free and modular networks , it has drawbacks that must be noted if it is to be considered a main mechanism for PPI network evolution . Primarily , only a small fraction of duplicate genes effectively contribute to the overall network topology . The key feature of the DD model originates from the fact that duplicate genes share a certain number of interaction partners . However , the interaction patterns of duplicate genes diverge rapidly [17] , and the vast majority of gene duplicates are shown to share no interaction partners [18]–[20] . Some duplicates , in fact , may have diverged so extensively that they can no longer be detected by sequence homology . These distant duplicates would share even fewer interaction partners , and thus they are essentially indistinguishable from non-duplicate pairs in terms of interaction patterns . To better understand the evolution of PPI networks , we analyzed a non-topological property—the age of each protein as estimated based upon the taxonomic distribution of its constituent domains [21] , [22]—and observe that yeast PPI networks show a unique interaction density pattern between different protein age groups . The density pattern of the yeast PPI network was compared with those generated by canonical network evolution models—preferential attachment ( the Barabási-Albert model ) , duplication-divergence ( DD ) , and anti-preferential attachment ( AP ) . Each model generates a unique interaction density pattern between the age groups; thus , the validity of the models could be effectively discriminated . Using this test , we observe that none of the canonical models are consistent with real yeast PPI networks . The age-dependent interaction density pattern nonetheless suggests growth by a stochastic process . We therefore propose an alternative model called the crystal growth ( CG ) model , which is based upon known physical and chemical principles and shows good agreement with real PPI networks in both topological and age properties as well as the 3-D subunit configurations of protein complexes .
First , we introduce the basic attachment rules of protein-protein interactions . The interaction densities , Dm , n , between two protein age groups ( m , n ) show unique patterns depending upon the attachment rule . Three basic rules are considered—random attachment ( RA ) , preferential attachment ( PA ) by Barabási and Albert [8] , [23] , and anti-preferential attachment ( AP ) . Here , we consider three protein age groups ( G1 , G2 , and G3 , from oldest to youngest ) , and assume a fixed number of new connections ( ΔE ) are made between a newly introduced node and the existing nodes as a network grows . In the RA model , a new node is randomly connected to existing nodes with equal probabilities . Initially , at time t = 1 , the first age group , G1 , makes only intra-group connections . Then a new group , G2 , is introduced and connected randomly either to G1 ( inter-group ) or within G2 ( intra-group ) . In the RA model , the expected interaction density , D , is the same between D1 , 2 and D2 , 2 . Similarly , G3 connects to G1 , G2 , and within G3 , showing the pattern of D1 , 3 = D2 , 3 = D3 , 3 . More generally , the RA model shows a pattern of Dm , n = Dm+1 , n ( m<n ) ( Figure 1A ) . In the PA mode , new proteins are preferentially connected to highly connected nodes . Thus , G2 proteins are more likely to be linked to G1 than G2 because G1 proteins have previously made connections and have a higher average degree . Likewise , G3 proteins are more likely to be connected to older groups , showing D1 , 3>D2 , 3>D3 , 3 . Thus the typical pattern of the PA model is Dm , n>Dm+1 , n ( m<n ) ( Figure 1B ) . The AP model shows an inverse pattern to the PA model , Dm , n<Dm+1 , n ( m<n ) , because new nodes prefer to connect to less-connected nodes ( Figure 1C ) . As a measure of age-dependency of interaction density , ΔD is defined as the average value of Dm+1 , n - Dm , n ( m<n ) ( see Methods ) . A positive ΔD indicates that protein interactions are more likely between similar age groups . The sign of ΔD effectively discriminates each model—it is positive in PA , negative in AP , and near zero in the RA model . We collected two independent sets of yeast PPIs - literature curated ( LC ) and high-throughput ( HTP ) PPIs , using the method of Batada et al . [23] , [24] ( Dataset S1 and Dataset S2 ) and inspected both the network topology and the age-dependency of interaction density . The number of nodes , N ( proteins ) and edges , E ( interactions ) in the LC and HTP networks are NLC = 3268 , ELC = 12058 and NHTP = 2488 , EHTP = 6766 respectively . The union ( LC+HTP ) of the two networks has 3780 nodes and 16505 edges . As HTP and LC+HTP show highly similar characteristics ( Figure S2 ) as well as the original set by Batada et al . [23] , [24] , we mainly discuss the LC data set as the yeast PPI network ( PPIyeast ) here . The recently compiled set ( Y2H-union ) by Vidal and colleagues [25] from large-scale yeast two-hybrid experiments showed the same trend ( Figure S2 ) . The PPIyeast recapitulates known topological features such as a scale-free degree distribution , hierarchical modularity , and degree-dissortative mixing property [8] , [26]–[28] , which were characterized by the various network property indices shown in the first column ( PPI ) in Figure 2 ( summarized in Table S1 ) . The probability of a node having degree k shows a scale-free or power-law degree distribution in P ( k ) ∼ k−γ plot ( the row I in Figure 2 ) . The PPIyeast is shown to be highly modular , with a high degree of clustering coefficient , C and modularity index , Q defined by Newman [29] . In particular , the PPIyeast has a scaling property in C ( k ) ∼ k−β plot ( β>0 ) , suggesting hierarchical modularity [27] ( the row II in Figure 2 ) . In a dissortative network , high-degree nodes ( hubs ) tend to connect with low-degree nodes and hub-hub interactions are suppressed , as called the Maslov-Sneppen rule [30] . The degree-dissortativity was characterized by a negative correlation in <knn> ( k ) ∼ kδ ( δ<0 ) plot ( the row III in Figure 2 ) , where <knn> ( k ) is the average degree of the nearest neighbors of the nodes with degree k . Surprisingly , the interaction density of PPIyeast is also highly age-dependent . Yeast proteins were assigned to one of the age groups ABE , AE/BE , E and F depending on the taxonomic distribution of constituent domains among archaea ( A ) , bacteria ( B ) , eukaryote ( E ) and fungi ( F ) ( see Methods , Figure S1 ) . We measured the interaction density between the age groups and observe a positive ΔD similar to AP model ( the row IV in Figure 2 ) . The pattern of positive ΔD is highly robust regardless of the sources of data ( LC , HTP and LC+HTP ) and the random addition or deletion of edges , e . g . by 50% . It suggests that the positive ΔD is a genuine feature of PPIyeast . We next simulated PPI network evolution using the three canonical models—PA ( preferential attachment ) , DD ( duplication and divergence ) , and AP ( anti-preferential attachment ) and tested compatibility with PPIyeast in terms of both topology and age-dependency . In all three models , the network starts from a small number , N0 = 4 of seed nodes and a new node is added until the total number of nodes reaches N = 3 , 000 , which is comparable to the PPIyeast ( LC ) with 3 , 268 nodes and 12 , 058 edges . In the PA and AP models , a fixed number of edges ( ΔE = 4 ) are added for each new node , which makes the final network size similar to the PPIyeast . The link probability ( P ) is proportional to the degree in the PA model ( P ∼ k ) and inversely proportional in the AP model ( P ∼ k−1 ) . For the DD model , we employ one of the simplest models by Vázquez et al . [12]: One node ( i ) is duplicated randomly , the new node ( i' ) is connected to all of the neighbors of i , and then the duplicates ( i and i' ) are linked with a small probability p . For each neighbor ( j ) of the duplicates , one of the two links ( i , j and i' , j ) is chosen randomly and deleted with the divergence probability q . Because this model may generate orphan nodes that are not connected to any other nodes , orphan nodes were removed in each duplication step . Surprisingly , none of the three models satisfied all of the characteristics of PPIyeast ( the 2nd , 3rd and 4th columns in Figure 2 for the PA , DD and AP model respectively ) . The PA and DD models generate scale-free networks and show degree-dissortativity and the DD model also shows some degree of hierarchical modularity . However , both the PA and DD models show an inverse interaction density pattern with negative ΔD . In contrast , although the AP model shows positive ΔD similar to PPIyeast , it deviates greatly in terms of topological characteristics . That is , the PPIyeast seem to show mixed characteristics , with the network topology resembling that of the DD ( PA ) model but with the interaction density similar to the AP model . Also , all three models generally show much lower levels of modularity than the PPIyeast ( the row II in Figure 2 ) . We further examined two more variants of DD models , where the divergence of edges between the duplicates is asymmetric ( DDasym ) by Ispolatov et al . [14] and allow rewiring as well as asymmetric ( DDasym-rw ) by Pastor-Satorras et al . [11] . None of the tested DD variants were in good agreement with PPIyeast , showing negative ΔD and lower clustering coefficient . In yeast , whole genome duplication ( WGD ) occurred relatively recently after speciation of Kluyveromyces waltii and Saccharomyces cerevisiae [31] . Simulation of WGD at the last stage of DD model did not improve the model either ( data not shown ) . As a global topological index , the shortest path length was also examined but provided little discrimination among the tested models due to high variability depending on model parameters ( DD model ) and the choice of yeast PPI data set . Each model was simulated 100 times and the summary of the network properties is given in Table S2 . While additional variants of each model might be considered [13] , [20] , [32] , the critical characteristics of each model are largely captured by these canonical models , e . g . the DD model has no mechanism to generate positive ΔD . The inconsistency of these models with the interaction age density of real PPI networks clearly suggest that none of these canonical models is sufficient in itself to qualify as a valid model for the evolution of the yeast PPI network . To better address both topological and age properties of real networks , we developed an alternative model for PPI network evolution called the crystal growth model ( CG ) , in which we view the growth of a PPI network as analogous to incorporating new proteins into crystals grown in solution ( Figure 3A ) . The two key ideas are as follows . First , the connection probability increases with the availability of unoccupied surface , and thus the model follows anti-preferential attachment rule ( AP rule ) . Second , the connections of a new node tend to be limited within a network module , as observed in growing crystals and here termed as localized connection . The procedure of the CG model is illustrated in Figure 3B . As in the PA and AP models , the CG model starts with a few seed nodes ( N0 = 4 ) , and a new node makes a fixed number of connections ( here , ΔE = 4 ) to existing nodes . For each new node added , network modules are redefined as local dense regions in the network . As modules emerge as a result of network growth and are not pre-defined artificially , the number of modules ( M ) is not fixed but may increase or decrease in each step . With a small probability Pnew , a new node becomes a new module by itself and makes connections ΔE times to other nodes in accordance with the AP rule . Otherwise , an existing module is selected randomly , and the new node is committed to the module by making connections exclusively within the selected module . The connection takes two steps , dubbed “anchoring and extension” . In the anchoring step , the new node connects to an anchor node in the module in accordance with the AP rule , and then , in the extension step , the new node further connects only to the neighbors of the anchor node in the module . Connections are created randomly to neighboring nodes until ΔE connections are made . The anchoring and extension steps are analogous to the node e in Figure 3A ( stage II ) . Therefore , the CG model is inherently highly module-oriented . In case that the neighbors of the anchor node are fewer than ΔE in the chosen module , the module selection and connection step is repeated until ΔE connections are made and the new node becomes connected to multiple modules . The CG model introduces two parameters , how to define the network modules and how frequently a new module is created ( Pnew ) . A network module is generally defined as a densely connected sub-network , and there are various ways to partition a network into modules . Most stringently , modules can be defined as complete subgraphs or cliques , and more loosely they can be defined as k-cores , triangularly connected components ( TCC ) and so on . We tested two different module definitions , one by Newman [33] and the other by TCC . We mainly discuss the results by the Newman definition , but results using TCC were highly similar ( Figure S3 ) . Also , Pnew was assigned as M−1 because the chance of creating a new module generally decreases with the number of existing modules ( M ) . Setting a small , fixed value of Pnew also show a similar result ( data not shown ) . Networks generated by the CG model show a remarkable similarity to real PPI networks for all tested network properties . A typical result of the CG model is shown in the 5th column in Figure 2 . The topology of the CG model shows a scale-free , a hierarchical modular , and a degree-dissortative characteristic . Interestingly , both the magnitude and the shape of clustering coefficient was similar to the PPIyeast in the C ( k ) ∼ k plot ( the row II in Figure 2 ) . The CG model also shows a similar pattern of degree-dissortativity and interaction density with a positive ΔD ( the row III and IV in Figure 2 ) . These characteristics were robust with varying network sizes , e . g . , N = 1 , 000 and N = 5 , 000 ( data not shown ) . The canonical models were shown to significantly deviate from the PPIyeast , but the CG model shows a good agreement not only qualitatively but also quantitatively ( Figure 4 ) . For objective comparison of the models , various indices were used to summarize the network characteristics , including power-law degree distribution ( γ ) , hierarchical modularity ( Q , C , C ( k ) ∼ k curve shape and triangle density , T ) , dissortativity ( δ ) , and the age-dependency of interaction density ( ΔD ) . DD and PA show an inverse age-dependency of PPIyeast and much less modularity in terms of clustering coefficient and triangle density although they show scale-free degree distributions ( Figure 4B and 4C ) . The AP model was not able to generate a scale-free network and significantly deviates from the PPIyeast for all the network indices tested except ΔD ( Figure 4B ) . Only the CG model was comparable to the PPIyeast in terms of all the network indices tested , including both scale-freeness ( γ ) and age-dependency ( ΔD ) ( Figure 4D ) . In particular , only the CG model shows an extremely high degree of modularity comparable to the PPIyeast in terms of both clustering coefficient and triangle density due to its inherently module-oriented mechanism . The mixing exponent ( δ ) is intermediate between LC and HTP . Therefore , of all models considered , the CG model agrees best with both topological and age-dependencies of the actual yeast PPI network . In Table S2 , the network property indices are summarized for all the models tested after 100 simulations of each model . In the CG model , homodimers would be more frequent in older groups because there are simply fewer proteins with which to make connections in earlier stages . The age distribution of homodimeric interactions was exactly in the order of ABE>AE/BE>E>Fu among the 166 homodimeric yeast proteins collected from UniProt [34] and the literature ( Figure 5 , Dataset S4 ) . This result is also consistent with previous studies from protein 3-D structures , in which ancient proteins were shown to be highly enriched with homodimeric or paralogous interactions [35] , [36] . Although the PA and AP would also generate a similar trend , the resulting topology and/or interaction density greatly deviate from PPIyeast to be considered as a realistic model . In the DD model , a fixed interaction probability , p is set for interactions between duplicates ( paralogs ) , therefore implicitly predicts homodimeric formation is age-independent because most paralogous interactions originate from homodimeric interactions and were not created de novo [37] , [38] . Thus , the age-dependency of homodimeric frequencies is a good support for the CG model , which has not previously been applied as a criterion for valid network evolution models . Within the sub-networks of known complexes from MIPS , protein subunits tend to be either more likely to be connected among similar age groups in agreement with the general tendency of positive ΔD in the full yeast PPI networks ( Figures S4A and S4B ) or consist mostly of the same age group , reflecting the creation of a new protein module at a certain evolutionary lineage e . g . actin-associated proteins ( Figure S4E ) . Other complexes form densely connected sub-networks , where age-dependency was not evident , e . g . RNA polymerase I and III ( Figures S4C and S4D ) . We further validated the CG model by inspecting the 3-D subunit arrangement of protein complexes according to age . Obviously , a protein subunit of a stable complex interacts mostly with the subunits of its participating complex . When a subunit is in contact with multiple other subunits in a protein complex , it is most likely that the partner subunits are spatially close , often interacting among themselves as well . For transient interactions , the member proteins can interact with fewer spatial constraints but the interactions are much denser within each biological module , e . g . as for a MAP kinase signaling pathway or transcription initiation complex . Therefore , a protein tends to interact in a highly “localized” manner within the biological modules it belongs to . None of the canonical models has such a module-oriented mechanism as the CG model . In the CG model , older subunits of protein complexes would tend to be more centrally located than younger ones because each protein is attached in the order of its age . Therefore , it is more likely that older subunits are aggregated centrally and younger subunits are scattered at the periphery in a protein complex . To examine this trend among known protein complexes , we collected protein complexes from the Protein Databank ( PDB ) which consisted of at least 3 protein chains , with at least 2 age groups represented; these are stringent criteria that strongly limit the number of available complexes . After removing inappropriate complexes , such as non-protein structures , viral proteins , antibodies and small peptides , a non-redundant set of 12 multi-protein complexes was collected that met these criteria ( detailed descriptions are in Methods ) . In general , older subunits tend to be aggregated centrally ( red tone ) , while younger ones are separated peripherally ( green and blue ) ( Figure 6 ) . In Figure 6A , older subunits form trimeric aggregates but younger ones were separated . There were four linear complexes and no younger subunit intervened between the older ones ( Figure 6B–6E ) . That is , the contacts were always in e . g . the ABE-ABE-AE configuration but not the ABE-AE-ABE , as predicted by the CG model , in which ABE-ABE is connected first and ABE-AE later . The other three complexes contain trans-membrane helix bundles , where the younger helix chain is located at the periphery ( Figure 6F–6H ) . Of the remaining four complexes , two had all subunits contacting each other and were thus non-informative ( Figure 6I–6J ) , and two had ambiguous age assignments for subunits , although the putatively younger subunits were spatially separated ( Figure 6K–6L ) . Considering the eight informative complexes ( Figure 6A–6H ) , the observed subunit arrangements significantly support the CG model at P = 0 . 019 , based on random permutations of chain arrangements within the asymmetric unit of each complex . It is notable that the total degree of PPIyeast is underestimated relative to the actual degree due to homomeric interactions and subunit stoichiometry . For example , the APRIL-TACI complex ( Figure 6A ) was the form A3B3 with the degree kA = 3 ( two homomeric , one heteromeric ) and kB = 1 ( one heteromeric ) . In contrast , only one interaction ( A–B ) would be counted for each subunit in PPIyeast .
The validity of network evolution models have been measured mainly by the resulting network topology , such as a power-law degree distribution , hierarchical modularity and dissortativity as observed in real PPI networks . Accordingly , the DD model has been thought of as the principal mechanism for PPI network evolution . Here , we dissect the history of PPI network evolution by inspecting several protein age-dependent patterns such as interaction density , homodimeric frequency , and the 3-D spatial arrangement of subunits within multiprotein complexes . The age-dependencies are shown to be very effective in discriminating the validity of different models as summarized in Table 1 . The tested aspects of age-dependency were independent of topologies as well as of each other , and are thus highly useful as orthogonal criteria for valid models . Importantly , the age-dependent interaction patterns provided insights on PPI evolution , suggesting evidence against the DD model as the dominant mode of PPI network evolution , instead supporting an alternative model , the CG model . In the CG model , we view the PPI network as sparse and dynamic protein crystals per se . The CG model mimics the process of growing protein crystals in solution by sequentially adding each protein . Despite the huge differences in time scale and heterogeneous composition , PPI network evolution likely obeys similar constraints on growing protein crystals . In the CG model , a protein complex or a tightly linked module is analogous to individual crystals , and the number and membership of modules are not pre-defined but rather emerge naturally in each growing step . Crystals grow around multiple nuclei just as protein networks consist of multiple modules/complexes . New modules are generated as the genome size increases and novel function evolves in higher organisms , in a manner similar to how a new crystal forms occasionally through new nucleation events . The CG model exploits two keys ideas , the first being that the chance of new connection is proportional to the availability of free surface , which is a feature readily recognized by a new protein molecule; this results in an anti-preferential attachment ( AP ) rule . Although the same surface of a protein can be involved in multiple interactions with different partners through spatial and temporal differentiation , such a factor uniformly increases the capacity of interactions in any protein . Therefore , the connection probability is still positively correlated with the available surface area . These results agree with those of Kim et al . [39] , which show that the evolutionary rate is anti-correlated with available surface area . There , multi-interface hubs were nearly four times more frequent than single-interface hubs , reflecting the dominant connection mode of the AP rule . The second key idea is that once an initial connection is made , the subsequent connections are localized to the neighbors of the initial partner within the same module . This localized connection enforces high modularity , similar to that observed in real PPI networks . At the basis of the crystal growth model is the notion that new interactions form preferentially within existing physical complexes ( enforcing modularity ) , and thus are limited by available protein surface area ( the AP rule ) . Thus modularity & the AP rule both arise due to simple physical constraints of which proteins are most accessible to each other . Recently , Levy and colleagues has shown that the successive steps of homo-oligomeric assembly mimics the evolutionary pathway [38] . The CG model expands this idea , where crystal growth reproduces the evolution of the entire PPI network . Given that the CG model follows an AP rule , how does it generate scale-freeness or “the rich get richer” connectivity ? In the CG model , the network grows by anchoring and extension , where a node increases its degree either by becoming an anchor node ( anchoring ) or by being the neighbor of the anchor node ( extension ) . Therefore , the highly connected nodes have greater chances to increase their degree within each module because they have more opportunities to have anchors as their neighbors . Therefore , the CG model implicitly implements the preferential attachment ( PA ) rule within each module in a manner similar to the DD model , where the nodes increase their degree by having duplicating genes as their neighbors . Our result suggests that the CG model is a more plausible mechanism for PPI network evolution than the DD model . First , all the age-dependent aspects tested agree well with the CG model but disagree with the DD model . Second , the CG model is more comprehensive than the DD model in that the CG model can accommodate both gene duplication and horizontal gene transfer as the origins of new nodes ( genes ) . Practically , the DD model may be applicable only to ∼20% of the yeast proteome having identifiable duplicates [40] . The CG model also embodies the rapid divergence of gene duplicates [17] by the AP rule , which avoids competition for the same interface on common partners and connects to new partners with less occupied surfaces . Finally , the CG model is more robust than the DD model . The DD model shows a highly variable degree distribution depending upon parameters and network sizes [14] , [41] . In contrast , the CG model shows stable characteristics regardless of network size or different module definition methods . Taken together , these strongly suggest that the DD model is unlikely to be the principal , and strongly unlikely to be the sole , mechanism of PPI network evolution . The age-dependency of interaction density also sheds light on a more fundamental question regarding the mechanism of PPI network evolution . It has been hypothesized that inherent features of proteins , such as stickiness and hydrophobicity are dominant factors in shaping the global network structure [6] . However , the observed age-dependency is inconsistent with such a hypothesis and suggests that a stochastic process played a major role . For example , the yeast PPI network shows the patterns of both DABE , AE/BE>DABE , E and DAE/BE , Fu<DE , Fu ( the row IV in Figure 2 ) . The connection probability cannot depend solely upon a feature such as protein length or surface hydrophobicity because no single feature ( F ) can satisfy FAE/BE>FE ( with common FABE ) and FAE/BE<FE ( with common FFu ) simultaneously . Power-law distributions have been commonly observed in various types of networks , such as the Internet , social networks , and biological networks . However , the growth of a PPI network poses unique constraints compared to other types of networks . For example , in an airline or railroad network , each new connection is made by considering the context of global network topology ( e . g . , to minimize average path length ) , which seems intuitively unlikely to be the case in PPI networks . The CG model follows two simple constraints of available free surface and localized connection , which are physically plausible and depend only on local context but not global topology . With these minimal assumptions analogous to growing protein crystals , the CG model recapitulates remarkably well the age-dependencies as well as the network topologies of the yeast PPI networks .
Two independent sets of yeast protein-protein interaction data were collected using a method essentially identical to that described by Batada et al . [23] , [24] , only differing in that the HTP set was collected from the original publications instead of from BioGrid [42] . We compiled the HTP set from Uetz et al . [43] , Ito et al . [44] , the merged set of Gavin et al . [45] , [46] , Ho et al . [47] , and Krogan et al . [48] , and then filtered out the interactions supported by only a single experiment . Repeated and reciprocal assays were considered as independent experiments even if they were performed in the same publication . The LC data set was collected from the latest release of BioGrid , excluding high-throughput data . Ribosomal proteins were removed from both LC and HTP data sets . All protein-RNA interactions and interactions supported only by co-localization or co-fractionation were removed . We further removed interactions supported only by Ptacek et al . [49] , Grandi [50] , Collins et al . [51] , or Fields et al . [52] . Pfam domains were assigned for yeast proteins using BioMart ( http://www . biomart . org ) . The taxonomic distributions of Pfam domains were obtained for archaea ( A ) , bacteria ( B ) , eukaryotes ( E ) , and fungi ( F ) ( http://www . sanger . ac . uk/Software/Pfam ) . According to these distributions , each Pfam domain was assigned to one of the age groups ABE , AE/BE , E , and F . The group ABE includes the oldest proteins common to all three kingdoms , while group F is the youngest , being specific to fungi . As yeast is a eukaryote , groups A , B , and AB do not occur . A protein's age group was assigned as the youngest age of its constituent Pfam domains—e . g . , E for a protein with domains from ABE and E ( Dataset S3 , Figure S1 ) . Interaction density Dm , n measures the normalized interaction density between two age groups m , n ( m<n ) . ΔD measures the interaction preference of a new node by the age differences . A positive value of ΔD indicates that a new node makes connections more frequently with close age groups than with distant ones . First , the normalized interaction density Dm , n between two age groups m , n ( m<n ) is calculated aswhere lm , n is the number of edges between the two age groups m and n , and Em , n is the number of all possible interactions between the two groups . Nm and Nn are the number of nodes in the age groups m and n , respectively , L is the total number of edges , and N is the total number of nodes in the network . Then the average interaction density gradient , ΔD , of a network is defined aswhere G ( G≥2 ) is the number of age groups . The modularity of a network is measured by the modularity index Q by Newman [29] after its modules are defined using the method described in [33]:where M = the total number of modules , L = the number of total edges in the network , ls = the number of edges within the module s , and ds = the sum of the degrees of the module s . The modularity index Q measures the difference between the intra-module interaction density and the expected interaction density at random for a given partition , where Q≈0 for a random network and Q = 1 for a completely modular network [53] . The list of PDB entries and 3-D coordinates were obtained from PQS ( Protein Quaternary Structure Server , ftp://ftp . ebi . ac . uk/pub/databases/msd/pqs ) . First , we took the PDB entries having three or more protein chains . The PDB entries annotated as crystal packing interfaces by PQS or from non X-ray crystallographic method were excluded . The protein chain clusters at 30% sequence identity cut-off were downloaded from PDB ( Protein Data Bank , ftp://ftp . wwpdb . org ) . PDB entries consisting of the same set of NR30 clusters were grouped together regardless of the number of chains and one representative PDB entry was selected in each group as NR30 entries . For NR30 entries , the age group of each PDB chain was assigned using BLAST against NR90 set of archaea , bacteria and eukaryote sequences from UNIPROT ( ftp://ftp . uniprot . org/pub/databases/uniprot ) using >30% identity and >30 alignment length as criteria . We took only the PDB entries consisting of two or more protein age groups and further applied a number of filters manually , excluding the entries with DNAs , RNAs , viral proteins , small peptides ( <30 amino acids ) and immunoproteins such as antibodies and MHCs with antigens . Where available , ambiguous quaternary structures were removed by comparing the data from PQS , PDB biological units and 3D complex databases [54] . | Proteins function together forming stable protein complexes or transient interactions in various cellular processes , such as gene regulation and signaling . Here , we address the basic question of how these networks of interacting proteins evolve . This is an important problem , as the structures of such networks underlie important features of biological systems , such as functional modularity , error-tolerance , and stability . It is not yet known how these network architectures originate or what driving forces underlie the observed network structure . Several models have been proposed over the past decade—in particular , a “rich get richer” model ( preferential attachment ) and a model based upon gene duplication and divergence—often based only on network topologies . Here , we show that real yeast protein interaction networks show a unique age distribution among interacting proteins , which rules out these canonical models . In light of these results , we developed a simple , alternative model based on well-established physical principles , analogous to the process of growing protein crystals in solution . The model better explains many features of real PPI networks , including the network topologies , their characteristic age distributions , and the spatial distribution of subunits of differing ages within protein complexes , suggesting a plausible physical mechanism of network evolution . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"biochemistry/molecular",
"evolution",
"biophysics/macromolecular",
"assemblies",
"and",
"machines",
"biophysics/theory",
"and",
"simulation",
"computational",
"biology/evolutionary",
"modeling",
"biochemistry/bioinformatics",
"biochemistry/theory",
"and",
"simulation",
"biochemistry... | 2008 | Age-Dependent Evolution of the Yeast Protein Interaction Network Suggests a Limited Role of Gene Duplication and Divergence |
Fusion peptides from influenza hemagglutinin act on membranes to promote membrane fusion , but the mechanism by which they do so remains unknown . Recent theoretical work has suggested that contact of protruding lipid tails may be an important feature of the transition state for membrane fusion . If this is so , then influenza fusion peptides would be expected to promote tail protrusion in proportion to the ability of the corresponding full-length hemagglutinin to drive lipid mixing in fusion assays . We have performed molecular dynamics simulations of influenza fusion peptides in lipid bilayers , comparing the X-31 influenza strain against a series of N-terminal mutants . As hypothesized , the probability of lipid tail protrusion correlates well with the lipid mixing rate induced by each mutant . This supports the conclusion that tail protrusion is important to the transition state for fusion . Furthermore , it suggests that tail protrusion can be used to examine how fusion peptides might interact with membranes to promote fusion . Previous models for native influenza fusion peptide structure in membranes include a kinked helix , a straight helix , and a helical hairpin . Our simulations visit each of these conformations . Thus , the free energy differences between each are likely low enough that specifics of the membrane environment and peptide construct may be sufficient to modulate the equilibrium between them . However , the kinked helix promotes lipid tail protrusion in our simulations much more strongly than the other two structures . We therefore predict that the kinked helix is the most fusogenic of these three conformations .
Membrane fusion is critical to eukaryotic cellular function and also provides the mode of entry for enveloped viruses such as influenza and HIV . Influenza viral entry is mediated by the hemagglutinin protein . As part of this process , short fusion peptides are inserted into the host membrane and act to promote fusion . Influenza is a distinctive system for studying fusion because hemagglutinin mutants have been generated that can insert fusion peptides and pull viral and target membranes together but not complete the fusion process [1]–[5] . This can be accomplished via mutations in the fusion peptide region or deletions in the transmembrane domain . These mutagenesis results suggest a specific role for fusion peptide-membrane interactions in promoting influenza membrane fusion . The mechanism by which fusion peptides act on membranes remains unknown , but some possibilities that have been previously suggested include inducing membrane curvature , altering local membrane composition , and inducing local disorder in membrane lipids [6]–[9] . Because of the dynamic and heterogeneous nature of membrane assemblies and the transience of fusion intermediates , molecular dynamics simulations have been used to generate and examine physical hypotheses for fusion mechanisms . These simulations have suggested that hydrophobic tail protrusion into the polar layer between two apposed bilayers ( Fig . 1 ) may be an important feature and indeed a transition state for fusion stalk formation [10]–[13] . We have previously shown that influenza fusion peptides can promote lipid tail protrusion in simulations without loss of overall lamellar structure [10] . However , these predictions and their consequences for influenza fusion are difficult to test spectroscopically . Several structural models exist for native hemagglutinin fusion peptides in membranes , obtained via different approaches and under different conditions . NMR experiments in micelles and EPR experiments in bilayers have provided a kinked helix model for the structure of the fusion peptide of X-31 ( H3 ) hemagglutinin in bilayers . This model is consistent with additional infrared and circular dichroism spectroscopic data [6] , [14] , [15] . Solid-state NMR experiments in bilayers have yielded structures that are grossly similar but have a slightly more pronounced kink [16] . More recently , NMR studies in micelles using a longer construct from A/swine/Scotland/410440/94 ( H1 ) hemagglutinin have yielded a helical hairpin structure [17] , [18] . Finally , simulation studies have also suggested that a relatively flat helical model may be appropriate [19] , [20] , though other simulations have yielded a kinked helix [21] , [22] or rapid exchange between the two [19] , [23] . Additional NMR structural data are available for a series of fusion peptide mutants [24] , including the N-terminal glycine mutants discussed below . Here , we wish to better understand how influenza fusion peptides can drive fusion . We first report an indirect test of tail protrusion as a surrogate outcome for catalysis of fusion stalk formation in influenza membrane fusion . This is accomplished by simulating mutant fusion peptides with known experimental phenotypes and comparing lipid protrusion in these simulations to experimental fusogenic activity . We then use tail protrusion as a means to examine how different fusion peptide conformations may contribute to fusogenic activity . We find that all three major postulated conformations are accessible in our simulations . Since the precise conformational equilibria are likely controlled by membrane environment , peptide sequence , and sample preparation , this study does not address the question of which peptide conformation predominates under the membrane conditions of fusion . Indeed , both bilayer lipid composition and choice of detergent for membrane protein structure determination can substantially affect protein structure [25] , [26] . Instead , we use simulations to analyze how each of these conformations may contribute to lipid protrusion and fusion activity .
We performed molecular dynamics simulation of influenza fusion peptides and mutants in lipid bilayers and evaluated the resulting trajectories for lipid tail protrusion close to the peptides . A protrusion event was defined as any carbon in the lipid tail protruding more than 0 . 1 nm beyond the phosphorus atom of that lipid ( Fig . 1 ) . Lipid protrusion probabilities are plotted in Fig . 2a as a function of distance from the nearest fusion peptide . As hypothesized , lipids near the G1V mutant peptide showed significantly less protrusion ( p<0 . 01 , Kolmogorov-Smirnov test ) than lipids near the wild-type X-31 peptide , displaying only a minimal increase over baseline . Lipids near the G1S mutant peptide showed protrusion intermediate between the wild-type and G1V peptides ( p<0 . 01 via Kolmogorov-Smirnov with Bonferroni correction ) , consistent with the experimental observation that G1S hemagglutinin catalyzes lipid mixing between fusion partners ( a measurement of early fusion intermediate formation ) but less efficiently than wild-type ( Fig . 2b ) . Further testing of additional mutants will be helpful , but our simulation results to date support lipid tail protrusion as a predictive measure of fusogenic activity . Protrusion is a not-infrequent event in pure lipid bilayers , although it appears to be specifically enhanced by fusogenic peptides . Based on the probability of protrusion in simulations of a pure 1-palmitoyl 2-oleoyl phosphatidylcholine ( POPC ) bilayer , we estimate the free energy cost for any acyl chain carbon of a lipid to protrude at least 0 . 1 nm at approximately 0 . 5 kT; this increases to 2 . 2 kT if a protrusion threshold of 0 . 2 nm is used . We have previously tested protrusion in the POPC bilayer surrounding an ion channel , and no significant increase was observed over a protein-free bilayer , suggesting that protrusion is not a general phenomenon near inserted proteins but likely more specific to membrane-disordering and perhaps fusion peptides [10] . Furthermore , at the 1∶167 peptide∶lipid ratio used , enhancement of protrusion was a local effect: at long distance from the peptides , protrusion probabilities are identical within error to protein-free bilayers . Calculated SCD order parameters for peptide-free and peptide-inserted bilayers indicate that the membranes remain lamellar in the presence of peptides and are not grossly disordered ( Fig . S1 ) . To test for direct peptide-lipid interactions that might result in protrusion , we examined interaction energies between fusion peptides and each lipid in their immediate surroundings . These energies were computed at 1 ns intervals for all lipids within 12 Å of a peptide , using the AMBER03 force field , over the course of a 100-ns simulation of the X-31 fusion peptide and one of the G1V mutant . Lipids with a protruding acyl tail did not show a significant difference in interaction energy with the protein compared to lipids that did not protrude ( Fig . S5 ) . This suggests that the majority of lipids with protruding acyl tails are not participating in strong interactions with the peptide that would drive such behavior . Fusion peptides appear to promote acyl tail protrusion via a local change in bilayer order rather than either a global disordering of the bilayer or highly specific intermolecular interactions . While average SCD order parameters were displayed only a slight shift between peptide-containing and peptide-free bilayers , there was a significant change in order parameters of lipids very close to the peptide ( Fig . 3 ) . To distinguish generic interfacial effects from those specific to active fusion peptides , we compared lipid order parameters in the X-31 wild type peptide simulations and those of the fusion-null G1V mutant . Lipids closest to the X-31 peptide ( within 7 Å of the nearest peptide atom , approximately the first lipid shell ) showed a significant reduction in order parameters , while lipids in the same leaflet but farther than 7 Å did not ( Fig . 3a; p<0 . 001 , Kolmogorov-Smirnov test for each carbon C3–C14 ) . Lipids close to the G1V mutant peptide displayed a significantly smaller effect ( p<0 . 001 , Kolmogorov-Smirnov; Fig . 3b ) . This finding provides additional evidence of a highly localized disordering effect specifically in the presence of active fusion peptides . Fusion-active peptides thus appear to induce substantial disorder in a small fraction of nearby lipids . One possible mechanism for this is partial insertion of the peptide into the bilayer outer leaflet , displacing volume in the hydrocarbon region and leaving a localized pressure imbalance in the polar region . In our simulations , we occasionally observe a lipid “straddling” the partially inserted fusion peptide helix ( Fig . S4 ) , with one acyl tail on each side of the helix . Such straddling conformations have been previously observed in crystallographic structures of aquaporins [27] . In our simulations , straddling lipids did not always protrude past the phosphate group , but when they did , the protrusion time appears longer . The average protrusion probability for lipids straddling a kinked helix conformation was significantly higher than distance-matched non-straddling “control” lipids ( p<0 . 01 , Kolmogorov-Smirnov test ) . Straddle-related acyl chain protrusion does not account for the full difference between fusion-active and fusion-null peptides . However , it may be an example of a more general phenomenon for how partial insertion of a kinked fusion peptide helix leads to lipid tail disordering and protrusion . We also performed simulations to examine the fusogenic activity of different structural models for the influenza fusion peptide . In our simulations of the X-31 fusion peptide , the peptide sampled kinked-helix conformations , flat helical conformations , and helical hairpin conformations , coming within 1 . 2 Å C-alpha RMSD of NMR structural models for each of these ( see Text S1 for details ) . We therefore performed additional simulations where the same X-31 fusion peptide was restrained to each of these conformations using an elastic network model . Six restrained simulations of up to 200 ns each were run per structural model . We measured lipid tail protrusion in these simulations as a surrogate for fusogenic activity . The constructs used for different NMR structural studies varied in length; all simulations reported here use residues 1–20 of influenza hemagglutinin . The kinked helix produced by far the greatest probability of lipid protrusion ( Fig . 4 ) , significantly more than either the flat helix or hairpin simulations ( p<0 . 01 , Kolmorogov-Smirnov test ) . The increase in tail protrusion in our simulations takes the form of more frequent protrusions rather than longer persistence of each protrusion ( Fig . S2 ) . These results suggest that among the models tested here , whatever the equilibrium probability of each conformation at the conditions of fusion , the kinked helix has the greatest fusogenic activity . As an additional test of the effect of fusion peptide conformation on tail protrusion , we analyzed our unrestrained simulations of the X-31 fusion peptide and asked what conformation the peptide adopted at the start of each protrusion event . 60% of all protrusion events originated from a kinked helix conformation ( Fig . 5 ) , 32% from a flat , predominantly helical conformation , and 8% from a helical hairpin conformation . Mean helical kink angles for each of these conformational groups were 110° for the kinked helices , 154° for the flat helices , and 74° for the hairpin-like structures ( see Text S1 ) . Similarly , the protrusion events that did occur in simulations of the G1V mutant peptide were most likely to occur when the peptide adopted a kinked-helix conformation ( Fig . 6 ) , which happened much less often than for the X-31 peptide . This assessment of which conformations are associated with the most protrusion events is of course biased by the fraction of time the simulations spent in each conformation . Nonetheless , we obtain the same results from two orthogonal approaches: the restrained simulations where we pre-suppose a set of structural models and equalize the sampling of each model and the unrestrained simulations where we have uneven sampling but discover protrusion-associated conformations in an unbiased fashion . Both suggest that a kinked helix is the most fusion-active conformation among those visited by our simulations .
Influenza fusion peptides clearly act to promote membrane fusion , but the mechanism of their action has been extremely difficult to probe . We have demonstrated that lipid protrusion probability near a mutant fusion peptide correlates with the ability of that hemagglutinin mutant to drive lipid mixing in cell transfectants and indeed with lipid mixing rates . This correlation supports the hypothesis that the transition state for stalk formation involves lipid tail protrusion and suggests that tail protrusion may be an effective surrogate outcome in evaluating fusogenic activity of peptides . Since our simulations of the X-31 fusion peptide sample conformations close to the kinked helix , flat helix , and helical hairpin structural models that have been previously proposed for influenza fusion peptide structure , we have used lipid tail protrusion to probe the fusogenicity of each conformation . Our results suggest that 1 ) the free-energy differences between these three states may be low enough that lipid environment and peptide sequence may have a substantial effect on the predominant population and 2 ) the kinked-helix state is significantly more fusogenic than the flat helix or helical hairpin . Lipid tail exposure to the water layer , as would occur with tail protrusion , has been observed experimentally via NMR in POPC vesicles [28] as well as DPPC bilayers [29] and via neutron diffraction in DOPC bilayers [30] . In the NMR experiments , these are detected as Nuclear Overhauser Effect cross-relaxation between acyl-chain methyl and choline methyl groups . An increase in lipid tail protrusion due to fusion peptide activity should be detectable via such methods . However , two factors complicate the experiment . The first is that our simulations predict the effect on tail protrusion to be most profound in the upper region of the acyl chain , so a difference in tail exposure might be suboptimal probe for fusion peptide activity . Second , detecting an increase in average tail exposure over the entire sample might require a higher peptide∶lipid ratio than is physiological , since we predict tail protrusion to be a localized effect . If these two challenges can be overcome , these spectroscopic experiments would provide a good means to test our hypotheses . Amphipathic helices have been observed to be membrane-active in a wide variety of physiological contexts . For influenza , it appears that the precise conformation of the helical residues has an effect on local membrane properties and may control fusion activity . Here , we have explored the relationship of fusion activity to leading structural models for influenza hemagglutinin . The physical determinants of fusion activity in systems like parainfluenza , where the prevailing model is that of a transmembrane helix for the fusion peptide [31] will be an interesting subject for future investigations . Another challenge for the future is to integrate simulation with spectroscopic data to examine how different membrane environments alter the conformational equilibria of fusion peptides and how this may act to regulate the efficiency of viral entry in cellular systems .
To evaluate the relationship between lipid protrusion and fusogenic activity , we simulated a set of hemagglutinin mutants for which extensive structural and fusion activity data are available . The N-terminal glycine of HA2 is heavily conserved—among deposited human influenza hemagglutinin sequences [32] , >99 . 9% have glycine in this position . Cells transfected with G1S hemagglutinin display a terminal hemi-fused phenotype and reduced lipid-mixing rates compared to wild-type X-31 , and the G1V mutation blocks lipid mixing altogether [2] . This series of mutations thus provides a sensitive test of our protrusion hypothesis: the G1S mutant displays a block after stalk formation and would be expected to promote tail protrusion , while the fusion-null G1V mutant would not . We therefore simulated three copies of each fusion peptide in a POPC bilayer at a 1∶167 peptide∶lipid ratio . All simulations were started from the kinked helix structure ( PDB code 1IBN ) docked into the bilayer according to EPR data [33] with appropriate mutations; the structural differences among mutants detected via NMR were not assumed . Lipids strongly overlapping the inserted fusion peptide were removed ( approximately 4 per peptide ) ; the criteria used for lipid deletion were phosphate groups within 7 Å of the peptide as previously reported [33] . Any protrusion propensity resulting from such a pressure imbalance would be a factor 2 lower than what we observe ( Fig . S3 ) . Three copies of the fusion peptide and approximately 500 POPC molecules were solvated with TIP3P water in a box with sides 13 nm and height 6 . 5 nm in length . The AMBER03 force field [34] was used for protein parameters in conjunction with the Berger lipid parameters [35] as previously described [10] . This combination of the Amber force field , Berger lipid parameters , and TIP3P water has been used previously [10] , [36] . As an additional validation , we calculate the area per lipid head group in equilibrated POPC bilayers to be 0 . 69 nm2 , closely matching the experimental value of 0 . 683±0 . 15 at 303K [37] . Data reported are for simulations with charged N- and C-termini on the peptide . Simulations were performed using GROMACS 4 . 5 [38]; 200 simulations were run per mutant with an average length of 215 ns per simulation for an aggregate of 368 peptide-microseconds . Simulation details were as follows: a constant temperature of 310K ( 37C ) was maintained using the velocity-rescaling thermostat [39] , and pressure was maintained semi-isotropically at 1 bar via the Berendsen method . All covalent bond lengths were constrained using LINCS [40] , and long-range electrostatics were computed every step using Particle Mesh Ewald ( PME ) [41] . The extensive simulations reported here were designed to probe equilibrium properties of the fusion peptide conformational ensemble . However , in examining the simulation trajectories , we detected some extremely long-lived states ( decorrelation times >200 ns ) , such that we believe estimates of equilibrium properties may be prone to error . This finding is perhaps not surprising—for as small a protein as Trp-zip in water , prior simulation studies have found long-lived misfolded states [42] , [43] . It is possible that decorrelation times in a membrane environment may be even slower . We have thus elected to report the results of a more conservative analysis; methods to analyze the equilibrium distribution of peptide conformations remain a subject of active inquiry and future work . Unrestrained simulations were run using the Folding@Home [44] distributed computing network . In the restrained simulations , distance restraints were applied between every pair of backbone atoms within 7 Å of each other at a force constant of 1000 KJ mol−1 nm−2 . The following conformational models were used as targets for the restrained simulations: the X-31 NMR structure in micelles by Han and co-workers [45] ( PDB code 1IBN ) for the kinked helix , the A/swine/Scotland/410440/94 NMR structure in micelles by Lorieau and co-workers [17] ( PDB code 2KXA ) for the helical hairpin , and the G1V X-31 NMR structure in micelles ( PDB code 1XOP ) by Li and co-workers [24] as an idealized straight helix . Simulations were run with and without protonation of residue Glu11; the same relationship in tail protrusion probability was observed with and without protonation of Glu11 . Figures in the main text display analysis performed on the unprotonated systems , and data for protonated Glu11 is plotted in Fig . S3 . All simulations are summarized in Table 1 , and additional details of the analysis are given in the Text S1 as well as sequences of all simulated peptides . | Membrane fusion is a common process critical to both cellular function and infection by enveloped viruses . Influenza is a particularly useful model system for studying fusion because the fusion reaction is accomplished by a single protein , hemagglutinin . Furthermore , mutations to the membrane-inserted portion of hemagglutinin have been identified that do not detectably alter the rest of the protein but can either arrest fusion halfway or block it entirely . For influenza at least , it seems that the membrane-inserted hemagglutinin peptide plays a critical role in promoting fusion , perhaps by increasing the local disorder of lipid bilayers . However , we lack a mechanistic understanding sufficient to predict the activity of fusion peptide mutants from their sequence . Here , we have used lipid tail protrusion as a way to measure how much fusion peptides disorder their surrounding bilayer; we see a strong relationship between lipid tail protrusion and the ability of fusion peptide mutants to promote lipid mixing between membranes . Our simulations also predict that this lipid tail protrusion is much more common when the peptides adopt a kinked helix structure than when they are straight or hairpin-like . We therefore hypothesize that , while all three types of structure likely undergo conformational exchange , the kinked helix structure is most active in promoting fusion . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"biology",
"computational",
"biology"
] | 2013 | Lipid Tail Protrusion in Simulations Predicts Fusogenic Activity of Influenza Fusion Peptide Mutants and Conformational Models |
The extent by which different cellular components generate phenotypic diversity is an ongoing debate in evolutionary biology that is yet to be addressed by quantitative comparative studies . We conducted an in vivo mass-spectrometry study of the phosphoproteomes of three yeast species ( Saccharomyces cerevisiae , Candida albicans , and Schizosaccharomyces pombe ) in order to quantify the evolutionary rate of change of phosphorylation . We estimate that kinase–substrate interactions change , at most , two orders of magnitude more slowly than transcription factor ( TF ) –promoter interactions . Our computational analysis linking kinases to putative substrates recapitulates known phosphoregulation events and provides putative evolutionary histories for the kinase regulation of protein complexes across 11 yeast species . To validate these trends , we used the E-MAP approach to analyze over 2 , 000 quantitative genetic interactions in S . cerevisiae and Sc . pombe , which demonstrated that protein kinases , and to a greater extent TFs , show lower than average conservation of genetic interactions . We propose therefore that protein kinases are an important source of phenotypic diversity .
Genetic variation , in the form of point mutations , gene duplication/loss , and recombination serves as the raw material upon which natural selection acts during the evolution of a species . To understand this evolutionary process , we must in turn be able to understand how this variation translates into phenotypic changes that have a measurable impact on fitness . The great advances in DNA sequencing and comparative genomic analysis have brought us tremendous insight into the organization of genomes and the extent of genomic variation across species [1]–[4] . Similarly , gene expression studies have recently been used to study the evolution of transcriptional regulation [5]–[7] . Still , expression levels offer a very limited view of the inner workings of the cell . Other technologies are now maturing that allow us to analyze , in high-throughput fashion , how molecular components such as proteins are modified [8]–[11] and interact , either physically [12]–[18] or genetically , to enable the cell to carry out its essential functions . Recently , comparison of protein interaction networks in different species has been used to propose that protein–protein interactions change at a fast evolutionary rate after gene duplication [19] , [20] . In particular , interactions of lower specificity , such as those mediated by short linear motifs ( i . e . , peptide-binding domains ) , were postulated to have a higher rate of change and might therefore display greater potential to generate functional diversity [20] , [21] . In parallel with these efforts , the study of particular cellular functions has provided us with fascinating examples of the evolution of cellular interactions [22] , [23] . Tsong and colleagues [23] have shown that although the mating processes in Saccharomyces cerevisiae and Candida albicans are phenotypically similar ( both controlled by a conserved MAT locus ) , the regulatory arrangements that specify the mating types are different . These authors were able to trace mutations in one of the proteins involved ( alpha-2 ) that have contributed to the changes in regulation . Similarly , Moses and colleagues have shown that regulation of the nuclear localization of the MCM complex by Cdk phosphorylation of Mcm3 was acquired in the Saccharomyces lineage but does not occur in C . albicans [22] . Therefore , solutions to evolutionary problems , originating at the DNA level , may be manifested in different ways at the protein network level . In this study , we focus on the role of one of these mechanisms , that of protein phosphorylation . Protein phosphorylation is a ubiquitous and reversible modification that is crucial for the regulation of cellular events [24] . Protein kinases phosphorylate their peptide substrates by recognizing motifs that consist of a few key residues surrounding the target amino acid . The high regulatory and evolutionary potential of protein kinases make protein phosphoregulation a prime candidate for evolutionary studies . Recent technological developments now permit us to comprehensively study the in vivo phosphorylation of proteins for multiple species [8]–[10] , [25] , [26] . Comparison of these results shows that they contain significant overlap that relates to species taxonomy [27] . However , this approach has not yet been used to study the evolution of phosphoregulation on a large scale . We have carried out a mass spectrometry ( MS ) analysis of the in vivo phosphoproteome of three fungal species ( S . cerevisiae , C . albicans , and Schizosaccharomyces pombe ) , and we used these data to generate a cross-species analysis of phosphoregulation . We quantified the rate of evolutionary change of protein phosphorylation and analyzed the divergence of kinase–substrate interactions for particular protein complexes . Finally , we tested and validated the observed evolutionary trends through comparative genetic interaction studies .
We used a MS approach to globally determine the in vivo phosphorylation status of the S . cerevisiae , C . albicans , and Sc . pombe proteomes under exponential growth in rich media . The dataset is of high quality , with false positive rates ( FPRs ) varying from 1 . 3–1 . 7% ( see Methods ) . In total we could identity 1 , 185 , 1 , 449 , and 850 phosphoproteins in S . cerevisiae , C . albicans , and Sc . pombe , respectively , and within these , we identified 3 , 486 , 4 , 715 , and 1912 phosphosites ( Table 1 and Dataset S1 ) . The distributions of phosphorylation in these three screens among serine , threonine , and tyrosine is similar to those observed previously for studies in budding yeast [9] , [26] , [28] , [29] with the majority of phosphorylation occurring at serine ( 73–83% ) , followed by threonine ( 15–25% ) , and small numbers of tyrosines ( 0 . 8–1 . 9% ) . The small fraction of detected phosphotyrosines is expected given the absence of identifiable tyrosine kinases in these species . To estimate the coverage of these datasets , we calculated the overlap with previous phosphorylation studies of S . cerevisiae [9] , [26] , [28] , [29] and Sc . pombe [10] . The estimated coverage of our phosphorylation sets ranges from 51–71% for detection of phosphoproteins , 43–62% for detection of phosphorylated peptides ( 10-amino acid peptide ) , and 20–31% for correct detection of previously known phosphosites ( see Protocol S1 ) . One potentially confounding effect is abundance bias in the determination of phosphoproteins , with phosphoproteins being potentially over- or under-sampled because they are more or less abundant than other proteins . To address this issue , we used experimentally determined concentration values that were systematically generated for individual proteins in S . cerevisiae [30] . Although phosphorylated proteins are on average three times more abundant when compared to all others ( p-value = 6 . 3×10−13 with a t-test ) , this difference is small compared to the eight orders of magnitude spanned by the abundance of all proteins . In fact , the known phosphoproteins also span similar orders of magnitude ( Protocol S1 ) , and therefore this small abundance bias is unlikely to explain observed differences in protein phosphorylation across the different species . Therefore , we assembled a high-quality cross-species phosphoprotein database that is suitable for addressing questions concerning the evolution of phosphoregulation . Using this dataset , we first attempted to quantify the rate of change of individual phosphoproteins across species to estimate the rate at which species change kinase–substrate interactions during evolution . To calculate this rate , we first compiled the majority of previously published in vivo protein phosphorylation data generated for S . cerevisiae [9] , [26] , [28] , [29] . The coverage of the combined set ( estimated using leave-one-out analysis ) ranged from 81–92% , indicating that the combined set of 1 , 956 S . cerevisiae phosphoproteins is reaching completeness , at least for exponential growth in rich medium with currently available MS approaches . We assumed an estimated coverage of 92% and used the phosphorylation information for other species to calculate the rate of change of protein phosphorylation during evolution ( Table 2 , Methods ) . For each test species , we calculated the number of phosphoproteins expected to be observed in S . cerevisiae by homology as 92% of the number of orthologous phosphoproteins in that species . We then defined as the number of evolutionary changes in phosphorylation the difference between the observed conserved phosphoproteins and the expected value by homology . We estimated that , on average , 1×10−4 proteins changed their phosphorylation status per protein per million years ( My ) . Assuming that the gain or loss of a phosphoprotein corresponds to the gain or loss of up to five kinase–substrate protein–protein interactions , we estimate that kinase–substrate interactions change at a rate of approximately 1×10−6 to 1×10−5 interactions per protein pair per My ( Methods ) . Interestingly , these estimates are similar to previously calculated rates of change for protein–protein interactions after gene duplication [19] , [20] . This value likely represents a lower bound estimate , because changes of kinase–substrate interactions can occur without changing the total number of phosphoproteins . We next considered that evolutionary changes in phosphosite position should also be considered a change of kinase regulation . To estimate the rate of change in kinase–substrate interactions considering also changes in phosphosite locations , we aligned S . cerevisiae proteins to their corresponding orthologs in other species using a general purpose sequence alignment tool ( TCoffee , http://www . tcoffee . org ) . We considered that a phosphosite in an orthologous protein had diverged when no phosphosite was observed in the S . cerevisiae protein within an alignment window ranging from 20 to 200 alignment positions centered on the phosphosite of the orthologous protein . The rate of change of kinase–substrate regulation calculated in this way is 5 to 7 times faster ( depending on the alignment window size ) than the same calculations based on the phosphorylation status of the full proteins . Our calculations can be compared with estimates for the rate of change of transcriptional regulation . This rate can be obtained from data of binding of three transcription factors ( TFs ) to promoter regions for different yeast species [17] , [18] , and similar information available for human and mouse [31] . Based on these studies , we estimate that TF binding to promoters change at an order of 1×10−4 to 3×10−4 per TF–gene interaction per My , at most two orders of magnitude faster than kinase-substrate turnover ( Methods and Protocol S1 ) . The results above suggest that , as a whole , kinase–substrate interactions can change quickly during evolution . We then asked if functionally related sets of proteins show significant differences in level of phosphorylation across species . We transferred the gene ontology and protein complexes information available for S . cerevisiae to other species using orthology assignments . In this way , we defined , for each species , sets of proteins grouped according to their functional categories or protein complex membership . We then calculated the number of phosphosites per protein within each group , normalized by the average number of phosphosites per protein in the proteome . We observed a generally high correlation of the number of phosphosites per protein across different functions for all three species studied ( Figure 1A ) . For instance , proteins involved in budding , cytokinesis , and signal transduction , which are well known to be processes regulated by phosphorylation , were highly phosphorylated in the three yeast species . We can conclude , therefore , that although individual kinase–substrate interactions might change quickly , phosphorylation levels within specific processes are highly conserved , even for the relatively large divergence times considered here . Importantly , we could also use this information to discover functions and complexes that show significant changes in the average number of phosphosites per protein across species ( Figure 1B and 1C and Methods ) . We identified 12 functional groups ( e . g . , cellular respiration , cell budding , pseudohyphal growth , vitamin metabolic process ) and nine complexes ( e . g . , clathrin-associated complex , outer kinetochore complex , H+ transporting v-ATPase , etc . ) with significant cross-species variation in levels of phosphorylation ranging from 1 . 5 to 7 times the average number of phosphosites as expected by orthology . For example , we could detect ten phosphosites in the conserved proteins of the outer kinetochore complex in S . cerevisiae , whereas only three were found in Sc . pombe , which was close to four times less than expected by orthology . A potential pitfall of analyzing phosphorylation levels as the number of phosphosites per functional group is that it may miss cases where phosphorylation levels within that group of proteins remain the same across species , but the exact proteins that are phosphorylated have diverged . One striking example of this is the phosphorylation of the pre-replication complex . Although the level of phosphorylation of this complex is conserved , the proteins that are phosphorylated have changed . For this complex , phosphorylation of the S . cerevisiae orthologs in Sc . pombe is less conserved than expected by chance ( p-value <0 . 005 , hypergeometric distribution ) , and vice-versa ( p-value <0 . 04 , hypergeometric distribution ) . The orthology definitions used include cases of one-to-one assignments and also cases of one-to-many assignments due to species-specific gene duplication . For this reason , the functional groups mapped by orthology from S . cerevisiae to the other fungal species do not necessarily have the same number of proteins in all species . Because of this , gene duplication could account for some of the observed changes in the average number of phosphosites per protein across species . To examine this , we analyzed the functions and complexes showing significant differences in phosphorylation levels that also show significant differences in the number of proteins assigned to them ( Figure 1B and 1C ) , which applied to six out of 19 functional groups . However , even in these cases , it is clear that changes in the total numbers of proteins do not explain the changes in phosphorylation levels . For example , the expansion of a respiratory chain complex in C . albicans does not explain the observed differences in phosphorylation across the three species . Because protein abundance biases and protein duplication account for only a small fraction of the observed variation in phosphorylation , we conclude that most of the changes in the groups identified here are due to the evolutionary gain or loss of phosphorylation sites . Protein complexes are stable assemblies of proteins that cooperate in the cell to carry out specific functions , many of which are conserved across species [32] . We used the results presented above to ask whether the regulation of protein complexes by phosphorylation diverged across the three species . Compared to the broader ontological groups defined above ( that may encompass more than one pathway ) , changes in the regulation of complexes—given their smaller size—might be more readily explained by changes in regulation by one or a few kinases . To study the evolution of phosphoregulation and complement the experimentally derived MS results , we developed a sequence-based phosphorylation propensity predictor and a kinase–substrate predictor that allowed us to study lineage specific divergence of kinase–substrate relationships ( see Methods ) . To predict the phosphorylation propensity from protein sequence , we used two different approaches: ( 1 ) likelihood ratios ( LRs ) for kinase motif enrichment and spatial clustering following the method of Moses and colleagues [33] and ( 2 ) phosphosite propensity predictions using the GPS 2 . 0 algorithm [34] . For each fungal protein sequence , we define the phosphorylation propensity either as the sum of all kinase LRs using the motif enrichment method or the sum over all phosphosite likelihoods using the GPS 2 . 0 algorithm . We benchmarked these two approaches using the known phosphoproteins of S . cerevisiae and we use the area under the receiver operating characteristic ( ROC ) curve ( AROC value ) as a measure of the method's performance . We obtained an AROC value of 0 . 69 for the motif enrichment method and 0 . 73 using GPS 2 . 0 . For each protein complex , we used the prediction method that would best predict the phosphoproteins experimentally determined for S . cerevisiae , C . albicans , and Sc . pombe . In parallel to this , we trained a naïve Bayes predictor for kinase–substrate interactions for S . cerevisiae . We used a set of features that include the number of shared ( physical and genetic ) interaction partners between a kinase and a putative substrate , the existence of a phosphosite matching the substrate recognition motif of the kinase , etc . ( see Methods ) . We obtained an AROC value of 0 . 84 for this predictor using as a benchmark a set of curated kinase–substrate interactions . For each divergent complex identified above , we first calculated the predicted phosphorylation propensity for the orthologous group across 11 ascomycota species . In addition , we tried to determine the most likely kinase ( s ) responsible for the observed phosphorylation of each complex across the three species in a three-step process: ( 1 ) we use the kinase–substrate predictor to rank all 116 S . cerevisiae protein kinases according to the likelihood that they phosphorylate the members of this complex in S . cerevisiae; ( 2 ) we retain the top five kinases and for each we predict the phosphoproteins observed in the three species ( S . cerevisiae , C . albicans , and Sc . pombe ) using their substrate recognition motif and the motif enrichment method; ( 3 ) we then assume that the kinase that best predicts the phosphoproteins would be the most likely regulator . We present below the results obtained for the pre-replication complex and for the clathrin-associated complex . The analysis of the remaining complexes as well as individual kinase–substrate predictions for S . cerevisiae can be found in Protocol S1 and Dataset S2 . The evolution of cell-cycle control has previously been studied by analyzing gene expression data for multiple species [35] . One key finding from this study was that although there was little overlap between the sets of genes that are periodically expressed in different species , a similar physiological outcome is maintained . That is , the timely assembly of the different cell-cycle complexes is attained by regulated expression of one component , but the exact protein that is periodically expressed may differ across species [35] . These same authors also found a significant association between genes that are periodically expressed and under kinase regulation , showing that there is significant co-evolution of gene regulation and protein phosphorylation [35] . As noted above , our results support their conclusions at the level of post-translational regulation of the pre-replication complex . Although the pre-replication complex as a whole shows similar levels of phosphorylation across three yeast species , the specific phosphoproteins detected appear to have diverged significantly . The MCM and ORC complexes are a part of pre-replication complex and are among the few examples were evolutionary studies of phosphoregulation have been conducted [22] . Regulation by phosphorylation of these complexes is also well studied , making them a good starting point for the evaluation of our methods . Among the top five kinases predicted to regulate these complexes in S . cerevisiae ( Rad53p , Cdc28p , Dun1p , Fus3p , and Cla4p ) , Cdc28p , a well-known regulator of these complexes [36]–[39] , was predicted to best explain the phosphorylation pattern observed ( Figure 2B ) . For S . cerevisiae we correctly predicted phosphorylation by Cdc28p of Mcm3p and Mcm4p [37] , [39] . Although it was not apparent from the calculated Cdc28 phosphorylation propensity , we do find conserved Cdc28 motifs in Orc6p that would predict known regulation patterns [38] . Importantly , we correctly predict the divergent regulation of Mcm3 by Cdc28 . This interaction displays high phosphorylation propensity in the Saccharomyces lineage that it is not observed in more divergent species [22] . The phosphorylation event regulates nuclear localization of the whole MCM complex in S . cerevisiae by masking nuclear localization and export sequences that work in coordination with localization signals in Mcm2 [22] , [39] . Interestingly , we predict a strong N-terminal cluster of Cdc28p target sites in C . albicans' Mcm2 , which overlaps with an experimentally observed phosphorylation and shows strong homology to a conserved nuclear localization sequence . Therefore we postulate that in C . albicans , the localization of the MCM complex might be regulated via phosphorylation of Mcm2p instead of Mcm3p as occurs in the Saccharomyces lineage . However , there are known regulatory events that we fail to predict . We do not correctly predict the known Cdc28p regulation of Orc2p [38] , nor do we place Cdc7p among the top five most likely kinase regulators of this complex , although it is known that it phosphorylates Mcm4p [40] and Mcm2p [41] . We think further experimental work in cross-species phosphoregulation of protein complexes will create better benchmarks and further improvements in these computational methods . Having established that we could use our approach to predict known kinase–substrate interactions and a known case of evolutionary divergence of phosphoregulation , we used this method to analyze complexes that show divergent levels of phosphorylation across species ( Figure 1C and Protocol S1 ) . In Figure 3A , we show the experimentally determined phosphoproteins and the predicted phosphorylation propensity of the clathrin-associated AP-1/2/3 complexes . The top five kinases predicted to be associated with the S . cerevisiae complexes were Cka1p , Yck1p , Yck2p , Cka2p , and Cdc7 . Contrary to the example above , the observed phosphorylations could be explained equally well by the binding specificity of the five kinases so we selected the top kinase associated with the complex in S . cerevisiae , casein kinase type I ( both isoforms Yck1 and Yck2 ) as the most likely kinase responsible for the observed phosphorylations ( Figure 3B ) . The resulting predictions are consistent with observations made in other species . For example , we predict a conserved casein kinase I regulation of the C terminus of APL6 and , in fact , this phosphorylation event has been observed in human cells [42] . Our results also suggest that a kinase casein isoform regulates the miu2-like subunit of AP-1 ( APM2 ) with highly conserved target motifs at amino acids 150 to 160 . Again , it is known that phosphorylation of the human miu2 isoforms of the AP2 complex at Thr156 can regulate the complex [43] . Finally our analysis points to a casein kinase I-dependent phosphorylation of the C terminus of APL2 that is not observed in the Saccharomyces lineage , but we predict it to occur in the yeast species that diverged from budding yeast prior to the whole-genome duplication event ( Figure 3B ) . These results show that the new phosphorylation information provided here , coupled with our computational approach , can confirm known cases of conserved and diverged kinase–substrate interactions , and predict new ones . A detailed analysis of the remaining complexes is provided in Protocol S1 and can provide a starting point for future evolutionary studies of protein-complex regulation by protein kinases . The results presented above show that the changes of phosphorylation during evolution might contribute significantly to evolutionary divergence , possibly at levels similar to transcriptional regulation . One could postulate that , if a large fraction of the phosphorylation sites played no significant functional role , then the observed changes in phosphorylation could represent mostly neutral variation with no impact on species fitness . In contrast , if most changes in phosphorylation observed here have an impact on fitness , then we would expect also to see significant divergence of protein kinase function . In order to test for functional changes , we decided to study the genetic interactions of protein kinases in two different yeast species ( S . cerevisiae and Sc . pombe ) . Two genes are said to genetically interact if concurrent mutations in these genes produce phenotypes that are different from the expected combined effect of the individual mutations [44] . These epistatic or genetic interactions are used as way to identify functional relationships between genes . We assume that there is a correlation between the conservation of a gene's function in two different species with the conservation of its genetic interactions . We used quantitative genetic interaction screening to ask whether protein kinases do indeed evolve new functions more rapidly than average genes . We excluded from this analysis kinases that phosphorylate cellular components other than proteins ( e . g . , lipid kinases ) . We assembled genetic interaction maps for S . cerevisiae and Sc . pombe from the BioGRID database [45] and quantitative genetic interactions obtained with the E-MAP technology [46]–[50] , [51]–[53] . To expand the total number of genetic interactions that we could compare across the two species , we performed additional assays in Sc . pombe and S . cerevisiae using the E-MAP method as previously described , adding an additional 2 , 000 genetic interactions to the dataset [49] , [50] ( data provided in Dataset S3 ) . In total we compiled a set of 5 , 322 pairs of genes that genetically interact in S . cerevisiae that were also tested in Sc . pombe ( see Figure 4 ) . We observed that on average , 14% of the S . cerevisiae genetic interactions ( 761 pairs ) were conserved in Sc . pombe , whereas only 8% ( 38 out of 472 ) of genetic interactions with protein kinases and 4% ( 6 out of 141 ) of genetic interactions with TFs are conserved . This shows that indeed the functional roles of protein kinases and TFs are less conserved than average genes ( p-value = 5×10−6 and 6×10−5 , respectively , with hypergeometric distribution ) . We have previously observed that positive genetic interactions between genes coding for physically interacting proteins are much more conserved than for average gene pairs [48] . However , we found that genetic interactions among genes coding for physically interacting kinase–protein pairs are significantly less conserved than those for all physically interacting partners ( p-value = 0 . 007 with hypergeometric distribution ) . This trend is stronger for genetic interactions among transient physical interactions partners ( p-value = 2×10−7 with hypergeometric distribution ) . Interactions were defined as transient based on the experimental methods used ( see Methods ) . Finally we observed that genetic interactions between kinase-protein interaction partners and between TF–promoter interactions ( from ChIP–chip experiments ) show similar levels of conservation ( 8% ) . We conclude that kinase–substrate interactions change at a fast evolutionary rate and that this leads to functional divergence that is more rapid than for average genes . According to our results , protein kinases diverge in function at a similar rate ( when testing direct physical targets ) or somewhat slower ( when testing all genes ) than TFs . Therefore we suggest that protein kinases , given their high regulatory potential and rapid divergence in their interactions , are an important source of phenotypic diversity .
Comparing cellular interaction networks across different species is crucial for understanding how DNA variability drives functional potential . Changes in the regulation of gene expression have , to date , been seen as a prime mechanism leading to phenotypic divergence . This stems from the early studies of molecular evolution and a large body of work on the study of the evolution of morphology [54] . Recently , methods have been developed to detect protein–protein interactions in high-throughput fashion [12]–[16] . The resulting protein interaction networks have been studied alongside an increasing number of solved protein complex structures to shed new light into the evolutionary potential of protein–protein interactions . It has been observed that protein complexes are indeed well conserved across species , and changes in complex formation occur typically by duplication or deletion of complex components , rather than through rewiring of existing proteins [32] , [55] . Still , on average , protein interactions were observed to change at a fast rate after gene duplication [19] , [20] . This apparent discrepancy can be explained by noting that transient interactions of lower specificity , like interactions mediated by short peptide motifs , are much more likely to change than stable interactions are [20] , [21] . We hypothesized that protein kinases , given their crucial regulatory role and transient interactions , could be an important source of phenotypic variability across species . To study this , we have experimentally determined phosphorylation sites by MS analysis for three yeast species ( S . cerevisiae , C . albicans , and Sc . pombe ) spanning 400 to 600 million years of evolution . We have used this information to estimate the global rates of change of phosphoproteins . Based on these rates , our estimated kinase–substrate interaction changes are within an order of magnitude of previous estimates for gain or loss of interaction after gene duplication . Furthermore , kinase–substrate interaction evolution is at most two orders of magnitude slower than TF–promoter interactions . These observations are further supported by the comparative analysis of quantitative genetic interactions between S . cerevisiae and Sc . pombe genes . We observed a lower-than-average conservation of genetic interactions for protein kinases and TFs , suggesting that the observed divergence of phosphorylation correlates with functional changes of protein kinases . Interestingly the level of conservation of genetic interactions between kinases and their interaction partners is similar to that observed for TFs and the genes they bind to . However , it should be noted that the current overlap between genetic interactions and physical interactions for kinases and TFs is still small . Also , the different nature of physical interaction ( protein–DNA versus protein–protein ) could potentially result in differences in the genetic interactions observed between interacting partners . For these reasons , further studies are needed to determine the exact relative functional divergence rate . Our results indicate that there is a high level of conservation of phosphorylation for different functional groups across the broad time scale studied . This would mean that even if individual kinase–substrate interactions differ , the overall phosphorylation levels of a given functional group might be strongly predicted by homology . It is conceivable that this conservation of phosphorylation levels is maintained by physical proximity of kinases and substrates due to shared interaction partners or sub-cellular localization . Given that the in vivo targets of a protein kinase are determined , in large part , by factors other than its own substrate recognition ( i . e . , gene expression , localization , scaffolding , etc . ) [56] , it is possible that differential association to kinases serves to maintain the levels of phosphorylation among different functional groups . In this study , we have combined experimental phosphorylation information with computational methods to predict kinase–substrate interactions and their evolution . We used this approach to study eight protein complexes that show significant changes in phosphorylation and we predict putative kinase regulators responsible for these observed changes . Analysis of well-studied pre-replication complexes showed that we predict known examples of conserved and divergent phosphoregulation . In addition to our analysis , the study of human phosphorylation sites has recently shown that highly conserved phosphorylation networks are associated to disease ( C . S . H . Tan and R . Linding , personal communication ) . These results highlight the importance of studying the evolution of kinase regulation and our work offers a starting point for further studies . Selection pressure acts on the preservation or acquisition of phenotypes , rather than the mechanisms by which these phenotypes are implemented . A picture is emerging of highly conserved modules ( i . e . , complexes ) that are regulated and organized in different ways in different species . For instance , the conservation of timed assembly of cell-cycle complexes , regulation of mating , or co-expression of ribosome subunits may be conserved , although details of the implementation diverges in different species [23] , [35] , [48] , [57] . Similarly we show here that kinase–substrate interactions have a large potential to change , and that care should therefore be taken in projecting information about these interactions using cross-species homology . Importantly , kinase–substrate interactions are just one type of essential transient regulatory interaction [24] , and recent work by Neduva and colleagues point to the existence of other undiscovered interactions mediated by small linear peptide motifs [58] . There has been a long-standing debate , in particular in the field of developmental biology , as to the types of adaptive mutations that contribute most to phenotypic changes [54] , [59] . This debate has tended to focus on studies of the evolutionary history of individual biological systems . In contrast , we have used large-scale phosphorylation and genetic data to place quantitative bounds on the relative rate of change of TF–gene and kinase–substrate interactions . We believe that our approach , that of combining physical and genetic interaction mapping on a large scale across multiple species , will allow us to systematically probe the evolutionary potential of different cellular components .
Proteins were precipitated from yeast lysates using TCA on ice and washed once with acetone at 4°C . Protein pellets ( approximately 24 mg protein ) were resuspended in 3 ml of freshly deionized 8 M urea . Samples were incubated for 1 h at 57°C with 2 mM Tris ( 2-carboxyethyl ) phosphine hydrochloride to reduce cysteine side chains , these side chains were then alkylated with 4 . 2 mM iodoacetamide in the dark for 45 min at 21°C . The mixture was diluted 8-fold with 25 mM ammonium bicarbonate and 1% ( w/w ) modified trypsin ( Promega ) was added . The pH was adjusted to 8 . 0 and the mixture was digested for 12 h at 37°C . The digests were desalted using a C18 Sep Pak cartridge ( Waters ) and lyophilized to dryness using a SpeedVac concentrator ( Thermo Electron ) . Phosphorylated peptides were enriched using an ÄKTA Purifier . Peptides run over an analytical guard column ( Upchurch Scientific ) loaded with 5-µm titanium dioxide beads ( GL Sciences ) . Peptides were re-suspended in 750-µl wash solution ( 35% acetonitrile , 200 mM NaCl , 0 . 3% TFA ) , and the enrichment was done on three separate 250-µl aliquots . Each aliquot was injected over the titanium dioxide column , with an additional 3 . 9 ml wash solution to remove non-phosphorylated peptides . This was then followed by 3 . 5 ml of rinse solution ( 5% acetonitrile , 0 . 1% TFA ) . Phosphorylated peptides were eluted from the titanium dioxide column using 1 ml of elution solution ( 1 M KH2PO4 ) . High-pH reverse-phase chromatography was performed using an ÄKTA Purifier ( GE Healthcare ) equipped with a 250-×4 . 60-mm column packed with 3-µm Gemini C18 resin ( Phenomenex ) . Phosphopeptide-enriched fractions were loaded onto the column in 2 mM ammonium trifluoroacetic acid , pH 9 . 5 ( buffer A ) . Buffer B consisted of 2 mM ammonium trifluoroacetic acid in acetonitrile . The gradient went from 1% B to 60% B over 20 ml , and from 60% B to 100% B over 5 ml . Between 30 and 40 fractions were collected and dried down using a SpeedVac concentrator . Samples were desalted using C18 ziptips ( Millipore ) . Individual fractions were separated using a 75-µm×15-cm reverse-phase C18 column ( LC Packings ) at a flow rate of 350 nl/min , running a 3–32% acetonitrile gradient in 0 . 1% formic acid over 1 h on an Agilent 1100 series HPLC equipped with an autosampler ( Agilent Technologies ) . The LC eluent was coupled to a micro-ionspray source attached to a QSTAR Elite mass spectrometer ( Applied Biosystems ) . Peptides were analyzed in positive ion mode . MS spectra were acquired for 1 s . For each MS spectrum , the two most intense multiple charged peaks were selected for generation of subsequent collision-induced dissociation MS . For precursor ion selection , the quadrapole resolution was set to “low , ” which allows for transmission of ions within approximately 2 mass to charge ( m/z ) units of the monoisotopic mass . The collision-induced dissociation energy was automatically adjusted based upon peptide charge and m/z ratio . A dynamic exclusion window was applied which prevented the same m/z from being selected for three minutes after its initial acquisition . Data were analyzed using Analyst QS software ( version 1 . 1 ) and MS/MS centroid peak lists were generated using the Mascot . dll script ( version 1 . 6b18 ) . The MS/MS spectra were searched against the entire Uniprot database of the respective species ( downloaded 19 April 2007 ) using the following parameters . Initial peptide tolerances in MS and MS/MS modes were 200 ppm and 0 . 2 Da , respectively . Trypsin was designated as the enzyme and up to two missed cleavages were allowed . Carbamidomethylation was searched as a fixed modification . Oxidation of methionine , protein N-terminal acetylation , pyro-glutamine formation , and phosphorylation of serine/threonine/tyrosine residues were allowed as variable modifications . All high-scoring peptide matches ( expectation value <0 . 01 ) from individual LC-MS/MS runs were then used to internally recalibrate MS parent ion m/z values within that run . Recalibrated data files were then searched with a peptide tolerance in MS mode of 50 ppm . The false-positive rates were estimated by conducting the search using a concatenated database containing the original Uniprot database as well as a version of each original entry where the sequence has been randomized . Functional groups for S . cerevisiae were defined using the gene ontology mapping provided by SGD ( http://www . yeastgenome . org/ ) . The complexes definitions for S . cerevisiae were obtained from the MIPS database ( http://mips . gsf . de/ ) . For the other fungal species studied , complexes and functional groups were defined by transferring these annotations using the orthology definitions from the Synergy algorithm [1] . For the remainder of this methods section we will use “functional group” to describe both the gene ontology groups and complexes for brevity . In order to calculate the global rate of change of phosphoproteins in S . cerevisiae with respect to another species , we considered only the set of orthologous proteins between species i and S . cerevisiae ( denomined as ortProteins and ortKinases ) . We assumed that the coverage ( c ) of our compiled set of S . cerevisiae phosphoproteins is 92% , the largest value obtained from leaving out one of the previously published sets . We define the number of expected phosphoproteins ( “expPhospho” ) the number of orthologous phosphoproteins in species i and the conserved phosphoproteins ( “consPhospho” ) the number of ortologous phosphoproteins in species i detected as phosphorylated in S . cerevisiae . The number of divergent phosphoproteins ( “divPhospho” ) was thus defined as the difference: ( expPhospho×c ) −consPhospho . We defined the rate of change of S . cerevisiae phosphoproteins in reference to species i as:where divergenceTime is the time since the last common ancestor between S . cerevisiae and species i . Similarly , we defined the rate of change of kinase–substrate interactions as:where N is the assumed number of kinase–substrate interactions changed with every change in total phosphoproteins . We calculated similar rates for the change of TF–gene interactions using available information from the literature [17] , [18] , [31] . Detailed values for all species studied are available in Protocol S1 . For each species and for each functional group defined above , we determined the average number of phosphosites per protein . For this analysis , we used the phosphosites determined in this study and additional studies for S . cerevisiae and Sc . pombe growing in exponential phase [10] , [29] ( excluding condition-specific studies ) . For each species , we then normalized the results of each functional group by the average number of phosphosites per protein for the whole proteome . We define this normalized value as the phosphorylation level and used this measure for all the functional analysis presented in this manuscript . In similar fashion , we also calculated the fraction of phosphoproteins per functional group normalized by the fraction of phosphoproteins per proteome in each species . To search for significant cross-species differences in the average number of phosphosites per protein , we defined for each functional group and each species a measure of comparative phosphorylation ( compPhos ) as the relative contribution to the sum across the three species . For species i:where norPhos is the normalized average fraction of phosphosites per protein for that functional group in species i , as defined above , and n the set of three yeast species studied here . Defined in this way , functional groups with the same average fraction of phosphosites per protein , in the three species , would have a comparative phosphorylation value matching exactly 1/3 in the three species . As expected from the high-cross species correlation shown in Figure 1 , most of the functional groups show very similar levels of phosphorylation across species with an average comparative phosphorylation value near 0 . 33 for the three species . We then defined as a significant change comparative phosphorylation values that significantly deviate from 0 . 33 . For this purpose , we calculated z-scores and selected functional groups that had , for at least one species , z-score greater than 1 . 6 or smaller than −1 . 6 corresponding to significant changes in phosphorylation levels ( p-value <0 . 05 ) . z-scores for each functional group are provided in Protocol S1 . In order to find complexes with significant differences in average number of phosphosites , we considered only 28 complexes that had at least ten protein subunits to discard large variations due to small complex sizes . We used two different approaches to predict phosphorylation from sequence for all fungal proteins studied: ( 1 ) LRs for kinase motif enrichment and spatial clustering; ( 2 ) phospho-site propensity predictions from GPS 2 . 0 [34] . The LRs for kinase motif enrichment and spatial clustering were determined following the method of Moses and Colleagues [33] . We used kinase substrate motifs for 116 protein kinases predicted by Predikin [60] , including for each kinase , motifs that vary from the originally published by addition of one or two fully degenerate positions . For each sequence , the final prediction score was defined as the sum of the LRs of all kinases . For the second approach , we used GPS 2 . 0 to predict phosphorylation sites within all the fungal sequences studied . The final protein phosphorylation prediction score was defined as the sum over all the phosphorylation sites likelihood scores for any given protein . The two prediction scores were obtained for all protein sequences in the genomes of S . cerevisiae , S . bayanus , S . paradoxus , S . castellii , Kluyveromyces lactis , K . waltii , Debaryomyces hansenii , C . albicans , Yarrowia lipolytica and Sc . pombe . The prediction scores were benchmarked using the known phosphoproteins of S . cerevisiae . We plotted the ROC curve and determined the area under the ROC ( AROC ) curve for both methods ( see Protocol S1 ) . The LR method predicts phosphoproteins with an AROC of 0 . 69 while the GPS 2 . 0 method predicts phosphoproteins with an AROC of 0 . 73 . For each complex , we selected the method that could best predict the phosphoproteins determined for S . cerevisiae , C . albicans , and Sc . pombe for that complex . The exact AROC values for each complex are available in Protocol S1 . In order to predict kinase–substrate interactions for S . cerevisiae proteins , we used a naïve Bayes predictor integrating sequence based prediction of kinase interactions with available protein and genetic interaction data defined in the BioGRID database [45] version number 2 . 0 . 43 . Four features were used in the predictor: ( 1 ) substrate motifs enrichment LRs in putative target as determined above; ( 2 ) presence or absence of at least one phosphosite matching the kinase motif; ( 3 ) number of orthologs ( from 0 to 2 ) in C . albicans and/or Sc . pombe with at least one phosphosite matching the kinase motif; and ( 4 ) the number of shared physical or genetic interactions partners in common between the kinase and the putative target . These four indicators were integrated using a naïve Bayes algorithm , and its performance was evaluated by AROC using a set of 472 kinase–substrate interactions curated from the literature [46] as our set of positive interactions . The positive set was used both as training and testing sets using a 5-fold cross-validation . The sequence-based predictions has AROC value of 0 . 63 that improves significantly with the integration of physical and genetic interaction data to an AROC value of 0 . 84 ( see Protocol S1 for ROC curves ) . To predict the kinases most likely responsible for the phosphoregulation of a protein complex , we defined the kinase-complex association score as the sum of the S . cerevisiae kinase–substrate prediction score across all the complex subunits . For each complex , we selected ( from the 116 S . cerevisiae protein kinases ) the top five kinases predicted to regulate the complex for further analysis . These five kinases were then ranked on how well their substrate specificity explains the phosphorylation pattern of the complex subunits across the three species with available phosphorylation data . The ranking was done on the AROC value for phosphorylation prediction using the kinase–substrate LRs predicted from their binding motifs as described above . Detailed results for the complexes studied are provided in Protocol S1 . Genetic interaction information for S . cerevisiae and Sc . pombe were compiled from different quantitative high-throughput studies [45]–[48] and from the BioGRID interaction database . Genetic interactions from E-MAP studies were defined as any interactions with a positive S-score greater than 2 or a negative score lower than −2 . 5 . For the genetic interactions obtained from the BioGrid database that do not contain a quantitative score we assumed that those labeled as “Synthetic Rescue” or “Phenotypic Suppression” were positive interactions , while those labeled with “Synthetic Lethality” , “Phenotypic Enhancement” , “Synthetic Haploinsufficiency” , or “Synthetic Growth Defect” were negative interactions . To increase the overlap available for cross-species analysis , we determined 634 novel strong genetic interactions in S . cerevisiae and tested an additional 1 , 293 gene pairs in Sc . pombe using the E-MAP method as previously described [50] . The final set contains 5 , 322 pairs of genes that genetically interact in S . cerevisiae that were also tested in Sc . pombe . This set is provided in Dataset S3 . A genetic interaction was considered to be conserved when the corresponding orthologs in Sc . pombe also genetically interact according to the definition defined above ( S-score >2 or S-score <−2 . 5 ) having a similar phenotypic effect ( suppression or enhancement ) in both species . Physical protein–protein interactions were obtained from BioGRID database [45] version number 2 . 0 . 43 . In order to define a subset of physical interactions enriched for transient interactions we excluded those that were labeled in BioGRID as “Affinity Capture , ” “Reconstituted Complex , ” or “Co-crystal Structure . ” We considered for our analysis 114 sequence specific transcription factors annotated in SGD database ( http://www . yeastgenome . org ) . TF–promoter interactions were obtained from Harbison and colleagues [61] . | Natural selection at a population level requires phenotypic diversity , which at the molecular level arises by mutation of the genome of each individual . What kinds of changes at the level of the DNA are most important for the generation of phenotypic differences remains a fundamental question in evolutionary biology . One well-studied source of phenotypic diversity is mutation in gene regulatory regions that results in changes in gene expression , but what proportion of phenotypic diversity is due to such mutations is not entirely clear . We investigated the relative contribution to phenotypic diversity of mutations in protein-coding regions compared to mutations in gene regulatory sequences . Given the important regulatory role played by phosphorylation across biological systems , we focused on mutations in protein-coding regions that alter protein–protein interactions involved in the binding of kinases to their substrate proteins . We studied the evolution of this “phosphoregulation” by analyzing the in vivo complement of phosphorylated proteins ( the “phosphoproteome” ) in three highly diverged yeast species—the budding yeast Saccharomyces cerevisiae , the pathogenic yeast Candida albicans , and the fission yeast Schizosaccharomyces pombe—and integrating those data with existing data on thousands of known genetic interactions from S . cerevisiae and Sc . pombe . We show that kinase–substrate interactions are altered at a rate that is at most two orders of magnitude slower than the alteration of transcription factor ( TF ) –promoter interactions , whereas TFs and kinases both show a faster than average rate of functional divergence estimated by the cross-species analysis of genetic interactions . Our data provide a quantitative estimate of the relative frequencies of different kinds of functionally relevant mutations and demonstrate that , like mutations in gene regulatory regions , mutations that result in changes in kinase–substrate interactions are an important source of phenotypic diversity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"computational",
"biology/signaling",
"networks",
"genetics",
"and",
"genomics/comparative",
"genomics",
"evolutionary",
"biology/evolutionary",
"and",
"comparative",
"genetics",
"cell",
"biology/cell",
"signaling"
] | 2009 | Evolution of Phosphoregulation: Comparison of Phosphorylation Patterns across Yeast Species |
The analysis of double-strand break ( DSB ) repair is complicated by the existence of several pathways utilizing a large number of genes . Moreover , many of these genes have been shown to have multiple roles in DSB repair . To address this complexity we used a repair reporter construct designed to measure multiple repair outcomes simultaneously . This approach provides estimates of the relative usage of several DSB repair pathways in the premeiotic male germline of Drosophila . We applied this system to mutations at each of 11 repair loci plus various double mutants and altered dosage genotypes . Most of the mutants were found to suppress one of the pathways with a compensating increase in one or more of the others . Perhaps surprisingly , none of the single mutants suppressed more than one pathway , but they varied widely in how the suppression was compensated . We found several cases in which two or more loci were similar in which pathway was suppressed while differing in how this suppression was compensated . Taken as a whole , the data suggest that the choice of which repair pathway is used for a given DSB occurs by a two-stage “decision circuit” in which the DSB is first placed into one of two pools from which a specific pathway is then selected .
In the last ten years the study of double-strand break ( DSB ) repair has become more urgent and more difficult . The urgency arises from findings that defects in DSB repair are linked to elevated cancer risks [1] and phenotypes that resemble accelerated aging [2–4] . Meanwhile , the complexity of analyzing DSB repair has increased exponentially with the number of pathways , genes , and interactions that have been discovered [5–7] . Breaks can be repaired via nonhomologous end-joining ( NHEJ ) [8]or homologous recombination ( HR ) [9] , each of which can be further subdivided into several pathways , all requiring genes for checkpoints , signaling , and effecting the repair itself . In view of this complexity , it is fortunate that the high degree of conservation in DSB repair systems makes it possible to use model organisms such as Drosophila to obtain generalizable results concerning the basic processes . In Drosophila , a large-scale screen for mutagen sensitivity has been useful in identifying DSB repair genes [10] . A variety of methods is available for studying DSB repair in Drosophila . These include the use of excision of P transposable elements to create breaks [11–23] as well as transplanted endonucleases and recombinases derived from microorganisms [24–27] . The repair reporter construct ( Rr3 ) is designed to yield simultaneous measurements of multiple DSB repair outcomes in the Drosophila germline [28] . Other such reporters have been valuable in mammalian systems [29 , 30] and yeast [31–33] . Measurements obtained with Rr3 reflect the relative usage of NHEJ , single-strand annealing ( SSA ) , and homologous repair with conversion from the homolog ( HR-h ) . They also provide further quantitative information about specific events within these pathways , including the length of conversion tracts , deletion formation , and crossing over [34–36] . Rr3 has been used to show that the relative usage of DSB repair pathways changes with developmental stage [28] . Another surprising finding was that as adult flies age , their usage of HR for repair increases in the germline at the expense of other repair pathways [34 , 36 , 37] . Studies with Rr3 also provided evidence that the Drosophila version of BLM , the Bloom syndrome gene , is needed to resolve double Holliday junctions via dissolution [35] . A major advantage of using Rr3 as opposed to measuring one pathway at a time is that the Rr3 analysis reveals not only which pathway ( s ) are inhibited by a given condition or genotype , but also which other pathways are used to compensate . This compensation is easily seen as a negative correlation between the relative usage of the various pathways [28] such that the total of the various repair outcomes remains close to 100% . Here we use Rr3 to compare the effects of mutations in 11 genes related to DSB repair . The results provide information concerning how these mutations depress specific pathways , some of which is expected from what is already known about the genes . More interestingly , the results also show how other pathways compensate for each defect . We find that when a given pathway is suppressed , the defect can be compensated in more than one way depending on which mutant gene is responsible for the suppression . This analysis provides a fresh insight into the process by which each DSB is channeled into one of many possible repair outcomes and how the DSB repair system as a whole adjusts when one or more of its components is missing .
The use of Rr3 to measure multiple outcomes of DSB repair has been described in detail elsewhere [28 , 34–36] . Briefly , DSBs are created at the recognition site of the rare-cutting endonuclease , I-SceI , located within a red fluorescent reporter gene , DsRed , and flanked by a direct repeat of 147 bp ( Figure 1A ) . The endonuclease gene , located on another chromosome , is expressed continuously and in all tissues , but we analyze only breaks that occur in the germ cells . Figure 1B shows the five distinguishable outcomes that are observed among the progeny . If repair occurs via conversion with the sister chromatid as template ( HR-s ) , the recognition site is restored , and Rr3 is available for another round of breakage and repair . The cycle can continue until one of the five measured outcomes occurs , all of which destroy the recognition site . We identify these outcomes among the offspring by scoring ( i ) visible markers and sex to determine the presence of the Rr3-derived chromosome , the endonuclease transgene , and to detect crossing over between flanking markers; ( ii ) DsRed fluorescence to indicate collapse of the duplication in all or part of the fly; and ( iii ) single-fly PCR tests in a subset of the offspring to distinguish between specific outcomes . The measured outcomes are: One indication that there is compensation between DSB repair pathways is that the sum of the outcomes remains close to 100% across experiments despite wide variation in individual frequencies . This property can be seen in Figure 2A and 2B where the five measurements from cross 2 and the three from cross 1 are plotted as stacked columns . We see that the total column heights are relatively constant , staying within the range 93%–103% for cross 2 and 93%–106% for cross 1 even though some of the individual measurements vary over a much wider range . It should be noted that there is no artificial constraint on this sum , since the SSA frequency was computed using a different subset of the offspring from those used to compute NHEJ and HR-h . Another way to see compensation between pathways is by the negative correlation between individual outcomes . The frequency of SSA for each of the 30 experiments is plotted against NHEJ for cross 2 in Figure 2C and for cross 1 in Figure 2D . Those two outcomes were selected because , as mentioned above , they are computed from different subsets of the offspring , and because they are available from both crosses . A strong negative correlation is apparent from both plots . Note also the consistency between the two crosses at the extreme ends of the distribution ( highlighted regions ) . Experiments 2 , 3 , 11 , 14 , 15 , 23 , and 24 appear at the upper left ends of both plots . These are the experiments with mutants at the lig4 or ku70 locus . Our interpretation , as discussed below , is that these two loci depress NHEJ and are compensated by SSA . At the lower right ends of both scatterplots lie experiments 4 , 5 , and 27 , which represent mutations at DmATR , mus101 , and mus301 . These three mutations are apparently defective in SSA with compensation by NHEJ . However , as discussed below , these three loci differ from each other by whether they can also be compensated by HR-h . Finally , one can detect compensation through the negative correlation between measurements in individual males . As reported previously ( see Figure 6A in [28] ) , random differences in DSB repair pathway usage between individuals can also display compensation . Figure 2E and 2F show this effect for crosses 1 and 2 with the test males from experiments 2 and 4 . The data in Table 1 and Figure 2 show that certain genotypes display conspicuous suppression of NHEJ . Experiments 2 , 3 , 11 , 14 , and 15 , which test null mutants for lig4 , show reduction in NHEJ usage from 2- to 6-fold relative to the corresponding controls . Figure 3A shows the total cross 2 NHEJ frequencies and their controls , and Figure 3B shows the same results for cross 1 . All ten of these comparisons—including five in cross 2 and five in cross 1—showed a highly significant reduction in NHEJ frequency . We conclude that NHEJ repair is reduced to less than half of its normal usage in lig4 mutants . Similarly , ku70 mutants were tested in experiments 23 and 24 . Figure 3A and 3B compare their NHEJ frequencies to their heterozygous controls from experiments 25 and 26 . These mutants also showed a clear reduction in NHEJ usage , with all four comparisons being highly significant . Therefore , ku70 mutants are similar to lig4 in lacking most of the normal NHEJ repair capability . The results also show a maternal effect contribution affecting NHEJ for both lig4 and ku70 . By comparing experiment 2 with 3 , we see significantly less NHEJ usage when lig4 mutant males had mutant mothers in cross 2 ( p = 0 . 009 ) ( Figure 3A ) and in cross 1 ( p = 0 . 004 ) ( Figure 3B ) . The ku70 mutant males also showed less NHEJ usage when their mothers were ku70−/− by comparing experiments 23 and 24 in crosses 2 and 1 ( p = 0 . 070 and p = 0 . 018 , respectively ) ( Figure 3A and 3B ) . The ku70 mutant test males from cross 1 of experiment 24 also showed mosaicism in their eye color , which was not present in any other class . We interpret this mosaicism as a result of somatic NHEJΔ events . In addition to the results in Table 1 , we also used cross 1 to analyze test males with an extra copy of Ku70 , but no differences relative to the controls were detected . Despite the pronounced drop in “normal” end-joining repair for lig4 and ku70 mutants , there was no such decrease for the end joining accompanied by long deletions ( Figure 3C and 3D ) . In fact , experiments 14 and 24 showed significant increases in NHEJΔ ( p = 0 . 002 and p = 0 . 043 , respectively ) for cross 1 . These deletion events could result when the more typical NHEJ process fails . Unexpectedly , we found that the reduction in NHEJ frequencies in lig4 and ku70 mutants was entirely compensated by an increase in SSA , as seen in Figure 3E and 3F . All ten of the relevant comparisons for lig4 were highly significant , and three of the four comparisons for ku70 were also significant . In contrast , there was little or no compensation by HR-h , as shown in Figure 3G . In fact , four of the seven comparisons actually show a decrease in HR-h frequency relative to their controls , although in the case of experiment 14 this decrease is most likely a result of the DmBlm mutation also present in that genotype . We conclude that SSA provides all or nearly all of the compensation for the loss of NHEJ in lig4 and ku70 mutants . It should be emphasized that this conclusion applies only to breaks within the Rr3 construct where an opportunity for SSA is provided via the 147-bp duplication . The finding that SSA but not HR-h can compensate for the reduction in NHEJ in this class of mutations suggests that there is a pool of DSBs that can “choose” between NHEJ and SSA , but not HR-h . We shall refer to this hypothetical pool of breaks as pool 1 in subsequent discussions . HR-h was significantly reduced in the homozygotes of DmBlm , rad54 , and rad51 as well as in the heterozygotes of rad51 . These effects are seen in Table 1 and Figure 4A and 4B . The most pronounced reduction was in the rad51 homozygotes that reduced long-tract HR-h more than 100-fold , and eliminated short HR-h completely . Null homozygotes of DmBlm and rad54 also reduced both long- and short-tract HR-h ( Figure 4A and 4B and [35] ) . Homozygotes for a weak hypomorphic allele of top3α behaved in a qualitatively similar way to DmBlm , as reported previously [35] . Interestingly , rad54 did not behave analogously to the other HR-suppressing mutants in terms of how the reduction in HR-h was compensated by other pathways . The rad54 homozygotes showed significant increases in both SSA ( Figure 4E ) and NHEJ ( Figure 4C ) whereas SSA provided essentially all of the compensation in the DmBlm and rad51 mutants ( Figure 4C and 4E ) . The existence of two mutations , DmBlm and rad51 , where suppression of HR-h is compensated solely by SSA , suggests that there is a pool of DSBs where the choice of repair pathways is limited to HR-h and SSA . We shall refer to this hypothetical group of breaks as pool 2 in discussions below . This compensatory increase in SSA was pronounced in cross 2 but weak or nonexistent in cross 1 where there was no opportunity for HR-h ( Figure 4E versus 4H ) . This relative lack of effect in cross 1 implies that the primary phenotype of these mutants is the drop in HR-h , whereas the concomitant increases in SSA seen with all three mutants may be considered an indirect effect . The smaller cross-1 increase in SSA seen in homozygous rad51 mutants ( Figure 4H ) can be interpreted as an indirect consequence of a reduction of conversion off the sister chromatid , HR-s . This interpretation relies on the “decision circuit” model discussed below . All four mutant types showed an increase in long-deletion NHEJΔ repair ( Figure 4D and [35] ) . However , the overall frequency of these deletion events was insufficient to provide substantial compensation for the lack of HR-h . Instead we suggest that the increase in long deletions is the result of aberrant HR-h repair . Consistent with this hypothesis , note that the increase in deletions occurs even in the DmBlm and rad51 mutants where normal NHEJ is not enhanced . In addition , there were no significant changes in NHEJΔ frequencies in cross 1 where HR-h cannot occur ( Figure 4G ) . In addition to the measures reported in Table 1 , we also recorded the frequency of crossing over in cross 2 between markers flanking Rr3 for all 30 genotypes shown in the table . These values are not included in the table , because the only significant differences relative to the controls occurred in the DmBlm and top3α mutants that were reported previously [35] . Finally , Figure 4F shows that rad54 and rad51 mutants both showed a decrease in NHEJ in cross 1 . This effect stands in contrast to the result in cross 2 where rad51 had no effect on NHEJ and rad54 actually had a highly significant increase ( Figure 4C ) . In the case of rad51 this difference between crosses 1 and 2 can be explained in terms of the “decision circuit” model discussed below . We found that SSA was suppressed in mutants for DmATR , mus101 , mus301 , and , to a lesser extent , mei-9 . This suppression is apparent in both crosses , as seen in Figure 5A and 5D . The suppression of SSA was highly significant in a statistical sense , but note that none of the mutant genotypes reduced SSA to less than 67% of its control value . This result contrasts sharply with the mutants that suppress NHEJ and HR-h , sometimes reducing these outcomes to small fractions of their normal values ( Figures 3 and 4 ) . The effect of mei-9 ( Rad1p ) in suppressing SSA is only apparent in experiments 6 and 9 where the endonuclease is supplied from insertion ubiquitin-driven I-SceI endonuclease ( UIE ) -72C on chromosome 3 and not in experiments 17 , 18 , or 19 where the endonuclease source is insertion UIE-5B on the X chromosome . As noted previously [28] , different endonuclease sources can result in subtle differences in the timing and frequency of DSB formation . Such differences are more likely to be critical in the case of mei-9 whose effect on SSA is relatively weak . All four mutations utilized NHEJ to compensate for the drop in SSA . The increase in NHEJ was significant for DmATR , mus101 , and mus301 in both cross 2 ( Figure 5B ) and cross 1 ( Figure 5E ) and for mei-9 in cross 2 ( experiment 9 , Figure 5B ) . In addition , DmATR and mus301 also used HR-h to compensate ( Figure 5G and 5H ) , whereas no such compensation occurred in mutations for mus101 or mei-9 . This compensation by HR-h occurred primarily by long-tract conversion events ( Figure 5G ) rather than short-tract HR-h that was only significant for mus301 ( Figure 5H ) . One possible explanation for this specificity is that some DSBs , which would have been repaired via SSA were it not for the lack of DmATR or mus301 function , had undergone extensive gap widening before being repaired via HR-h . We also observed a significant increase in the NHEJΔ outcome for DmATR and mus101 in cross 2 ( Figure 5C ) and for mus301 in both crosses ( Figure 5C and 5F ) . Overall , the list of mutations that showed an increase in deletion formation includes both mutations that suppress NHEJ ( lig4 and ku70 ) , all four of those that suppress HR-h ( DmBlm , top3α , rad54 , and rad51 ) , and three of the four that suppress SSA ( DmATR , mus101 , and mus301 ) . No mutation showed a decrease in NHEJΔ frequency . We conclude that NHEJΔ can be increased by a wide range of mutational changes but not reduced . We interpret this finding to indicate that failure of repair via any of the pathways can result in long deletions , but that such deletions are not formed in the normal course of any pathway . We performed further tests on some of the mutants in addition to the Rr3 assays in Table 1 . For five of the loci we measured crossover frequencies related to DSB repair at an I-SceI cut site other than Rr3 . This cut site , located at cytological position 50C and described previously [35] , differs from Rr3 by the absence of a flanking duplication suitable as a substrate for SSA repair . Therefore , HR-h competes only with NHEJ . The results ( Figure S2 ) showed crossing over in mei-9 mutants was reduced to less than one-fourth the control levels . The other mutants tested , lig4 , mus81 , DmATR , and mus101 , showed no significant difference from the controls . Our interpretation is that mei-9 mutations are defective for some aspect of the repair process in which double Holliday junctions are formed and resolved via strand exchange as opposed to dissolution . The relative rarity of crossing over during DSB repair in Drosophila suggests that this sequence of events is not a major pathway in wild-type flies . We tested the ku70EX8 deletion shown in Figure S1 to verify that ku70 was mutated and to characterize other effects of the mutation . Figure S3 shows that homozygotes for this deletion are hypersensitive to the chemical mutagen methyl methanesulfonate and that this sensitivity is rescued by a 13-kb Drosophila Ku70 transgene [38] . Similarly , Figure S4 shows the same homozygotes may also be mildly hypersensitive to gamma irradiation and that the transgene also restores them to normal sensitivity . Finally , we noticed that many of the ku70EX8 homozygotes had an excess of macrochaetae ( large thoracic bristles ) that were deformed or reduced in size , and this phenotype was rescued by Ku70 transgenes ( Figure S5 ) . This macrochaetae phenotype may be similar to that seen by Brodsky et al . in mutants of mus304 [39] . The authors of that study interpreted the phenotype as the result of somatic mosaicism for haploinsufficiency at any of the Minute loci owing to excess production of deletions during development . This interpretation is consistent with the excess production of NHEJΔ deletions seen in our ku70 mutant tests ( Figure 3D ) .
Comparison of the relative usage of DSB repair pathways has been valuable in understanding oncogenic events in mammals [29 , 30] and screening for repair mutations in yeast [32] . We have made use of the Rr3 assay to detect changes in the mix of DSB repair pathway usage during development [28] , to identify age-related changes [34] and to analyze the effects of specific repair loci [35] . Here we apply the approach to a broad spectrum of repair mutations to compare their effects on the DSB repair system as a whole . By measuring multiple repair outcomes from a single pool of DSBs we can gather information that would be more difficult to obtain with an experimental system in which only one type of outcome is measured . For example , McVey et al . [18] found that Drosophila lig4 mutants had no significant drop in NHEJ repair , whereas our data show up to a 4-fold decrease ( Figure 3A and 3B ) . Several factors could contribute to this discrepancy . The enhanced sensitivity obtained when each measured DSB repair outcome is compared directly with competing outcomes provides a clear benefit for the Rr3 approach . Other differences are the nature of the endonucleases , P transposase versus I-SceI , and that the breaks in Rr3 can be repaired by SSA , which may allow recovery of some DSBs that would otherwise have been lost through cell death . The Rr3 analysis provides two pieces of information about each mutation: which pathway ( s ) are suppressed and which are used for compensation ( Table 2 ) . By classical genetic inference , suppression of a pathway in a mutant is taken to indicate a primary role of the gene . For some of the genes in our study , this suppression was predictable from previous work . For example , our finding that lig4 and ku70 both reduce NHEJ ( Figure 3A and 3B ) is not surprising given that both genes have long been linked to NHEJ repair [40 , 41] . Similarly , the loss of HR-h in mutations at the DmBlm , Rad51 , Top3α , and Rad54 loci ( Figure 4A and 4B; Table 1 ) was expected from previous work on these genes [6 , 42] , as was the decrease in SSA in DmATR and mei-9 ( Rad1p ) mutations [43 , 44] . On the other hand , suppression of SSA by mutations at mus101 and mus301 ( Figure 5A and 5B ) was not apparent from any previous knowledge about the functions of these genes , and the roles of these two genes in SSA remains to be determined . Of the 11 DSB repair loci in this study , ten of them suppressed exactly one of the three pathways ( NHEJ , SSA , and HR-h ) , and none reduced more than one . The only mutation with no differences from wild type in the Rr3 assay was mus81 , although synthetic lethality in the mus81-DmBlm double mutant [35] suggests a role for mus81 in DSB repair and/or recovery from replication fork collapse in Drosophila . Our finding that no mutation suppressed more than one pathway is perhaps surprising in view of the multiple roles found for many DSB repair genes . This apparent simplicity need not be taken to imply that each gene has a major role in only one pathway , but rather may be interpreted to reflect the robustness of the overall DSB repair system and its ability to compensate for defects in any one component . The second piece of information from the Rr3 analysis reveals how each defect is compensated by increases in usage of other pathways . This information opens a new dimension in which to compare DSB repair phenotypes . For example , the finding that lig4 and ku70 mutations are compensated by SSA and not HR-h ( Figure 3E–3G ) could not have been predicted from previous data . Furthermore , there are cases in which mutations in two genes knock down the same pathway , but differ from each other in how the effect is compensated . For example , rad54 , DmBlm , and rad51 mutants all cause severe reduction in HR-h repair , but rad54 is compensated by increases in both SSA and NHEJ whereas DmBlm and rad51 are compensated only by an increase in SSA ( Figure 4 ) . Another example is that of mus101 and mei-9 compared with DmATR and mus301 . All four mutations reduce the SSA outcome , but the first two are compensated only by an increase in NHEJ whereas the latter two are associated with increases in both NHEJ and HR-h ( Figure 5 ) . Furthermore , the HR-h component of this compensation occurs primarily with conversion tracts longer than 156 bp rightward ( i . e . , HR-h long ) . Differences such as these provide new insights into the complex process by which each DSB is channeled to one of the available repair pathways . In particular , as elaborated below , we suggest that how a given defect is compensated depends on where the defect lies within a stepwise decision process , as opposed to a biochemical pathway , in which each DSB is ultimately handled by one repair pathway . Compensatory changes in DSB repair pathway usage are clear from the negative correlations among Rr3 outcomes seen in Figure 2 and elsewhere in this report , as well as from previous publications [28 , 34–36] , but the underlying basis of this compensation is less clear . One possibility is that unsuccessful repair attempts are selectively eliminated , thus increasing the relative frequency of the successful repair processes among the survivors . Drosophila males can produce a sufficient excess of sperm to accommodate considerable selection without detectable loss of fertility . This selection could occur premeiotically , as through apoptosis . It could also occur through formation of aberrant repair products that survive gametogenesis but are eliminated as “dominant lethals” postmeiotically . This latter possibility is most easily tested , as it would result in decreased recovery of the Rr3 chromosome relative to its homolog . We found no evidence for such an effect . For example , in experiment 1 a total of 3 , 952 progeny carried Rr3 versus 4 , 630 with Rr3EJ1 , giving a recovery frequency for Rr3 of 46 . 0% in the absence of any DSB repair mutation . This frequency may be compared with experiment 2 where the lig4 mutation was associated with a marked loss of NHEJ events and compensation by SSA . If the observed decrease in NHEJ from 20 . 8% in experiment 1 to 8 . 6% in experiment 2 ( Table 1 ) is to be explained by postmeiotic selection , one would expect the recovery frequency of Rr3 to be reduced to 40 . 4% in experiment 2 . Instead , we found it was nearly identical to the controls: 45 . 8 ± 0 . 8% out of a total of 14 , 724 progeny . A permutation test to compare experiments 1 versus 2 using the Rr3 frequencies among the progeny of each individual male gave p = 0 . 416 . We conclude that postmeiotic selection accounts for little or none of the observed compensation . In the case of premeiotic selection , a testable prediction is that when one type of outcome is knocked down by mutation , all others would increase proportionately . That is , compensation would occur through an increase in all measured outcomes except those directly affected by the mutation . Instead , we found that compensation usually occurred through only one of the alternative outcomes ( Table 2 ) . In fact , seven of the ten loci where one outcome was diminished displayed compensation in only one other outcome . Furthermore , one of the remaining three loci ( DmATR ) showed compensation by NHEJ and HR-h , but the increase in HR-h occurred primarily with long-tract conversion events . One could argue that these eight genes have secondary effects on some of the other outcomes , but it would be necessary to postulate that these secondary effects counterbalance the increase just enough to leave a net result of no significant difference from the control . From the above observations , we conclude that selection , either post- or premeiotic , is unlikely to account for more than a small proportion of the observed compensation . Instead , we suggest that these compensatory changes reflect the process by which the DSB repair mechanism as a whole channels each break into one of the available pathways . Our finding ( Figure 3 ) that the loss of most NHEJ repair in lig4 and ku70 mutants is compensated only by SSA and not HR-h suggests the existence of a pool of DSBs for which SSA and NHEJ are the only options . A second line of evidence for pool 1 is provided by the mutations mus101 and mei-9 in which the reduction in SSA is compensated by an increase in NHEJ but not HR-h ( Figure 5 ) . The mutations at DmBlm , top3α , and rad51 suggest the existence of a second pool of DSBs . These mutants have reduced HR-h , which is compensated only by SSA and not NHEJ ( Figure 4 ) [35] , suggesting that these mutations act upon a subset of breaks to which only NHEJ and SSA are available . We refer to this hypothetical set of DSBs as pool 2 . Combining these observations leads us to suggest that breaks in Rr3 are channeled into one of these three repair types via a “decision circuit” as shown in Figure 6A . In this scheme , end-joining outcomes are derived only from breaks in pool 1 while HR-h outcomes come only from pool 2 . SSA outcomes , however , can be derived from either pool . Figures 6B–6F show how ten of the mutant loci in our study are interpreted by this scheme . Each mutation restricts one of the decision options ( or two of the options in the cases of DmATR and mus301 ) , which results in increased usage of the alternative option . For example , the rad51 mutation is hypothesized to prevent most DSBs in pool 2 from being repaired via HR-h , resulting in such breaks being handled by SSA ( Figure 6C ) . However , since pool 1 is not affected by rad51 , there is no change in the usage of NHEJ . This description fits the observed phenotype of rad51 ( Figure 4 ) . That is , we saw a severe reduction in HR-h , which was compensated fully by an increase in SSA , while NHEJ remained unchanged . The decision diagram in Figure 6A was originally drawn solely to visualize the compartmentalization of DSBs into the two hypothetical pools and to provide an interpretation for the single-mutant data as in Figures 6B–6F . However , an additional benefit of this scheme is that it makes two predictions that are met by existing data . First , the representation for single mutants of lig4 ( Figure 6B ) and DmBlm ( Figure 6C ) imply that the double mutant would have a greater frequency of SSA than either single mutant , since DSBs in both pools 1 and 2 are expected to utilize SSA more frequently . This expectation is shown in Figure 6G . The lig4; DmBlm double mutant was tested in experiment 14 ( Table 1 ) resulting in an SSA frequency of 89% in cross 2 , the highest of any of the experiments and in good agreement with the model . A second prediction is that mutants with reduced SSA will show a milder degree of reduction compared to those that reduce the other two outcomes . This is because SSA outcomes are drawn from both pools of breaks whereas NHEJ and HR-h each draw DSBs from only a single pool . Consistent with this prediction , we found that in the most severe restriction of SSA ( by mus301 in cross 2 ) the frequency was 67% of its heterozygous control ( experiments 27 and 28 ) . In contrast , the other two repair outcomes , NHEJ and HR-h , displayed much more severe reductions in the corresponding mutants: lig4 mutants without maternal product produced only 22% as much NHEJ as their controls ( experiments 3 and 1 ) , and rad51 homozygotes had less than 1% of the short- or long-tract HR-h frequencies as their controls ( experiments 29 and 22 ) . These observations provide a good fit to the model , although other explanations are possible . Cross 1 can also be represented by a decision circuit diagram , as in Figure 6H . HR-h is not available in cross 1 , and conversion from the sister chromatid ( HR-s ) yields a regenerated I-SceI cut site that is vulnerable to another round of DSB formation . HR-s would also restore the cut site in cross 2 , but this event is not shown explicitly in Figure 6A–G . The scheme in Figure 6H provides an explanation for the peculiar behavior of rad51 mutants that cause a reduction of NHEJ in cross 1 but not cross 2 ( Figure 4F versus 4C ) while increasing SSA in both crosses ( Figure 4E and 4H ) . We interpret these cross-1 effects as indirect consequences of a reduction in HR-s usage . We cannot measure HR-s directly , as it regenerates the original Rr3 structure . However , as shown in Figure 6I , a reduction in HR-s from pool 2 would enhance usage of SSA . Since NHEJ is the only other outcome from cross 1 , its frequency would then decrease owing to elevated competition with SSA . This reasoning , however , applies only to rad51 and not rad54 , since the results from cross 2 place the action of rad54 at a different point in the decision circuit ( Figure 6D ) . The reduction in NHEJ from rad54 in cross 1 is only weakly significant ( p = 0 . 04 ) and could be spurious . In addition , we do not see any effect on SSA or NHEJ in cross 1 for DmBlm or top3α even though they are postulated to act at the same segment of the decision circuit as rad51 ( Figure 6C ) . The reason may be that DmBlm and top3α have a weaker reduction in HR-h than rad51 , so their secondary effect on cross 1 ( indirectly via HR-s reduction ) would be less apparent . At minimum , the diagrams in Figure 6 provide a useful , albeit abstract , way to encapsulate the main points of a complex set of experiments . In discussing these diagrams , we were careful to avoid terms such as “pathway , ” which could be taken to imply that the diagrams represent specific biochemical steps . It is also important to emphasize that any process represented by these diagrams is strictly applicable only to breaks formed in the Rr3 construct . Most naturally occurring DSBs lack the direct duplication needed for SSA repair . Furthermore , specific parameters in the design of Rr3 , such as the length of its direct duplication ( 147 bp ) , are likely to influence the quantitative and even qualitative phenotypes . For example , we found that rad54 increased usage of SSA for duplications of moderate length , as in Figure 4E , but it actually decreased the SSA outcome or similar repair products for much longer duplications ( unpublished data ) . Finally , the genomic location of Rr3 in these experiments , cytological position 48C , is sufficiently far from the telomere to rule out the breakage-induced replication pathway [45] . We also emphasize that the representations in Figure 6 entail some simplifications of the data . In particular , the five measured outcomes ( Figure 1B ) are reduced to three , with the short- and long-HR-h outcomes combined and the NHEJΔ outcome not represented . This simplification masks some potentially important details , such as the finding that the reduction in HR-h by mutations at DmBlm , rad51 , and rad54 ( Figure 6C and 6D ) included both short- and long-tract HR-h equally ( Figure 4A and 4B ) , whereas the increase in HR-h effected by DmATR and mus301 ( Figure 6F ) applies primarily to the long-tract HR-h outcome ( Figure 5G and 5H ) . Despite the above limitations , it is hard to resist some speculation about the physical basis underlying these diagrams , especially the nature of the hypothesized pools 1 and 2 . We can think of pool 1 as containing DSBs for which HR-h ( and possibly HR-s ) is not available . This restriction could result from the timing of the break within the cell cycle , as recently reviewed [46] . Thus , pool 1 could represent breaks that occurred in G1 or early S phase where HR-s is unavailable , and HR-h is infrequent . Interestingly , the route from “DSB” to “pool 1” in our decision circuit scheme is the only one of the six routes in the diagram that is not reduced by any of the mutations studied , suggesting it may not be under genetic control . That suggestion is consistent with the interpretation that entrance into pool 1 in wild type individuals could be determined by the stage of the cell cycle rather than any enzymatic reaction . An alternative interpretation for pool 1 is that it represents breaks for which pairing with a potential template has not occurred . In that case , SSA and NHEJ remain available , but not HR-h or HR-s . Pool 2 consists of DSBs that are slated to be repaired by conversion ( HR-h and HR-s ) or SSA , both of which have extensive 5′ resection of the broken ends as a prerequisite [6 , 47] . Therefore , resection may be the defining characteristic of pool 2 . According to the interpretation in Figure 6D , rad54 restricts entry into pool 2 suggesting it has an early role in the decision process , acting at or prior to the resection stage . This example serves to emphasize the distinction between the decision process visualized in Figure 6 where rad54 appears to act early , versus the biochemical repair process itself where there is evidence that rad54 has one or more later roles [48 , 49] . The apparatus for repairing double-strand breaks is ancient and essential . Without this system , the onslaught of DSBs from replication fork collapse , oxidative damage , ambient radiation , and other sources would severely limit any cell lineage . Moreover , the weakening of genomic integrity would make multicellular organisms more susceptible to cancer [1] and probably accelerate the aging process [2 , 4 , 34 , 36 , 37] . The multiplicity of pathways available for DSB repair provides more than mere redundancy: each pathway comes with its own set of advantages and risks . NHEJ , for example , has the advantage of fewer prerequisites than HR-h or SSA , since it does not require a matching template or a flanking duplication . Its disadvantage , however , is that it usually results in base additions , deletions , or substitutions at the site of repair . At the other extreme is HR-s , which provides minimal risk of mutational changes following repair . A major disadvantage of HR-s , however , is that it is only available following replication . Also , its requirements for homology search and extensive DNA synthesis might be too time-consuming under some circumstances . The wide array of DSB repair methods available to the cell provides the flexibility to handle a variety of situations in an optimal way . We obtain a fuller picture of the entire apparatus by analyzing DSB repair mutations in an experimental system that provides accurate quantitative measurements of multiple outcomes . In particular , we can learn not only how each mutation reduces certain outcomes , but also how the system as a whole compensates for the defect by making increased use of alternative pathways .
Crossing and scoring procedures were as described previously [28 , 34–36] except for the presence of repair mutations . Rr3 was on Chromosome 2 , position 48C , for all experiments , but the ubiquitin-driven I-SceI endonuclease transgene was at one of three locations depending on which repair mutation ( s ) were also present . Earlier results have shown that different endonuclease locations can result in subtle differences in the DSB repair outcome frequencies [28] . Therefore , all comparisons were done between mutant genotypes and controls with the same endonuclease source . All DSB repair measurements pertain to the premeiotic germline of the test males whose progeny were scored . These males were always mated within two weeks of when they eclosed to avoid differences related to male age [34] . We mobilized nearby P element insertions to generate flanking deletions in the areas of lig4 and ku70 . Details of the procedure and the resulting deletion alleles are in Figure S1 . Each estimate of a DSB repair outcome frequency reported in Table 1 is the average of the indicated number of independent estimates from individual test males . The standard errors given in Table 1 and shown in Figures 3–5 were computed from these replicates . These standard errors are provided as a rough calibration of the precision of each estimate , but they were not used for the hypothesis tests in which mutant genotypes are compared with controls . Instead , we used the multiple independent replicates for each estimate in permutation tests [50] , which do not rely on assumptions of normality . Further details of how these permutation tests were performed are published elsewhere [35] . | DNA is a fragile thread that often breaks . When it does , the cell must find a way to splice the broken ends back together in order to continue its cycle of replication . Cells possess an array of ways to rejoin broken DNA ends , each with advantages and disadvantages . Some are “quick and dirty , ” sacrificing accuracy for robustness . They do the basic job of resealing the break but often result in random base changes at the site of the repair . At the other extreme are methods with greater fidelity but added restrictions , such as requiring chromosome replication . We used an experimental system to obtain highly accurate measurements of the relative usage of various repair methods in developing germ cells of fruit flies . The measurements were made in normal flies as well as those carrying mutations at each of 11 genes involved in DNA repair . Most previous studies of these genes focused on specific biochemical pathways . Our results looked at how the repair apparatus as a whole compensates for defects in individual components . The data point to a “decision circuit” for matching each break to a repair method and provide new insight into how our DNA repair machinery protects the genome's integrity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"molecular",
"biology",
"genetics",
"and",
"genomics",
"drosophila",
"melanogaster"
] | 2007 | Multiple-Pathway Analysis of Double-Strand Break Repair Mutations in Drosophila |
The control of virulence regulator/sensor kinase ( CovRS ) two-component system is critical to the infectivity of group A streptococcus ( GAS ) , and CovRS inactivating mutations are frequently observed in GAS strains causing severe human infections . CovS modulates the phosphorylation status and with it the regulatory effect of its cognate regulator CovR via its kinase and phosphatase activity . However , the contribution of each aspect of CovS function to GAS pathogenesis is unknown . We created isoallelic GAS strains that differ only by defined mutations which either abrogate CovR phosphorylation , CovS kinase or CovS phosphatase activity in order to test the contribution of CovR phosphorylation levels to GAS virulence , emergence of hypervirulent CovS-inactivated strains during infection , and GAS global gene expression . These sets of strains were created in both serotype M1 and M3 backgrounds , two prevalent GAS disease-causing serotypes , to ascertain whether our observations were serotype-specific . In both serotypes , GAS strains lacking CovS phosphatase activity ( CovS-T284A ) were profoundly impaired in their ability to cause skin infection or colonize the oropharynx in mice and to survive neutrophil killing in human blood . Further , response to the human cathelicidin LL-37 was abrogated . Hypervirulent GAS isolates harboring inactivating CovRS mutations were not recovered from mice infected with M1 strain M1-CovS-T284A and only sparsely recovered from mice infected with M3 strain M3-CovS-T284A late in the infection course . Consistent with our virulence data , transcriptome analyses revealed increased repression of a broad array of virulence genes in the CovS phosphatase deficient strains , including the genes encoding the key anti-phagocytic M protein and its positive regulator Mga , which are not typically part of the CovRS transcriptome . Taken together , these data establish a key role for CovS phosphatase activity in GAS pathogenesis and suggest that CovS phosphatase activity could be a promising therapeutic target in GAS without promoting emergence of hypervirulent CovS-inactivated strains .
The ability of bacteria to modify gene expression levels in adaptation to external influences is key to many aspects of bacterial pathogenesis [1] . Two-component regulatory systems ( TCS ) are a major mechanism by which bacteria detect and respond to diverse environmental factors [2 , 3] . TCS are absent in humans but abundant in a wide range of bacteria . They usually consist of a membrane-embedded histidine kinase that determines the regulatory activity of its cognate response regulator by altering its phosphorylation status [2 , 4] . The control of virulence regulator sensor ( CovRS , also called CsrRS for capsule synthesis regulator ) system of group A streptococcus ( GAS ) is one of the best-studied TCS in connection with bacterial pathogenesis [5 , 6] . GAS is a strictly human pathogen that causes a variety of diseases from relatively benign to life threatening such as necrotizing fasciitis [7] . GAS strains are classified into >200 serotypes based on variability in the key anti-phagocytic , cell-surface exposed M protein [8 , 9] . In tandem with the histidine kinase CovS , CovR is the central regulator of GAS virulence factor production [5 , 10] . Similar to other OmpR/PhoB family members , CovR is phosphorylated at a conserved aspartic acid residue ( D53 ) to create CovR~P , which is considered to be the active regulatory form of the protein [11–13] . Several signaling pathways converge to tightly regulate CovR~P levels . CovS primarily serves to increase CovR~P via its kinase activity [14] . As a member of the bifunctional HisKA-family of histidine kinases [15] , CovS also possesses phosphatase activity to reduce CovR~P . Extracellular signals ( e . g . Mg2+ or LL-37 ) influence CovS activity to modulate CovR~P levels and CovR-regulated virulence gene expression [16–18] . Deletion of covS reduces but does not completely eliminate CovR~P [14] suggesting a contribution of intracellular metabolites such as acetyl phosphate to CovR phosphorylation . In combination with CovS , the orphan kinase regulator of CovR , RocA , increases CovR~P levels [19] . It has been discovered that conserved mutations in the rocA gene mediate serotype-specific intrinsic CovR~P levels [20 , 21] . Finally , serine/threonine kinase ( Stk ) phosphorylates CovR residue threonine 65 in a fashion that antagonizes D53 phosphorylation [22] . This multi-faceted regulation of CovR~P status underpins its key position in GAS infectivity , which is further highlighted by the observation that the CovRS system is one of the hotspots for mutations in invasive clinical isolates [23 , 24] . Given that CovR mainly serves as a transcriptional repressor , emergence of CovRS inactivating mutations during human infection or animal passage result in hypervirulent GAS strains due to relieved repression of virulence gene expression [25–28] . It has been speculated that there is a serotype/strain-specific tendency for the acquisition of covRS mutations that comes along with the underlying invasive potential of different M-types [29] . This tendency is thought to be determined by serotype-specific expression of certain virulence factors , such as the hyaluronic acid capsule , the M protein and the secreted DNase Sda1 [30–33] . Likewise , the intrinsic CovR~P levels of GAS strains might play a role in determining the selective pressure to lower CovR~P via covRS mutations [14] . Although the CovRS system has been extensively studied , several key outstanding issues regarding its impact on GAS pathogenesis remain unanswered . First , it is well established that CovRS-inactivation increases GAS virulence in bacteremia models [5 , 21 , 22 , 25 , 34] . However , there are conflicting results regarding how CovR~P levels impact GAS skin/soft tissue infection and nasopharyngeal colonization [20 , 25 , 26 , 35 , 36] . Second , the specific influence of CovR kinase and phosphatase activity on GAS virulence is not clear since previous research has focused on the impact of CovR or CovS inactivation . We recently showed that a strain lacking CovS phosphatase activity resulted in increased CovR~P levels [14] . However , how increased CovR~P levels due to abrogating CovS phosphatase activity impact GAS infectivity remains unknown . Finally , virulence studies using strains with varying CovR~P levels have generally been restricted to a single GAS serotype leaving open the question of the applicability of findings in a single strain to the broader GAS population . To address these questions , we employed isoallelic GAS strains of two distinct M serotypes that are common causes of human disease to comprehensively investigate how defined CovR~P levels impact GAS virulence , global gene expression , and emergence of hypervirulent CovRS mutated strains .
We have previously created and characterized isoallelic serotype M3 strains in which single amino acids in CovRS are altered compared to the parental strain MGAS10870 [14 , 22] . We recreated the same set of mutations in the parental strain MGAS2221 , a contemporary serotype M1 GAS strain ( see Table 1 ) ( For clarity of the genetic background of the used strains , we refer to these strains as M1-wild type ( WT ) , M3-wild type ( WT ) etc . for the remainder of the manuscript ) . The CovR~P levels ( defined as [CovR~P]/[CovR]total ) in the respective M1 GAS strains during growth in THY medium were determined by Phostag-Western blot using anti-CovR antibodies ( Fig 1 ) . Similar to our M3 studies [14 , 22] , changing the CovR phosphorylation site aspartate 53 to an alanine resulted in no detectable CovR~P in strain M1-CovR-D53A . Abolishing CovS kinase activity via the E281A mutation decreased CovR~P levels to ~20% in M1-CovS-E281A , resembling a covS deletion strain , whereas the T284A change abrogating the CovS phosphatase activity increased CovR~P levels to ~80% in strain M1-CovS-T284A ( Fig 1 ) . Thus , the CovR~P levels observed in the isoallelic mutant strains of M1-WT closely mirrored those previously observed for the serotype M3 derivatives [14] . However , it is important to note that CovR~P levels are higher in the M1-WT strain ( ~70% ) relative to the M3-WT strain ( ~40% ) due to a naturally occurring mutation in the RocA protein in M3 strains [14 , 21] . This collection of strains allowed us to study in detail the impact of CovR phosphorylation on GAS virulence and covRS switching rates during infection in two of the most prevalent GAS serotypes causing pharyngitis and severe infections in the US ( Center for Disease Control and Prevention [37] ) . To investigate the relationship between CovR~P levels and GAS virulence , we first employed a skin/soft tissue mouse model that mimics cellulitis , a common manifestation of GAS infection . For each strain , 20 mice were subcutaneously challenged with 107 colony forming units ( CFU ) of GAS , and lesion size measurements were performed daily ( Fig 2A and 2B ) . In the serotype M3 background ( Fig 2A ) , all of the strains caused appreciable disease with strain M3-CovR-D53A causing the largest lesions over the course of the entire experiment ( P <0 . 05 in comparison to each of the other three strains ) . The difference in lesion size generated by strain M3-CovS-E281A compared to the M3-WT strain was not significantly different ( P = 0 . 99 ) . The smallest lesions were produced by strain M3-CovS-T284A ( P < 0 . 05 compared to each of the other three strains ) . However , it is notable that for some mice inoculated with strain M3-CovS-T284A , lesions started to appear more severe and showed ulceration after day 5 of the experiment , when healing of lesions had already set in for the other strains . We will address this finding later in the manuscript . Among the M1 strains , the wild type strain was the most virulent with maximal average lesion area of ~150 mm2 on day 4 ( P < 0 . 05 compared to the other three strains ) , followed by strain M1-CovS-E281A ( P < 0 . 05 compared to strains M1-CovS-D53A and M1-T284A ) ( Fig 2B ) . Strain M1-CovS-T284A was the least virulent ( Fig 2B ) ( P < 0 . 05 compared to the other strains ) . However , in contrast to its serotype M3 counterpart , only a few mice inoculated with M1-CovS-T284A evidenced any visible infection , and in these cases the lesions were small in size and without any sign of ulceration over the entire length of the experiment . Thus , we did not observe a consistent relationship between CovR~P levels and lesion size . Regardless , our results demonstrate that abrogation of CovS phosphatase activity in both GAS M1 and M3 serotypes attenuated virulence in the skin/soft tissue mouse model of infection . Next , we investigated the effect of CovR~P levels on oropharyngeal GAS colonization following nasopharyngeal mouse challenge . For serotype M3 strains , 20 mice per strain were inoculated with 108 CFU GAS in a 20 μl volume . The volume was reduced compared to previous studies to avoid aspiration of GAS into the lungs [36] . Unexpectedly , we observed a high death rate in mice inoculated with strains M3-CovS-E281A and M3-CovR-D53A , while no death occurred for mice inoculated with M3-WT or M3-T284A , albeit the same volume and CFU were used . For this reason , we subsequently reduced the inoculum to 107 CFU for all serotype M1 strains . Despite the different inocula used for the M1 and M3 strains , we consistently observed that GAS strains with higher CovR~P colonized at lower rates ( Fig 2C and 2D ) . Specifically , low CovR~P strains CovS-E281A and CovR-D53A colonized at significantly higher rates ( ~80% and 40–50% for serotype M3 and M1 , respectively ) compared to the wild type and CovS-T284A strains for both serotypes ( P < 0 . 05 for all specified comparisons ) . For serotype M3 , the wild type strain ( medium CovR~P level ) had a colonization rate intermediate to the low and high CovR~P strains ( ~30% , Fig 2C ) . For serotype M1 , the wild type strain and M1-CovS-T284A ( both high CovR~P ) were rarely recovered shortly after initial inoculation ( Fig 2C and 2D ) . Thus , we observed an inverse relationship between CovR~P levels and the capability of the GAS strains to colonize mouse pharyngeal tissue . To investigate whether changes in CovR~P levels affect GAS interaction with the human immune system , we performed a Lancefield bactericidal assay using heparinized whole human blood of three non-immune donors . GAS survival in whole human blood was determined by plating serial dilutions on BSA plates after 3 hours exposure of GAS cells to human blood , and multiplication factors compared to the inoculum were calculated for each strain ( Fig 2E and 2F ) . Multiplication factors were generally lower for serotype M1 compared to serotype M3 ( Fig 2E and 2F ) . This is consistent with the previous finding that an intact RocA protein negatively influences GAS survival in blood [39] . The multiplication factors between the respective wild type and CovR-D53A or CovS-E281A mutant strains were not significantly different ( P > 0 . 05 for all comparisons ) . In stark contrast , CovS-T284A derivatives of both serotype M1 and M3 GAS completely lost the ability to survive and propagate in whole human blood ( P < 0 . 001 for all comparisons ) ( Fig 2E and 2F ) . Thus , CovS phosphatase activity is crucial for GAS survival and propagation in whole human blood . In standard laboratory medium CovR~P levels of strain M1-CovS-T284A are only slightly elevated compared to those in M1-WT , yet there were dramatic differences in virulence between the two strains in the skin/soft tissue infection model and bactericidal assay ( Figs 1 and 2 ) . One possible explanation for this discrepancy is that host factors increase CovS phosphatase activity during infection to decrease CovR~P levels and augment virulence factor production . Indeed , CovS senses the human antimicrobial peptide LL-37 and responds by lowering CovR~P levels [14 , 18] . Thus , we next addressed the question whether LL-37 signaling is affected by the CovS-T284A mutation . To this end , we measured CovR~P and hasA ( hyaluronic acid capsule ) transcript levels in GAS strains grown in THY ( standard laboratory medium ) and THY supplemented with LL-37 ( Fig 3A and 3B ) . Consistent with our previous findings [14] , M1-WT CovR~P levels were reduced and consequently hasA transcript levels elevated in the presence of LL-37 compared to unsupplemented THY . In contrast , supplementation with LL-37 did not affect CovR~P or hasA transcript levels in strain M3-WT or strain M1-WT engineered to contain the truncated M3 version of RocA ( M1-rocAM3 ) [39] . In both serotypes , the CovS-T284A strains had higher CovR~P and lower hasA transcript levels compared to the respective wild type strains when grown in unsupplemented THY . However , neither CovR~P nor hasA transcript levels were influenced by LL-37 in the CovS phosphatase deficient strains ( Fig 3 ) . Further , CovR~P and hasA transcript levels of CovS phosphatase deficient strains were not affected by an additional mutation in the rocA gene in either medium ( see strain M1-CovS-T284A/rocAM3 , Fig 3 ) . We conclude that LL-37 increases CovS phosphatase activity in a GAS strain with an intact RocA . It is well established that strains with covRS mutations can emerge during human infection or animal passage , thereby giving rise to hypervirulent GAS isolates [23 , 25–27] . To investigate how CovR~P levels in the context of GAS serotype influence the emergence of CovRS inactivated mutants , we amplified and sequenced the complete covRS operon from at least 25 randomly picked GAS colonies per strain isolated during each infection study ( Table 2 ) . In the skin/soft tissue model of infection , we recovered GAS colonies from skin lesions of five mice per strain on day 4 . No additional mutations were found in colonies isolated from mice infected with strains M3-WT , M3-CovS-E281A , and M3-CovR-D53A . In contrast , a few of the colonies isolated from mice infected with strain M3-CovS-T284A had mutations in CovR , namely CovR-L155I and CovR-R66H . Since the lesions in some mice infected with M3-CovS-T284A looked more severe after day 5 , we sequenced additional colonies isolated from these animals on day 10 ( end point of experiment ) and were able to detect colonies that had a duplication of covS nucleotides 100 to 131 , which is predicted to result in a non-functional CovS due to frameshift ( Table 2 ) . Therefore , on rare occasions , it appears that mutations that abrogate CovS activity occurred late in the disease course for strain M3-CovS-T284A which likely accounted for the increase lesion size described above . As observed with the M3 strains , no GAS with additional covRS mutations were recovered from animals infected with strains M1-CovR-D53A and M1-CovS-E281A . Interestingly , however , unlike its serotype M3 counterpart , no additional mutations were found in GAS isolated from mice infected with strain M1-CovS-T284A . In contrast , a high number of colonies recovered from animals inoculated with the M1-WT isolate had mutations in the covRS systems as has been previously reported for this strain [25 , 40] ( Table 2 ) . Many of the recovered strains had changes that truncated the CovS protein , whereas several colonies had non-synonymous SNPs in covR ( A81T in CovR ) or covS ( R241S or P285S in CovS ) ( Table 2 ) . No additional mutations were detected in any GAS strain of either serotype during nasopharyngeal mouse challenge or during growth in whole human blood . Thus , elimination of phosphatase activity by the CovS-T284A change abrogated the emergence of covRS mutations in the M1 background during skin/soft tissue infection but increased such emergence in the M3 strain , albeit primarily late in the infection course . It has previously been shown that truncations in CovS mimic a covS deletion strain ( e . g . reduced CovR~P levels ) , but much less is known about the consequences of non-synonymous single nucleotide polymorphisms ( SNPs ) in either covR or covS that arise during mouse challenge or human infection [38 , 41] . Thus , we next sought to evaluate the effect of some of the previously uncharacterized SNPs isolated during our mouse infection study by generating the isoallelic GAS strains CovR-A81T , CovR-R66H , CovR-L155I , and CovS-P285A in the serotype M3 background ( Fig 4 ) . CovR~P levels in strains CovR-A81T and CovS-P285S were strongly reduced compared to the wild type and resembled that of a covS deletion strain . In contrast , strains CovR-R66H and CovR-L155I had CovR~P levels similar to the wild type ( Fig 4A ) . SpeB is an actively secreted broad-spectrum protease whose production is abrogated by CovS inactivation [42 , 43] . In accordance with the CovR~P levels , strains CovR-A81T and CovS-P285S had reduced SpeB activity on milk plates , while SpeB activity was not affected in CovR-R66H or CovR-L155I ( S1 Fig ) . Next , we performed TaqMan qRT-PCR of various known CovR-regulated genes that have previously shown to be regulated by CovR via different mechanisms [44] to evaluate the effect of the mutations on CovR-mediated transcription regulation ( Fig 4B–4D ) . Consistent with the CovR~P level analysis , in strains CovR-A81T and CovS-P285S transcript levels of spyM3_0105 , prtS , sagB , and cbp ( which encode a cell surface protein , an IL-8 degrading protease , a pore-forming toxin , and a pilus protein , respectively ) resembled that of a covS deletion strain . Specifically , the transcript levels of spyM3_0105 and prtS were >10 or ~2-fold elevated , respectively , compared to M3-WT ( Fig 4B and S2 Fig ) , while transcript levels of sagB and cbp did not differ significantly compared to M3-WT ( Fig 4C and 4D ) ( P > 0 . 05 ) . Interestingly , despite CovR~P levels being similar to the wild type , repression of all genes studied was strongly relieved in strains CovR-R66H and CovR-L155I compared to the wild type ( Fig 4B–4D ) implying that functions aside from CovR phosphorylation ( e . g . signal transduction , DNA binding ) are affected in these mutants ( P < 0 . 05 for all gene transcript levels ) . Next , we determined the transcriptomes of the eight strains used in the animal challenges to obtain mechanistic insights into the observed virulence differences . To this end , four biological replicates per strains were grown to late-logarithmic phase ( OD = 0 . 9 ) , and RNA was extracted and subjected to RNAseq analysis . Transcript levels were considered significantly different if the mean transcript level difference was ≥ 2 . 0 fold and the final adjusted P value ≤ 0 . 05 compared to the wild-type strain . By principle component analysis ( PCA ) biological replicates of each strain clustered together ( Fig 5A and 5B ) . Consistent with the higher CovR~P levels in M1-WT compared to M3-WT , there were more genes differentially regulated compared to the wild type in strains M1-CovR-D53A and M1-CovS-E281A and less genes in M1-CovS-T284A than in the respective M3 serotype strains ( Table 3 ) . That is , the number of differentially regulated genes paralleled the differences in CovR~P between the wild type and the distinct isoallelic strains . In accordance with our virulence data , the transcriptomes of M3-CovR-D53A and M3-CovS-E281A were highly similar with 100 and 97 genes being differentially regulated compared to the wild type ( Fig 5A , Table 3 ) . Genes up-regulated compared to M3-WT included the known virulence genes prtS and speA ( encoding a pyrogenic exotoxin ) . The transcriptomes of M1-CovR-D53A and M1-CovS-E281A were also similar . However , compared to their M3 counterparts , a larger number of known virulence factor encoding genes were up-regulated in these strains relative to M1-WT including the has operon , nga ( NAD glycohydrolase ) , slo ( streptolysin O ) , and speC ( exotoxin ) . Moreover , transcript levels of several DNA binding proteins and genes involved in amino acid and sugar transport and metabolism were increased compared to the wild type . In contrast , transcript levels of 41 and 23 genes were decreased in strains M3-CovS-T284A and M1-CovS-T284A compared to the wild type , respectively . Remarkably , genes further repressed in the CovS-T284A strains comprised those encoding nearly the complete repertoire of known GAS virulence factors ranging from secreted toxins to immune-modulating surface proteins ( see Fig 5 and S2 Table ) , which likely explains the striking reduction in virulence of CovS phosphatase deficient GAS strains seen in our infection studies . Many genes with reduced transcript levels in the CovS-T284A strains have not been previously identified as part of the CovRS-regulon including mga , which encodes the multi-gene activator of numerous virulence genes , Mga , and part of its regulon , e . g . emm coding for M protein or grm ( gene regulated by Mga ) [45] . On the other hand , with the exception of the dpp operon ( encoding a dipeptide permease ) we did not identify genes involved in metabolism and transport being further repressed in the CovS-T284A strains , suggesting specific modulation of virulence gene expression in these strains . Thus , results of our transcriptome analyses can explain the hypervirulent phenotype of CovS phosphatase deficient strains by identifying ( in part novel ) CovR-mediated repression of a broad array of virulence factor encoding genes at high CovR~P levels . We further used our multi-strain , multi-serotype RNAseq data to differentiate three broad classes of CovR-regulated genes depending on their repression ( T284A ) and de-repression ( E281A/D53A ) profile ( see examples in Fig 5C–5E ) . The first class ( class I ) encompasses genes that were de-repressed in the E281A and D53A strains in both serotype M1 and M3 . These genes have been identified as part of the CovRS regulon by previous transcriptome studies [10 , 17 , 22 , 25 , 28 , 38] and include well-known virulence genes like the has operon and prtS . Interestingly , these genes were only further repressed in the T284A strain compared to parent strain M3-WT while they were already fully repressed in M1-WT ( Fig 5C ) . In contrast , we defined class II genes as those , whose transcript levels were affected in the E281A/D53A strains only in the M1-WT background but showed repression in the T284A strains for either serotype . These genes included nga/slo and scpA ( C5 peptidase ) ( Fig 5D ) . As mentioned earlier , our transcriptome analysis identified novel virulence factor repression in both CovS-T284A strains . These genes , categorized as class III genes , were not increased in the CovS-E281A and CovR-D53A strains and have not previously been identified as part of the CovRS regulon in serotype M1 and serotype M3 GAS strains ( Fig 5E ) . Additionally , transcript levels of several prophage-encoded , serotype-specific virulence factors like speK ( exotoxin ) , sla ( extracellular phospholipase ) or sdaD2 , spd3 , and sdn ( all secreted DNases ) were reduced in the CovS-T284A mutant strains ( Fig 5F ) . Due to their serotype-specific nature these genes could not be unambiguously assigned to one of the described major classes . Extracellular DNases have been shown to contribute to GAS pathogenesis [46] . To gain insight into the physiological consequences of our transcription data , we performed DNase activity tests and found that indeed DNase activity was significantly reduced in supernatants derived from CovS-T284A cells compared to that of the respective parental strains ( S3 Fig ) . Next , we employed TaqMan qRT-PCR to confirm the transcript level pattern of selected genes exemplifying the different classes of CovR-regulated genes revealed by RNAseq ( Fig 6 and S4 Fig ) . To enable a better comparison of transcript levels vs . CovR~P in both serotypes , we also included strain M1-WT grown in THY supplemented with 100nM LL-37 ( CovR~P is ~40% as for M3-WT ) in our analyses . The qRT-PCR results were in concert with the RNAseq data . By contrasting the level of CovR~P with the degree of transcriptional regulation we were able to deduce differences in CovR~P dependency for the regulation of distinct promoter classes . The transcript levels of class I genes , exemplified by prtS ( Fig 6B ) and hasA ( S4B Fig ) , revealed differential regulation over a range of 20–70% CovR~P , above which no further repression was detectable . In contrast , gene regulation of class II genes , such as slo ( Fig 6C ) and scpA ( S4C Fig ) , was not affected by varying CovR~P below a level of 40% , but was dramatically impacted when CovR~P was increased from 40% to beyond 70% . Mga ( Fig 6D ) and emm ( S4D Fig ) , as examples for class III genes , showed the highest CovR~P dependency with CovR-mediated repression only observed at CovR~P above 70% . Mga and its regulated genes have previously not been identified as part of the CovR regulon in serotype M1 or M3 GAS [17 , 25] . Our transcriptome data , however , showed that the expression of mga and several Mga-controlled genes such as emm or grm were down-regulated in M3-CovS-T284A and M1-CovS-T284A compared to the respective parental strain ( Fig 5E and Fig 6D ) . Given the presence of potential CovR-binding sites ( ATTARA ) within the mga promoter , we next performed electrophoretic shift mobility analyses ( EMSA ) to address the question whether CovR binds to the promoter of mga and whether this binding is dependent on the phosphorylation status of the protein . A PCR fragment of ~500bp encompassing the mga promoter amplified from M1-WT genomic DNA ( DNA sequences of mga promoter regions from M1-WT and M3-WT have 94% nucleotide identity ) was incubated with increasing concentration of unphosphorylated or in vitro phosphorylated purified CovR protein , and samples were separated on a TBE-PAA gel . Although unphosphorylated CovR was able to bind the mga promoter DNA to create a low molecular weight complex , increasing protein concentrations up to 5 μM did not appreciably change the binding behavior ( Fig 7A ) . By contrast , increasing concentrations of CovR~P progressively resulted in complexes of higher molecular weight as typically seen in CovR/CovR~P promoter binding assays of genes known to be directly regulated by CovR [22 , 28] ( Fig 7B ) . In accordance with our transcription data , high concentrations of CovR~P were needed for effective binding of the mga promoter .
Although phosphorylation of response regulator proteins is critical for bacterial pathogenesis , there remains limited understanding of how variation in response regulator phosphorylation impacts bacterial virulence at diverse infection sites . Herein we employed transcriptome and virulence assays of an array of isoallelic serotype M1 and M3 GAS strains to assess the impact of multiple , distinct phosphorylation levels of the key response regulator CovR on GAS pathophysiology . For our virulence assays , we chose the mouse models of skin/soft tissue infection and nasopharyngeal colonization . Previous studies on the contribution of CovRS to GAS infection using these models have come to varying conclusions [20 , 25 , 26 , 35 , 36] , which may be explained by serotype-specific differences , the use of strains with diverse inactivated regulators with potentially additional functions ( ΔcovS vs . ΔrocA ) or possible emergence of hyper-virulent clones during infection . In contrast , all studies to date using the intraperitoneal ( i . e . bacteremia ) model had consistently found increased GAS virulence with decreasing CovR~P levels [5 , 21 , 22 , 25 , 34] . Confirming many previous observations , our assays indicate both serotype- and site-specific effects on the impact of CovRS inactivation on GAS pathogenesis . Similar to results obtained by Dalton et al . [35] , we saw an inverse correlation between CovR~P levels and virulence of our M3 strains in the skin/soft tissue model . This result was not mirrored by our M1 strains . However , the hypervirulent phenotype of M1-WT in our mouse skin/soft tissue model likely stems from the emergence of CovS inactivating , SpeB- mutations early in the infection course . It has been shown that a mixture of wild type and covS-inactivated M1 GAS produces larger skin lesions compared to a covS deletion strain [26] , consistent with skin-specific increase in virulence of wild-type vs . covS-inactivated M1 GAS described by Sumby et al . [25] . Although inactivating CovS is believed to reduce GAS fitness in the upper respiratory tract [36 , 47] , we observed a profound increase in nasopharyngeal colonization rates for strains with lower CovR~P levels for both serotypes . Despite the limitations of the nasopharyngeal mouse model , a similar study using an M18 strain found that increasing CovR~P levels by repairing a naturally occurring RocA mutation also decreased colonization rates [20] . While previous studies solely analyzed the effect of lowering CovR~P on GAS infectivity , the most striking conclusion of our virulence data was the consistent hypovirulent phenotype of the CovS-T284A strains . In both M1 and M3 GAS serotype , these CovS phosphatase deficient strains caused the smallest lesions in the skin/soft tissue model , had a strongly reduced capacity to colonize the mouse oropharynx and did not survive neutrophil killing in whole human blood . Thus , CovS phosphatase activity strongly influences GAS overall ability to cause disease . Another key finding from our animal studies was the effect of CovR~P levels on emergence of GAS strains with CovRS inactivating mutations . Consistent with the concept that the primary selective pressure for CovRS mutations is to decrease CovR~P levels , we only observed mutations in strains whose initial CovR~P levels were ≥ 70% . The absence of additional mutations in the E281A strains suggest that there is no selection pressure to reduce CovR~P beyond ~20% and is in accordance with human data that GAS strains with CovR inactivating mutations are rare compared to CovS . Surprisingly , increasing CovR~P levels via the T284A mutations did not evoke high levels of CovRS mutations . The latter occurred late in the infection course in a very limited number of animals infected with M3-CovS-T284A and not at all in mice infected with M1-CovS-T284A . Given the small lesion sizes , we speculate that the profound hypovirulence induced by the T284A mutation inhibited such emergence . Our transcriptome analysis offers an explanation for the hypovirulent phenotype by revealing specific down-regulation of nearly the entire repertoire of virulence factor encoding genes in the CovS-T284A strains . Interestingly , plasmid-derived overexpression of RocA in M1-WT produced a similar virulence gene repression profile [39] . Given our finding that RocA impairs CovS phosphatase activity ( Fig 3 ) , we speculate that RocA overexpression results in CovR~P levels similar to the CovS-T284A strains . Among the identified novel CovR-repressed virulence genes in the CovS-T284A strains were mga and Mga-regulated genes . Mga is a well-known activator of numerous GAS virulence factors , such as M protein , and is thought to be particularly important in the early stages of infection [45 , 48] . Thus it is likely that decreased mga expression played a pivotal role in the observed hypovirulence of our CovS phosphatase deficient strains . In addition , down-regulation of the Mga regulon has been shown to prevent in vivo selection of hypervirulent SpeB negative covRS variants [32] , and thus the low mga and emm transcript levels in the CovS-T284A strains may have negatively affected hypervirulent isolate emergence . Although an indirect relationship between CovR and Mga is possible [49] , our DNA binding analysis suggests that CovR could directly regulate mga . In accordance with this , we found several potential CovR-binding sites ( ATTARA ) within the mga promoter region , in particular an ATTARA sequence directly upstream of the P2 and downstream of the P1 promoter [50] as well as one partially overlapping a CodY binding site [51] . Nonetheless , in serotype M1 and M3 GAS , a functional protein-DNA complex seems to be only achieved at high CovR~P levels as seen in the CovS phosphatase deficient strains . Our multi-strain , multi-serotype approach during our transcription level analysis allowed us to distinguish three distinct classes of CovR-regulated genes on the basis of their CovR~P dependency for gene regulation . In addition , we have previously described a group of CovR-regulated genes ( e . g . sagB , cbp , covR ) , whose transcription regulation is independent of CovS [14 , 44] . Hence , for this group , which we designate as class 0 in this context , CovR~P of only 20% is sufficient to repress gene expression . Together our data suggests classes of promoters repressed under distinct CovR~P concentrations increasing from class 0 to class III . This CovR~P dependency is likely determined by a combination of different affinities for CovR binding sites ( as suggested by Jain et al . [39] ) and diverse DNA-binding mechanisms ( cooperativity , CovR oligomerization state , or interaction with other regulators or RNA polymerase ) [6 , 12 , 44 , 52–55] and requires further investigations . Regardless , the gradual repression of promoter groups establishes the basis for coordinated expression of GAS virulence factors in response to changing environmental cues . Adjusting gene expression in adaptation to environmental niches is pivotal for pathogenic bacteria . Recently , this function has been increasingly attributed to the phosphatase activity of bi-functional HisKA-family kinases [56–58] . Thus , besides limiting crosstalk between homologous TCS [59] , histidine kinase phosphatase activity evidently fulfills an important role in sensing extracellular signals . The currently established environmental signals that modulate the function of CovRS TCS likewise seem to target CovS phosphatase rather than kinase activity . Previous investigations revealed that inactivation of CovR by CovS is required for survival of GAS under stressful conditions such as the presence of LL-37 , iron starvation or acidic stress suggesting that stress signals activate CovS phosphatase activity in M3 or M6 strains [60–62] . Mg2+ reduces CovS phosphatase activity by an unknown mechanism thereby increasing intracellular CovR~P [14] . Herein we show that the presence of accessory protein RocA in MGAS2221 also diminishes CovS phosphatase activity . We speculate that RocA forms a hetero-oligomeric complex with CovS thereby stabilizing a CovS phosphatase incompetent conformation ( activated state ) [63] . Further , we show that the antimicrobial peptide LL-37 in turn increases CovS phosphatase activity . LL-37 has been demonstrated to bind directly to CovS [64] but its effect on CovR~P and transcription regulation is only observed in GAS strains expressing a functional RocA protein . Thus , we hypothesize that LL-37 increases CovS phosphatase activity indirectly by displacing RocA from the hetero-oligomeric complex to allow formation of a CovS phosphatase competent conformation . The antagonism between Stk mediated phosphorylation of CovR T65 and CovS phosphorylation at D53 adds an additional layer of complexity to regulation of CovR function [22] . The multi-faceted regulation of CovS phosphatase activity highlights its crucial function in adjusting CovR~P status and thus the expression of CovR-controlled virulence genes . Bacterial TCSs have been proposed as potential therapeutic targets [65–71] . General inhibition of CovS function is unlikely to be desirable in GAS given the hypervirulence of ΔcovS strains . However , the data presented in this study suggests that specifically targeting CovS phosphatase activity might be promising . CovS is present in all GAS serotypes , and abolishing CovS phosphatase activity markedly reduced GAS virulence in all three infection models . Notably , hypovirulence was even detected in serotype M1 GAS , a strain with low intrinsic CovS phosphatase activity and therefore high baseline CovR~P . Further , phosphatase activity of other bifunctional HisKA-family histidine kinases has been shown to play an important role in regulating infectivity of both Gram-positive and Gram-negative pathogens . For example , Liu et al . demonstrated that a T247A mutation ( homolog to CovS-T284A ) of Salmonella enterica EnvZ , mimicking a conserved pH-controlled mechanism of HK phosphatase activity ablation , increased macrophage infectivity due to accumulation of OmpR~P and downstream activation of ssrA-ssrB genes [58] . Mutations of WalK PAS domain that modulate WalK phosphatase activity also attenuated virulence of S . pneumoniae in a murine infection model [72] . The Spinola group has suggested the use of phosphatase inhibitors towards histidine kinase CpxA to treat certain urinary tract infections caused by uropathogenic E . coli [73 , 74] . These examples corroborate the importance of HK phosphatase activity in bacterial virulence and suggest a broader application of this approach with regards to other histidine kinases in pathogen bacteria . Together our study provides novel insights into mechanisms of GAS virulence factor regulation and establishes an important role of CovS phosphatase activity in controlling GAS pathogenicity .
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . Protocol #00001455-RN00 was approved by The University of Texas MD Anderson Cancer Center Institutional Animal Care and Use Committee . All efforts were made to minimize suffering . Human blood samples were drawn and used under IRB protocol #0110–0015 approved by Houston Methodist Research Institute Review Board . Written informed consent was obtained from all donors . All GAS strains used in this study are derivatives of either strain MGAS10870 ( serotype M3 , herein called M3-WT ) or strain MGAS2221 ( serotype M1 , herein called M1-WT ) , two clinical isolates that are known to have wild type covRS sequence ( see Table 1 ) . Primers for strain creation are listed in S1 Table . Single nucleotide exchanges were introduced into chromosomal DNA of strain M1-WT or M3-WT via homologues recombination using the integrative plasmids pBB740 or pJL1055 , respectively , as described in [28] to create isoallelic strains that differ only by the presence of a single amino acid replacement in CovR or CovS . GAS strains were statically grown in Todd Hewitt broth supplemented with 0 . 2% yeast ( THY ) at 37°C and 5% CO2 . When appropriate , chloramphenicol was added to 5μg/ml . Bacteria were plated on tryptic soy agar with 5% sheep blood ( BSA ) plates . Recombinant CovR was purified and phosphorylated as described [28] and served as control . GAS lysates were prepared and separated on 12 . 5% SuperSeq Phostag gels ( Wako , USA ) , and un/phosphorylated CovR species were detected using a polyclonal anti-CovR antibody and the Odyssey imaging system as described previously [14] . Independent Western blots were repeated at least twice . Strains were grown in 200 ml THY to mid-exponential phase and harvested by centrifugation at 9000 rpm . Cell pellets were washed twice with ice-cold PBS buffer , re-suspended in 4 ml PBS/20% glycerol solution and stored in aliquots at -80° C until use . CFU counts for each strain were determined by plating dilutions of the samples on BSA plates at least three times and confirmed after mice inoculation . All animal experiments were performed in a blinded fashion . Whole blood was drawn in sodium heparin tubes ( Becton Dickinson ) from three consented , healthy , non-immune donors under IRB protocol #0110–0015 . GAS growth in blood was performed as described in [21] . Indicated strains were grown in THY as described for animal experiments . 20–100 CFU of each GAS strain was inoculated in 300μl human blood containing 10% THY , respectively . Samples were incubated for 3h at 37°C in 5% CO2 with end-to-end rotation . CFU/ml were determined by plating serial dilutions in PBS on BSA plates for enumeration of β-hemolytic colonies . Multiplication factors were determined by dividing CFU/ml after 3h incubation by CFU/ml in the inoculum . The experiment was performed in triple biologic replicates on two separate occasions . Data were analyzed using one-way ANOVA followed by post-hoc analysis using Tukey’s correction for multiple comparisons . The DNase activity in filtered culture supernatants from M1 and M3 wild type and CovS-T284A strains grown to late-exponential phase ( OD = 0 . 95 ) was assayed . To this end , 100ng PCR-derived GAS DNA was incubated for 20 min at 37°C in 1x NEB 2 buffer ( New England Biolabs ) with 0 . 5μl of the respective supernatant . The remaining DNA was quantified using Quant-IT Pico Green dsDNA reagent ( Thermo Fisher Scientific ) according to the manufacturer’s instructions . Fluorescence was detected using a fluorescence microplate reader with 480 nm excitation and 520 nm emission wavelength . DNA concentrations were calculated using a standard curve with known DNA concentrations . Three biological replicates were assayed on two separate occasions ( n = 6 ) . The ~500bp encompassing mga promoter region was amplified from M1-WT genomic DNA by PCR using primers listed in S1 Table . The mga promoter region from M1-WT and M3-WT share 94% sequence identity , such that results from our EMSAs using serotype M1 DNA are likely applicable to serotype M3 . Purified PCR product was incubated in TBE-buffer with the indicated amount of CovR or CovR phosphorylated with acetyl phosphate at 37°C for 15 min as described [28] . Subsequently , samples were separated on a 5 . 5% TBE-PAA gel for 2h at 120V and stained with ethidium bromide . Per strain and experiment , at least 25 colonies were picked , and the complete covRS operon was PCR amplified and sequenced using the primers listed in S1 Table . Sequences were analyzed with Sequencher 5 . 4 . 6 using the covRS sequence of M3-WT or M1-WT , respectively , as template . Strains were grown in quadruplicate to late-exponential phase in THY . RNA was isolated using the RNeasy minikit ( Qiagen ) . RNA sequencing was performed at the MD Anderson Sequencing and Microarray Facility , and data analysis was performed as described using the M3-WT and MGAS5005 ( an M1 strain ) genome , respectively [22] . A total of 88 out of 1853 ( 4 . 7% ) genes ( M3 serotype ) and 73 out of 1849 ( 4% ) genes ( M1 serotype ) were excluded from the analysis due to low expression levels . Transcript levels were considered significantly different between the isoallelic and wild-type strains if the mean transcript level difference was ≥ 2 . 0 fold and the final adjusted P value was ≤ 0 . 05 . Transcriptome data have been deposited in the GEO database under accession number GSE121313 . Strains were grown as described for RNA seq . Approximately 300 ng RNA per sample was converted to cDNA using a high-capacity reverse transcription kit ( Applied Biosystems ) . TaqMan quantitative real-time PCR ( qRT-PCR ) was performed on an Applied Biosystems Step-One Plus system as described [22] using primers and probes listed in S1 Table . Two biological replicates were performed on two separate occasions ( n = 4 ) . Transcript levels between wild type and derivative strains were compared using a two-sample t test ( unequal variance ) with a P value of ≤0 . 05 following Bonferroni’s adjustment for multiple comparison and a mean transcript level of ≥ 2 . 0-fold change being considered as statistically significant different . | Group A streptococcus ( GAS ) , also known as Streptococcus pyogenes , causes a broad array of human infections of varying severity . Tight control of production of virulence factors is critical to GAS pathogenesis , and the control of virulence two-component signaling system ( CovRS ) is central to this process . The activity of the bifunctional histidine kinase CovS determines the phosphorylation status and thereby the activity of its cognate response regulator CovR . Herein , we sought to determine how varying CovR phosphorylation level ( CovR~P ) impacts GAS pathophysiology . Using three infection models , we discovered that GAS strains lacking CovS phosphatase activity resulting in high CovR~P levels had markedly impaired infectivity . Transcriptome analysis revealed that the hypovirulent phenotype of CovS phosphatase deficient strains is due to down-regulation of numerous genes encoding GAS virulence factors . We identified repression of additional virulence genes that are typically not controlled by CovR , thus expanding the CovR regulon at high CovR~P concentrations . Our data indicate that phosphatase activity of CovS sensor kinase is crucial for spatiotemporal regulation of GAS virulence gene expression . Thus , we propose that targeting the phosphatase activity of CovS sensor kinase could be a promising novel therapeutic approach to combat GAS disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"phosphorylation",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"enzymes",
"pathogens",
"gene",
"regulation",
"enzymology",
"phosphatases",
"animal",
"models",
"model",
"organisms",
"experimental",
"organism",
... | 2018 | Phosphatase activity of the control of virulence sensor kinase CovS is critical for the pathogenesis of group A streptococcus |
The anthelmintic emodepside paralyses adult filarial worms , via a mode of action distinct from previous anthelmintics and has recently garnered interest as a new treatment for onchocerciasis . Whole organism data suggest its anthelmintic action is underpinned by a selective activation of the nematode isoform of an evolutionary conserved Ca2+-activated K+ channel , SLO-1 . To test this at the molecular level we compared the actions of emodepside at heterologously expressed SLO-1 alpha subunit orthologues from nematode ( Caenorhabditis elegans ) , Drosophila melanogaster and human using whole cell voltage clamp . Intriguingly we found that emodepside modulated nematode ( Ce slo-1 ) , insect ( Drosophila , Dm slo ) and human ( hum kcnma1 ) SLO channels but that there are discrete differences in the features of the modulation that are consistent with its anthelmintic efficacy . Nematode SLO-1 currents required 100 μM intracellular Ca2+ and were strongly facilitated by emodepside ( 100 nM; +73 . 0 ± 17 . 4%; n = 9; p<0 . 001 ) . Drosophila Slo currents on the other hand were activated by emodepside ( 10 μM ) in the presence of 52 nM Ca2+ but were inhibited in the presence of 290 nM Ca2+ and exhibited a characteristic loss of rectification . Human Slo required 300nM Ca2+ and emodepside transiently facilitated currents ( 100nM; +33 . 5 ± 9%; n = 8; p<0 . 05 ) followed by a sustained inhibition ( -52 . 6 ± 9 . 8%; n = 8; p<0 . 001 ) . This first cross phyla comparison of the actions of emodepside at nematode , insect and human channels provides new mechanistic insight into the compound’s complex modulation of SLO channels . Consistent with whole organism behavioural studies on C . elegans , it indicates its anthelmintic action derives from a strong activation of SLO current , not observed in the human channel . These data provide an important benchmark for the wider deployment of emodepside as an anthelmintic treatment .
Emodepside is a cyclooctadepsipeptide compound introduced into veterinary medicine for the treatment of nematode infections in companion animals [1 , 2] . It inhibits nematode development [3] and elicits profound impairment of neuromuscular function [4 , 5] . Recently there has been interest in the use of emodepside for the treatment of human helminthiases [6] . In particular its toxicity to adult filarial worms [7] has raised the prospect of it providing a much needed macrofilaricidal anthelmintic drug for the treatment of filarial diseases . Investigation of the mode of action of emodepside revealed two targets , a latrophilin receptor [8 , 9] and a Ca2+-activated K+ channel SLO-1 [10] of which the latter appears to be the most important in terms of the efficacy of the compound at a whole organism level [10–13] . SLO-1 belongs to an evolutionary conserved family of K+ channels that are activated by cell depolarisation and cytosolic calcium . The first member of the family slowpoke ( slo ) was identified in the fruit fly Drosophila melanogaster [14] and the K+ channel was shown to have an unusually large conductance which prompted the name BK for ‘big K+ conductance’ . slo/BK channels were subsequently found across the animal phyla , spanning nematodes to human [15] . Their dual regulation by membrane voltage and intracellular Ca2+ imparts a phylogenetically conserved role in regulating cell excitability ( 9 ) . SLO channels have been extensively studied in the model genetic organism , the nematode Caenorhabditis elegans [13 , 16 , 17] and it has provided an excellent experimental platform in which to investigate the mode of action [10] and selective toxicity [12 , 18] of emodepside . These studies have expressed slo/BK channels from parasitic nematodes and human in a C . elegans slo-1 mutant to evaluate their role in conferring sensitivity to emodepside . These whole organism studies are consistent with the conclusion that emodepside preferentially activates the nematode isoforms of the calcium-activated K+ channel . More recently it has been shown that emodepside activates C . elegans SLO-1A heterologously expressed in Xenopus oocytes [19] . However a direct comparison of the effects of emodepside on heterologously expressed slo/BK channels from different phyla is currently lacking . This information is important to underpin its extended use as an anthelmintic drug . To address this knowledge gap , here we provide the first cross phyla study on the actions of emodepside , on nematode , insect and human channels . We used heterologous expression of recombinant C . elegans , Drosophila and human BK/SLO channels to demonstrate that emodepside directly modulates the nematode , insect and mammalian channels . Importantly we provide mechanistic insight which reveals a differential action of emodepside on the nematode channel , a profound facilitation , which underpins its selective toxicity as an anthelmintic . Taken together our data further support the wider use of emodepside for the treatment of helminthiases . Moreover , they are consistent with BK/SLO channels harbouring a cyclooctadepsidpeptide pharmacophore which could be exploited for further applications in veterinary and human medicine .
C . elegans slo-1a ( Ce slo-1 ) was cloned into a pIRES2-eGFP vector ( BD biosciences , Clontech ) for expression in HEK293 cells . slo-1a cDNA was PCR amplified from the pBK3 . 1 vector [16] and cloned into the pIRES2-eGFP vector . In preliminary experiments we found that high concentrations of pIRES2-eGFP::slo-1a cDNA were required to obtain functional expression of SLO-1 currents in HEK293 cells . Due to simultaneous transcription of egfp this resulted in high levels of expression of eGFP and cells looked unhealthy ( rounded , with dark inclusions ) and were not suitable for reliable electrophysiology . Therefore , we removed the eGFP coding sequence from the pIRES2-eGFP::slo-1a vector and instead used a separate plasmid , at lower concentration , to transfect the cells with the eGFP transformation marker . The mammalian orthologue for SLO-1a ( NP_001024259 ) , KCNMA1 was identified using Ensembl orthologue definitions as previously described [18] . It predicts 20 protein-coding transcripts . The protein product of Ensembl transcript KCNMA1-001 ( peptide ENSP00000286627 ) corresponds to KCNMA1 variant 2/isoform b in the NCBI database ( NP_002238 ) . The alignment and identity of KCNMA1 isoform b with C . elegans slo-1a is described in Crisford et al , [18]The transcript encoding this isoform ( NM_002247 ) was available to purchase from OriGene Technologies , USA . The kcnma1 gene is in the pCMV6-XL4 vector behind a mammalian cell specific pCMV promoter suitable for expression in HEK293 cells . Drosophila slo ( DmSlo ) was amplified by PCR with pfu Polymerase ( Stratagene ) from adult Drosophila melanogaster cDNA ( Clontech ) and cloned into pcDNA3 . 1 ( Invitrogen ) . Human embryonic kidney cells ( HEK293 ) were obtained from European Cell Culture , collection reference 85120602 . Cells were maintained in 25cm2 flasks in Dulbecco's Modified Eagle Medium ( DMEM GlutaMax , Gibco , Life Technologies UK ) supplemented with 10% Fetal Calf Serum ( FCS ) , 1% Penicillin/ Streptomycin ( 100units/100μg/ml ) and L-glutamine ( 2 mM ) . Cells were incubated at 37°C , 5% CO2 . HEK293 cells were passaged every 2–3 days ( when 70–80% confluent ) and kept for a maximum of 20 passages . Plasmid DNA was delivered to the cells using jetPEI transfection reagent ( Polyplus , Source Bioscience Autogen , Nottingham , UK ) . 1 g of DNA to express KCNMA1 or eGFP ( KCNMA1 control ) was mixed with 4 μl of jetPEI and each 1 μg of DNA to express SLO-1 or eGFP ( SLO-1 control ) was mixed with 2 μl of jetPEI . Transfections with eGFP DNA for control recordings were done in the same ratios as the transfections with the genes of interest . Cells were incubated with transfection mixtures for 18 hours at 37°C , 5% CO2 . DNA/JetPEI complexes were then removed and replaced with fresh DMEM/10% FBS . Cells transfected with plasmid DNA to express SLO-1 were transferred to 30°C , 5% CO2 incubator for 24 hours as this enhanced heterologous expression of the [20] invertebrate channel in mammalian cells . Cells transfected with plasmid DNA to express KCNMA1 were maintained at 37°C . CHO cell line ( CHO duk ) was obtained from ATCC , code ATCC CRL-9096 . For transfection with plasmid DNA to express Dm Slo CHO cells were passaged to 40% confluence before adding the transfection solution . The transfection solution contained 300 μL OptiMEM ( Life Technologies , Nr . : 31985 ) , 2 μL ( 6μg ) pcDNA3 . 1 ( - ) Dm Slo and 9μL FugeneHD ( Promega , Nr . : E2311 ) and were added to the cells and incubated for 48 hours at 37°C , 5% CO2 . The transfection medium was exchanged for the selection medium which contains additional G418 ( 2 mg/ml , Invitrogen , Nr . : 10131 ) , the cells were seeded into 384 well plates ( 300 cells/well ) . After a few weeks , the few surviving cells were tested with a voltage sensitive dye ( membrane potential assay kit Dye B , Molecular Devices Nr . : R80034 ) for K+ channel expression . Cells from the best clone ( strongest signal in reaction to a 70 mM K+ application in combination with 3 μM Ca2+ Ionophor A23187 , Sigma Nr . : C7522 ) , were plated out on glass coverslips for testing Dm Slo channel expression electrophysiologically using whole cell voltage clamp technique ( see section whole cell voltage clamping CHO ) . After validation of functional Dm Slo channel expression the best clone was subcloned in 384 well plates ( 0 . 7 cells/well ) in order to obtain clonal purity . The clone with the best membrane potential signal was assessed electrophysiologically for Dm Slo channel function . This generated a final stable CHO cell line expressing the Dm Slo . Whole cell currents from cells transfected with plasmid DNA to express Ce slo-1 were recorded 48 hours after transfection . Whole cell currents from cells transfected with plasmid DNA to express KCNMA1 were recorded 24 hours after transfection . Cells were transferred to the perfusion chamber and recordings were made at room temperature ( 20 to 25°C ) . Patch pipettes ( 4–6 MΩ ) were filled with internal solution containing 140 mM KCl , 1 mM MgCl2 , 2 mM Na2ATP , 3 . 3 mM CaCl2 , 10 mM Hepes , 5 mM EGTA , pH 7 . 2 ( adjusted with KOH ) , in which free [Ca2+] was 300 nM . The calculation of free Ca2+ was made by using the software “MaxChelator Sliders” ( from C . Patton , Stanford University ) and a dissociation constant for the Ca2+-EGTA-complex of 90nM ( pH 7 . 3 ) [21 , 22] . For 100 μM free [Ca2+] internal solution contained: 140 mM KCl , 1 . 2 mM MgCl2 , 2 mM Na2ATP , 5 . 4 mM CaCl2 , 10 mM Hepes , 5 mM EGTA . Osmolarity ranged from 280 to 300 mOsm . The external solution contained 137 mM NaCl , 5 . 9 mM KCl , 2 . 2 mM CaCl2 , 1 . 2 mM MgCl2 , 10 mM Hepes , 14 mM Glucose , pH 7 . 4 ( adjusted with NaOH ) . Osmolarity was 300 mOsm . Transiently transfected cells were identified by eGFP fluorescence . Whole cell currents were evoked by 50 ms voltage steps from -100 mV to +90 mV in 10 mV increments from a holding potential of -60 mV . Alternatively , whole cell currents were evoked by 50 ms voltage steps from -80 mV to +170 mV in 10 mV increments from a holding potential of -60 mV . Cells were perfused with extracellular solution , with or without drug , at a rate of 0 . 5–0 . 8 ml min-1 using a parallel tubing perfusion system driven by an electric pump . The time taken for the drug to access the vicinity of the cell being recorded was 30 to 45 s from the initiation of perfusion . Depolarising voltage steps were applied from -60 mV to +70 mV for 100 ms cells were held at -60 mV for 10 s between depolarising steps in order to analyse the current-voltage relationship during drug applications . Drug or vehicle ( control ) application was initiated when the current elicited by the voltage steps from -60 mV to +70 mV reached a steady state ( at least six current-voltage responses overlapped at +70 mV ) . Whole cell currents were recorded using an Axopatch 200B amplifier ( Axon Instruments , Foster City , CA ) and pClamp 10 software employing Digidata 1440A ( Axon Instruments ) . Data were filtered ( 4-pole Bessel ) at 1 kHz and digitized at 5 kHz . Current-voltage relationships were plotted from the plateau of each depolarising step using Graph Pad Prism computer software ( version 5 . 0 San Diego , California ) . The response to progressive depolarising voltage steps from -60 mV to a maximum of +170 mV was plotted as a time-dependent response . Leak subtraction voltage-command protocols were not applied . Traces for the whole cell currents where leak was more than 10% of the whole current were excluded from the analysis . Whole cell voltage-clamp was performed on CHO cells stably transfected with the Dm Slo channel . Patch pipettes ( borosilicate glass capillaries , 2 . 5–3 . 5 MΩ ) contained ( in mM ) : 150 KCl , 10 HEPES ( pH 7 . 3 adjusted with KOH ) , 10 K-EGTA , 0 , 3 , 5 , or 7 mM CaCl2 . The cells were placed in a perfusion chamber ( 2 . 5ml ) at room temperature ( 22–25°C ) and superfused continuously ( flow rate 3 ml min-1 ) with external bath solution driven by gravity . The fluid in the chamber renewed every 60 s . The external bath contained ( in mM ) : 150 NaCl , 4 KCl , 2 MgCl2 , 2 CaCl2 , 10 HEPES ( pH 7 . 3 adjusted with NaOH ) . Compounds were applied to the cells using the U-tube reversed flow technique . Currents were recorded from a holding potential of -70 mV with an L/M-EPC 7 patch clamp amplifier ( List , Darmstadt , Germany; pClamp software , Axon Instruments , Ver . 6 . 03 , Foster City , CA ) , low-pass Bessel filtered at 3 . 15 kHz and digitized at 5 kHz sample rate . HEK293 cells were incubated with 100 μl of 4 μM Fluo-3AM prepared with Pluronic F-27 ( Molecular Probes , Life Technologies , UK ) at room temperature for 30 min . Following dye removal cells were further incubated at 37°C , 5% CO2 with 180 μl of 1% Bovine Serum Albumin ( BSA ) per well for 30 min , washed with 200 μl of PBS ( phosphate buffered saline ) with subsequent addition of 100 μl of Hank’s Balanced Salt Solution ( HBSS ) ( Gibco , Life Technologies , UK ) containing 10 mM Ca2+ . Ca2+ levels in cells before and after the treatment with emodepside or ionomycin were then quantified by measuring the Ca2+ fluorescence at excitation of 480p and emission of 525p using FLUO Star OPTIMA . The experiment was repeated at least three times with independent cell cultures and the data were pooled for analysis . 10 mM stock solution of emodepside was prepared freshly for each experiment in 100% DMSO ( Dimethylsulfoxide ) . Emodepside was diluted in DMSO and then directly into external solution to give a final concentration of 100 nM , 10 nM and 1 nM ( 0 . 01% DMSO ) . For dissolving relatively high emodepside concentrations ( 1 μM and 10 μM ) 0 . 03% Pluronic F68 ( Sigma , P1300 ) was added as solvent . For injection studies emodepside was dissolved in 100% DMSO given a final concentration of 2 mg/ml . Emodepside was provided by Bayer Animal Health , Monheim Germany . 10 mM stock solution of penitrem A ( Enzo Life Sciences , Exeter , UK ) was prepared in 100% DMSO . Aliquots of stock solution were kept at -20°C . The final concentration of penitrem A was 1 μM ( 0 . 01% DMSO ) . Ionomycin was prepared by diluting 4 μl of ionomycin stock ( 10mM ) in 100% DMSO into 36 μl of HBSS . 1 μl of this ionomycin solution was added to the microtitre cell in 100 μl of HBSS to give a final concentration of 1% DMSO and 10 μM ionomycin . 2% Triton-X 100 was prepared by diluting 10 μl of Triton-X in 490 μl of HBSS and 10 mM Ca2+ solution , 2 . 5 μl of this preparation was added per well ( volume ) . For electrophysiological measurements a 1 mM stock solution of verrucologen ( Cfm Oskar Tropitzsch e . K . , Marktredwitz , Germany ) was prepared in DMSO and diluted into external solution to give required concentrations between 0 . 1 pM and 30 μM just before the experiment . Data points in graphs are presented as the mean ± standard error of the mean for the number of experiments as shown in individual figures . Current-voltage and current-time relationships were plotted using either Graph Pad 5 software ( San Diego , California ) or Origin 6 . 0 Software ( Microcal Software Inc . , Northampton , MA , USA ) . Statistical significance was determined either by unpaired Student’s t-test , one-way or two-way ANOVA as appropriate; significance level set at P< 0 . 05 , followed by Bonferroni post-hoc tests as appropriate . Boltzmann analysis was performed using GraphPad Prism , version 6 . 05 ( San Diego , California ) and the equation G/Gmax = 1/ ( 1+exp ( ( V50-Vm/slope ) ) . Conductance , G , for each membrane potential ( Vm ) was calculated using an equilibrium potential for K+ of -80 mV from the equation G = I/Vm-EK . Gmax was defined as the average maximal conductance for each experimental group . Values are given with 95% confidence intervals .
Ce slo-1 ( C . elegans slo-1 ) was expressed in HEK293 cells in the presence of different intracellular Ca2+ concentrations . Voltage-activated K+ currents of very low amplitude were recorded in 300 nM free intracellular Ca2+ ( Fig 1A and 1B; mean peak current 0 . 49 ± 0 . 05 nA , n = 25 ) and were only marginally higher than the current recorded from control eGFP transfected cells ( 0 . 33 ± 0 . 03 nA; n = 13; P = 0 . 0192 unpaired Student’s t-test ) . Previously it has been reported that C . elegans slo-1 requires micromolar Ca2+ for activation [16 , 23 , 24] . Therefore , we tested the slo-1 transfected cells in the presence of 100 μM Ca2+ . Under this condition the voltage-activated K+ currents recorded from HEK293 cells were markedly increased ( Fig 1A and 1B; mean peak current for SLO-1 in 100 μM Ca2+ was 3 . 48 ±0 . 19 nA; n = 52 ) . Increasing the intracellular Ca2+ to 100μM had no significant effect on the currents recorded from cells expressing the control egfp plasmid ( Fig 1A and 1B; mean peak current at +90 mV 0 . 30 ± 0 . 04 nA , n = 14; P = 0 . 57 unpaired Student’s t-test compared to egfp expressing cells recorded with 300 nM Ca2+ ) indicating that the effect of 100μM Ca2+ in facilitating channel activation is specific to the slo-1 transfected cells . It is established that mammalian BK channels can be activated by Ca2+ with a Kd of 0 . 8–11 μM [25] and as low as 0 . 5–100 nM [26] . Consistent with this , and in contrast to the Ca2+ dependence observed for slo-1 transfected cells , we observed robust voltage-activated K+ currents from cells expressing kcnma1 with 300 nM intracellular Ca2+ ( Fig 1A and 1B; mean peak current at +90mV of 2 . 74 ± 0 . 11nA , n = 50; P<0 . 001 compared to egfp alone , unpaired Student’s t-test ) . Similarly CHO cells expressing Dm Slo ( Drosophila melanogaster Slo ) showed strong activation of voltage-activated K+ currents at low free intracellular Ca2+ ( Fig 2A and 2B ) with a threshold around 52 nM and robust activation at 290 nM . Therefore , in further experiments to characterise all three channels we routinely used 100 μM , 300 nM and 290 nM free intracellular Ca2+ for Ce slo-1 , human kcnma1 and Dm Slo respectively in order to obtain robust currents of similar amplitude for each of the channels against which to make the most accurate comparison of the effects of emodepside on the different isoforms of channel . We confirmed the identity of the Ce slo-1 , kcnma1 and Dm Slo channel currents by testing their sensitivity to selective BK channel blockers penitrem A [27] or verruculogen [28] ( Fig 3A and 3B; Fig 4 ) . Only cells for which a consistent current-voltage relationship was obtained for the first 2 min of recording were subjected to drug application . For C . elegans SLO-1 in which recordings were made with 100 μM intracellular Ca2+ such stable recordings were only achieved for 10% of the total number of transfected cells from which whole cell recordings were made , most likely reflecting the impact that the high Ca2+ concentration has on the integrity of the cells . Once a stable recording was established the drugs were applied and recordings were made throughout the drug application for a period of up to 15 min . This was the maximum duration of recording that could be made for SLO-1 as after this time the recordings became unstable as indicated by the holding current and the cells were routinely lost . Visual inspection of the cells indicated a shrunken appearance , presumably due to the high ( 100 μM ) intracellular Ca2+ that was required for these recordings . For SLO-1 , the recordings were interleaved with controls on the same day; control recordings were made from transfected cells that were perfused with vehicle control ( 0 . 01% DMSO ) . Both SLO-1 and KCNMA1 currents were blocked by penitrem A ( 1 μM; Fig 3A and 3B ) . The mean peak current at +90 mV for SLO-1 was 3 . 95 ± 0 . 73 nA before and 0 . 63 ± 0 . 17 nA after 10 min of treatment with penitrem A ( n = 7 ) . For KCNMA1 the mean peak current at +90 mV was 2 . 66 ± 0 . 20nA before and 0 . 53 ± 0 . 09 nA after 10 min of treatment with penitrem A ( n = 7 ) . Similarly , in CHO cells expressing Dm Slo the voltage-activated K+ currents elicited in the presence of 290 nM free intracellular Ca2+ were completely blocked by the BK channel antagonist verruculogen at 3 μM ( Fig 4; n = 5 ) and the dose response curve revealed an IC50 value of 6 . 2 nM for inhibition of Dm Slo by verruculogen ( 95% confidence intervals 2 . 52 to 8 . 57 nM ) and an IC50 value of 0 . 5 nM for inhibition of Dm Slo by penitrem A ( 95% confidence intervals 0 . 43 t0 12 . 9 nM; Fig 4 ) . The same protocol that was optimised to determine the penitrem A inhibition of KCNMA1 was modified to test for an interaction of emodepside with the channels . Thus , for these experiments emodepside ( 1 , 10 or 100 nM ) or DMSO ( 0 . 01% ) was applied to the cells after 2 min initial baseline recording and recordings were followed for up to a further 23 min i . e . as long as the stability of the recording permitted . This range of concentrations of emodepside was chosen as the EC50 for the inhibitory effect of emodepside on C . elegans behaviours is between 10 and 100 nM [10 , 18] . In HEK293 cells expressing KCNMA1 emodepside elicited a biphasic , concentration-dependent effect on current amplitude ( Fig 5A , 5B and 5C ) . The effect was of a similar magnitude for 10 nM and 100 nM emodepside but was only statistically significant for 100 nM emodepside . The early facilitation was transient , lasting less than ten min , and followed by an inhibition of current amplitude . Again the inhibition was observed at both 10 and 100 nM emodepside but was only statistically significant for 100 nM emodepside . The mean peak currents for KCNMA1 at +90 mV were increased from 2 . 94 ± 0 . 36 nA ( n = 8 ) to 4 . 31 ± 0 . 34 nA ( n = 8 ) after 5 min emodepside and were then reduced to 1 . 67 ± 0 . 26 nA ( n = 7 ) after 23 min of emodepside . Whole cell currents recorded from transfected cells treated with 0 . 01% DMSO did not change significantly over a 25 min time-course . The time-course of the response to emodepside was slow in onset and slow to reach the peak effect similar to the time-course of the effect of emodepside reported in other systems [13] . Given the effect of 100 nM emodepside on hum KCNMA1 ( huma KCNMA1 ) we decided to test the effect of this concentration on C . elegans SLO-1 currents . We were constrained to testing this concentration of emodepside rather than a dose range because of the technical difficulty of recording SLO-1 currents from HEK293 cells with 100 μM Ca2+ . Notably , for Ce SLO-1 the current amplitude was increased , but not inhibited , by 100 nM emodepside ( Fig 6A and 6B ) . The facilitation followed a slow time-course and continued to increase throughout the time-course of the experiment ( Fig 6B ) . The mechanisms that might underpin a slow time-course of action for emodepside , pharmacokinetic versus mechanistic , have previously been discussed in Holden-Dye et al [13] . The mean peak currents at +90 mV for Ce SLO-1 recordings were 2 . 95 ± 0 . 35nA before and 4 . 39 ± 0 . 55nA after 10 min of treatment with emodepside ( n = 9; Fig 6B ) representing a facilitation of 73 . 0 ± 17 . 4% ( n = 9; p<0 . 001 compared to DMSO control ) . Over the same time-course the 0 . 01% DMSO control treated transfected cells exhibited mean peak currents at +90 mV of 3 . 14 ± 0 . 36nA before DMSO compared to 2 . 17 ± 0 . 25nA after DMSO ( n = 5 ) . The voltage activation of Ce SLO-1 and hum KCNMA1 was tested before and during the peak of the emodepside facilitation ( i . e . after 12 min application of 100 nM emodepside for Ce SLO-1 and after 5 min application of 100nM emodepside for hum KCNMA1; Fig 7 ) . A significant shift in the voltage activation curve was observed for Ce SLO-1 ( p<0 . 05 ) after emodepside treatment and a smaller significant shift for hum KCNMA1 ( p<0 . 05 ) . The data from Fig 7 were subjected to Boltzmann analysis to estimate V50 , the membrane potential for half-activation of the channels using Gmax values of 0 . 0174 , 0 . 0258 , 0 . 0177 and 0 . 0253 μS for Ce SLO-1 control , Ce SLO-1 with emodepside , hum KCNMA1 control and hum KCNMA1 with emodepside , respectively . For Ce SLO-1 V50 was +49 . 74 mV ( 95% confidence intervals 45 . 46 to 54 . 01; V slope 20 . 66 ) and this was shifted in the presence of emodepside to +32 . 95 mV ( 95% confidence intervals 28 . 82 to 37 . 07; Vslope 28 . 36 ) . For hum KCNMA1 the V50 was +56 . 51 mV ( 95% confidence intervals 52 . 82 to 60 . 20; Vslope 17 . 45 ) which was unchanged in the presence of emodepside , + 57 . 88 mV ( 95% confidence intervals 54 . 09 to 61 . 68; Vslope 17 . 12 ) . This analysis confirms the selective effect of emodepside on the voltage-activation for Ce SLO-1 . Recordings made from CHO cells expressing Dm Slo revealed complex effects of emodepside ( 10 μM ) . In the presence of low ( 52 nM ) Ca2+ , a situation in which voltage steps do not activate K+ currents , emodepside revealed a K+ current ( Fig 8A , middle panel ) . However , in the presence of higher ( 290 nM ) Ca2+ , a situation in which the voltage steps generate typical outwardly rectifying K+ currents , emodepside inhibited the amplitude of the outward current ( Fig 8B , bottom panel ) . We also noted , that in the presence of either 52 nM or 290 nM Ca2+ , emodepside caused rapid membrane hyperpolarisation ( n = 15; emodepside shifted the resting membrane potential from -55 ± 6 to -79 ± 3 mV , mean ± s . e . m . ) . During recordings made in current clamp mode we observed that a few minutes after emodepside application the membrane potentials reach steady-state values which were close to the K+ equilibrium potential ( Em = -79 . 8 ± 6 . 5 mV , n = 15; EK = -91 . 2 mV ) . Moreover , test pulses in the presence of emodepside for both 52 nM and 290 nM Ca2+ revealed linear current-voltage behaviour over the voltage range from -120 to + 60 mV in all experiments ( n = 15; Fig 8A and 8B ) . Whilst activation and inhibition of the voltage-activated K+ current by emodepside was modified by the Ca2+ concentration , the emodepside-induced modification of the current-voltage relationship was largely independent of intracellular Ca2+ concentration . Thus in the presence of low intracellular Ca2+ ( 52 nM ) where the Dm Slo channel was closed at most voltages and also at higher intracellular Ca2+ ( e . g . 290nM ) where the Dm Slo channel shows outward rectification , emodepside consistently induced a linear current-voltage behaviour of the Dm Slo channel ( Fig 8B ) indicating a reduction in the rectifying properties of the channel without any shift in reversal potential . It was important to substantiate that the actions of emodepside are mediated via a direct interaction with the channel rather than indirectly via an increase in intracellular Ca2+ , possibly by acting as an ionophore [29] . In order to test this HEK293 cells were loaded with the Ca2+ fluorophore Fluo-3 , AM and exposed to emodepside . Whilst a positive control , treating cells with ionomycin ( 10 μM ) , elicited a robust increase in Ca2+ fluorescence , no increase in intracellular Ca2+ was observed even at high micromolar concentrations of emodepside ( Fig 9 ) .
In this study we have investigated the cyclooctadepsipeptide compound emodepside as a novel modulator of the BK/SLO-1 Ca2+ -activated K+ channels from nematode , insect and human . We established the requisite Ca2+ concentration for channel activation in order to establish experimental conditions in which to investigate the ability of emodepside to act as a channel modulator . The nematode Ca2+ -activated K+ channel differs from the insect and human channel in terms of its Ca2+ sensitivity . Thus we observed that 100μM Ca2+ was required to record Ce SLO-1 currents in contrast to 290–300 nM for Dm Slo and hum KCNMA1 . This is consistent with the micromolar concentration of Ca2+ used to characterise the C . elegans SLO-1 channel in Xenopus oocyte expression studies [16] and patch clamp recordings from C . elegans body wall muscle which showed that 100μM internal Ca2+ concentration is required to observe SLO-1 channel openings [30] . The gating of the channel by Ca2+ is regulated by the gating ring [31] and thus these observations suggest functional differences between the nematode channel compared to human and insect in this region . The requirement of the nematode channel for high micromolar concentrations of Ca2+ for activation indicates that physiological regulation may be achieved only by close juxtaposition of the K+ channel with voltage-gated Ca2+ channels in the plasma membrane so that local , transient elevations in cytosolic Ca2+ may activate the channel as suggested by the Ca2+ domain hypothesis [32] . Emodepside showed efficacy at C . elegans SLO-1 and human KCNMA1 channels heterologously expressed in HEK293 cells . However , whilst emodepside affected both SLO-1 and KCNMA1 currents , there were distinct differences in the effect observed . Emodepside enhanced C . elegans SLO-1 currents by up to 80% and shifted the voltage-activation to a lower threshold . This resembles the effect of increased Ca2+ concentration on the voltage-activation range of BK channels [16 , 23 , 24 , 33] . In contrast , KCNMA1 currents were subject to biphasic modulation by emodepside , an initial transient facilitation followed by a more prolonged inhibition . Moreover , the maximal facilitation exerted by 100 nM emodepside was only 30% , considerably less than that observed for SLO-1 . This difference in the efficacy of direct modulation of the nematode and the human channel by emodepside is consistent with the selective toxicity of the compound as an anthelmintic . There is corroboration for this from experiments in which either nematode slo-1 or human kcnma1 were expressed in C . elegans slo-1 null mutants which were otherwise insensitive to emodepside . Whilst expression of nematode slo-1 conferred sensitivity to emodepside on the slo-1 null mutants [12 , 18] , 10 to 100 fold higher concentrations of emodepside were required to impair mobility in transgenic lines expressing the human channel [18] . We also noted that the qualitative features of the impaired mobility caused by emodepside in the lines expressing the human channel were different from the effects seen in lines expressing C . elegans slo-1 . In the latter a flaccid paralysis was observed when worms were exposed to low concentrations of emodepside and this effect was phenocopied by a slo-1 gain of function mutant consistent with potent activation of the channel by emodepside [11] . In contrast , micromolar concentrations of emodepside applied to C . elegans expressing the human channel KCNMA1 impaired mobility but did not elicit a flaccid paralysis [18] . This led us to suggest that emodepside at micromolar concentrations is an inhibitor of the mammalian channel [18] and this is directly supported by the inhibition of KCNMA1 currents observed here . Whilst the inhibition of KCNMA1 by emodepside may be of relevance in predicting side-effects that might be encountered in its future use in tropical medicine . The experiments with Dm Slo provide evidence for an additional effect of emodepside on Ca2+ -activated K+ channels , namely an ability to diminish the rectifying properties of the channel . This effect of emodepside on Dm Slo was independent of the intracellular Ca2+ concentration . Thus , in either low or high Ca2+ , emodepside exposure led to the appearance of inward current i . e . there was a loss of the characteristic outward rectifying properties of the channel . It will be interesting to explore whether this phenomenon is observed in isoforms of slo channels from species of nematodes other than C . elegans . The effect of emodepside on nematode muscle has been extensively investigated [4 , 9] and the observations made from muscle recordings are consistent with those from the recombinant expressed channels described here . Electrophysiological studies on the parasitic nematode Ascaris suum have shown that a cyclooctadepsipeptide compound , PF1022A , causes a Ca2+ and K+-dependent relaxation of body wall muscle [4] . Latterly it was shown that emodepside increases activation of voltage-dependent K+ currents in the muscle of A . suum [5] , again consistent with the shift in voltage-activation seen in heterologously expressed slo-1 observed in this study . BK channels , including C . elegans SLO-1 , assemble with accessory subunits in vivo and it is possible these may impact on emodepside modulation [24] but the similar actions of emodepside on the heterologously expressed channels and currents recorded from native tissue argue that this is unlikely to play a significant role . Our observations , and previous reports , indicate that the modulation of the Ca2+ -activated K+ channels in HEK293 cells by emodepside cannot be ascribed to an elevation of intracellular Ca2+ . The parent compound for emodepside , the cyclooctadepsipeptide PF1022A , does not increase intracellular Ca2+ in CaCo-2 cells [34] . Here we conducted a similar study on HEK293 cells and showed that a supramaximal concentration of emodepside ( 100μM ) had no significant effect on the intracellular concentration of Ca2+ . There is evidence from A . suum that in vivo emodepside may act indirectly to modulate SLO-1 via an intracellular signalling cascade involving nitric oxide and protein kinase C [5] . However , given that we have shown that emodepside differentially modulates Ce SLO-1 , Dm Slo and hum KCNMA1 expressed in HEK293 cells , this is an unlikely explanation for the modulation we have observed . Whilst it cannot be discounted that additional indirect mechanisms might operate in vivo , our data indicate that emodepside exerts a modulatory action on SLO-1 and KCNMA1 that is independent of a change in intracellular Ca2+ and likely reflects a direct modulation of the channel . Taken together , our data substantiate a direct interaction of emodepside with C . elegans slo-1 [19] . This first cross-phyla analysis of C . elegans , human and Drosophila slo/BK indicate a complex modulation of these Ca2+ -activated K+ channels by emodepside in which only the nematode channel exhibits a marked facilitation . This is underpinned by a shift in voltage-activation and provides molecular insight into its potency as an anthelmintic . The data reinforce the impetus to investigate the utility of this compound , or related drugs which target slo channels , for the treatment of human helminthiases . Recently it has been shown that there is a synergistic action of the filaricidal compound diethylcarbamazine with emodepside on Ascaris suum muscle slo currents and it would be interesting to investigate the mechanism of this in heterologously expressed slo channels [35] . From a broader perspective , the effects of emodepside reported here show that the cyclooctadepsipeptides provide a route to understanding a new pharmacophore harboured by BK/SLO-1 channels [13] which has important relevance for the therapeutic exploitation of this target . | Filarial diseases affect an estimated 200 million people and the Drugs for Neglected Diseases initiative ( DNDi ) has identified development of macrofilaricidal drugs as a priority . Emodepside , currently used in companion animals , paralyses adult filarial worms and may address this unmet need for human medicine . Its receptor is an evolutionary conserved Ca2+-activated K+ channel , SLO-1 . In this paper we address an important knowledge gap in terms of understanding the interaction of emodepside with its target receptor SLO-1 in nematodes in comparison to the human orthologue KCNMA1 and provide the first cross phyla analysis of the interaction of emodepside with slo channels , in nematode , insect and human . Intriguingly , this shows that emodepside modulates slo/BK currents from heterologously expressed channels from all three organisms , however there are discrete differences in the feature of modulation; only the nematode channel exhibits a sustained facilitation by emodepside . This is consistent with the effects of emodepside on C . elegans behaviour and indicates that this differential action of emodepside on the nematode channel likely underlies its potent anthelmintic effects . These data provide an important benchmark for the wider deployment of emodepside as an anthelmintic treatment . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | The Cyclooctadepsipeptide Anthelmintic Emodepside Differentially Modulates Nematode, Insect and Human Calcium-Activated Potassium (SLO) Channel Alpha Subunits |
Insect molting and metamorphosis are intricately governed by two hormones , ecdysteroids and juvenile hormones ( JHs ) . JHs prevent precocious metamorphosis and allow the larva to undergo multiple rounds of molting until it attains the proper size for metamorphosis . In the silkworm , Bombyx mori , several “moltinism” mutations have been identified that exhibit variations in the number of larval molts; however , none of them have been characterized molecularly . Here we report the identification and characterization of the gene responsible for the dimolting ( mod ) mutant that undergoes precocious metamorphosis with fewer larval–larval molts . We show that the mod mutation results in complete loss of JHs in the larval hemolymph and that the mutant phenotype can be rescued by topical application of a JH analog . We performed positional cloning of mod and found a null mutation in the cytochrome P450 gene CYP15C1 in the mod allele . We also demonstrated that CYP15C1 is specifically expressed in the corpus allatum , an endocrine organ that synthesizes and secretes JHs . Furthermore , a biochemical experiment showed that CYP15C1 epoxidizes farnesoic acid to JH acid in a highly stereospecific manner . Precocious metamorphosis of mod larvae was rescued when the wild-type allele of CYP15C1 was expressed in transgenic mod larvae using the GAL4/UAS system . Our data therefore reveal that CYP15C1 is the gene responsible for the mod mutation and is essential for JH biosynthesis . Remarkably , precocious larval–pupal transition in mod larvae does not occur in the first or second instar , suggesting that authentic epoxidized JHs are not essential in very young larvae of B . mori . Our identification of a JH–deficient mutant in this model insect will lead to a greater understanding of the molecular basis of the hormonal control of development and metamorphosis .
The number of larval instars in insects varies greatly across insect taxa , and can even vary at the intraspecific level [1] , [2] , [3] . In general , phylogenetically higher insects tend to have fewer larval instars ( three to eight ) compared to species from basal lineages , such as Ephemeroptera , Odonata and Plecoptera ( more than ten ) [1] , [2] , [3] . In many species , the number of larval instars is affected by genetic and environmental factors , such as temperature , nutritional conditions , photoperiod , humidity , injuries , and sex [1] , [2] . The variation in the number of larval instars in the insect lifecycle is generally considered to be an adaptive response to diverse environmental conditions in order to ensure the attainment of a threshold-size for metamorphosis [1] , [2] , [3] , [4] . The silkworm Bombyx mori , a classic model organism for endocrinology , has been reared by humans for thousands of years , and more than 1 , 000 strains are currently maintained [5] , [6] , [7] . Among these , several “moltinism” strains have been identified that exhibit variations in the number of larval instars [6] , [7] . Silkworms typically have five larval instars , but the moltinism strains vary between three and seven [6] , [7] . For example , precocious larval-pupal metamorphosis is observed in the mod ( dimolting , chromosome 11–27 . 4 cM ) , rt ( recessive trimolting , 7–9 . 0 ) and M3 ( Moltinism , 6–24 . 1 ) strains , while extra larval molting is observed in the M5 ( Moltinism , 6–24 . 1 ) strain [6] , [7] . To date , however , none of these loci has been characterized at the molecular level . Given the availability of whole genome data and post-genomic tools in B . mori [8] , [9] , [10] , these strains offer a valuable resource for elucidating the molecular mechanism that underlies plasticity in the number of larval instars . Here we report the identification and characterization of the gene responsible for the mod mutation that causes precocious larval-pupal metamorphosis in the third or fourth instar [11] . Most mod larvae form larval-pupal intermediates , but some individuals can become miniature moths with normal fertility . Thus , the mod mutant strain can be maintained as homozygous stocks [6] , [11] , [12] . We demonstrate that the mod locus encodes CYP15C1 , a cytochrome P450 involved in the biosynthesis of juvenile hormones ( JHs ) , whose “status quo” action allows the progression of multiple larval-larval molting until the larva attains the required size for metamorphosis [13] , [14] , [15] . CYP15C1 is specifically expressed in the corpus allatum ( CA ) , an endocrine organ that produces and secretes JHs . Enzymological analysis revealed that CYP15C1 converts farnesoic acid ( FA ) to JH acid ( JHA ) in a highly stereospecific manner . We further demonstrated that CYP15C1 plays an indispensable role in JH biosynthesis , and its molecular defect results in the loss of JHs in the hemolymph , thereby causing precocious metamorphosis in the mod strain . Remarkably , precocious larval-pupal transitions in mod larvae always occur after the larval third instar , but not in the first or second instar . Our data provide further evidence supporting the hypothesis that authentic ( epoxidized ) JHs are essential for the classic “status quo” molting in late larval stages ( third and fourth instar ) , but not in early larval stages ( first and second instar ) of B . mori [16] .
Larvae of standard B . mori strains undergo molting four times and thus have five larval instars; these larvae are conventionally termed “tetramolter” in silkworm genetics . The spontaneous mutant mod was identified in a standard strain [11] and mod larvae undergo precocious metamorphosis in the third ( dimolter ) or fourth instar ( trimolter ) . First , we obtained a detailed developmental profile of larvae from two batches of the mod strain . All mod larvae underwent precocious metamorphosis in the fourth instar and no individuals reached the fifth instar ( Figure 1A and 1B ) . We plotted the timing of the onset of spinning in the mod larvae ( Figure 1C and 1D ) . Consistent with previous reports [11] , [12] , we found that spinning occurred at two clearly distinguishable timings: ( 1 ) from 56 to 80 h and ( 2 ) from 112 to 144 h after the third molt: these larvae were termed early- and late-maturing trimolters , respectively . This segregation in the timing of the onset of spinning was not observed in the standard strain p50T ( Figure 1C ) or other moltinism strains [11] , and thus is a unique characteristic of the mod strain . Importantly , development in almost all early-maturing trimolters was arrested and they remained as larval-pupal intermediates ( 93 . 4% , 85/91 larvae ) ; only 3 of the 91 larvae ( 3 . 3% ) successfully survived to adulthood ( Figure 1B ) . In contrast , the late-maturing trimolters did not show such severe developmental impairment and 88 . 5% ( 77/87 ) became miniature adults ( Figure 1B ) . In the larval-pupal intermediates , we usually observed prothetelic phenotypes such as a mixed pupal cuticle on the exoskeleton of animals having overall a larval appearance ( Figure 1A ) , suggesting that hormonal switching of molting and metamorphosis may be aberrant in the mod strain . Notably , despite their small body size , reproduction in mod moths seemed normal , and their eggs hatched without apparent abnormalities . In the silkworm , premature metamorphosis can be induced by the loss of or low levels of JH signaling , which can occur due to the surgical removal of the CA [17] or to overexpression of the JH-degrading enzyme [16] . We therefore hypothesized that precocious metamorphosis in the mod strain was caused by the prevention of JH biosynthesis or signaling . To examine this hypothesis , we first determined whether the mod phenotype could be rescued by treatment with methoprene , a JH analogue . We topically applied several doses of methoprene to newly-molted third or fourth instar mod larvae and found that a fourth larval molting was induced by the treatment ( Figure 1E ) . Fifth instar larvae that had undergone fourth larval ecdysis grew normally , began to spin after ∼6 days , and eventually metamorphosed to pupae and adults that were normal and fertile . This result suggests that JH reception and subsequent JH signaling is normal in the mod strain . Therefore , we next compared the JH titers in the hemolymph of third instar larvae of mod and p50T strains at 24 h after molting to the third instar . JHs were extracted from the hemolymph and their methoxyhydrin derivatives were analyzed by liquid chromatography-mass spectrometry ( LC-MS ) . We detected JH I and JH II in the hemolymph of p50T , whereas the JH titer in the hemolymph of the mod strain was below the detectable level ( Figure 1F ) . These results indicate that the mod strain is a JH-deficient mutant in which complete ( or almost complete ) loss of JH caused precocious metamorphosis . To identify the gene responsible for the mod locus , we performed positional cloning using backcross 1 progeny ( BC1 ) obtained from crossing females of the mod strain ( t011 strain , see http://www . shigen . nig . ac . jp/silkwormbase/index . jsp ) with F1 heterozygote males of mod and p50T strains ( see Figure S1 ) . We mapped the mod locus within ∼400 kb region on the scaffold Bm_scaf16 ( chromosome 11 ) [8] using 792 BC1 individuals . Twenty-five genes were predicted to be present within this region . Among them , we focused on BGIBMGA011708 , a gene encodes a cytochrome P450 monooxygenase . Based on sequence homology and phylogenetic analysis ( Figure 2B ) , the gene was designated as CYP15C1 . We found that CYP15C1 shares high homology with the CYP15A1 of the cockroach Diploptera punctata , which is involved in JH biosynthesis in CA of the cockroach [18] . Given that the mod phenotype is a result of the loss of the JH titer ( Figure 1F ) , we speculated that the mod phenotype is due to the loss of function of CYP15C1 . To examine this possibility , we first determined the nucleotide sequence of the full-length CYP15C1 cDNA from p50T and mod strains . We identified a 68-bp deletion in the mod allele that introduces a premature stop codon in the coding region of CYP15C1 ( Figure 2C–2E ) . This deletion seemed to produce a functionally null mutation in CYP15C1 , since a heme-binding motif , which is essential for enzymatic activities in P450s [19] , was eliminated in the mod allele ( Figure 2D ) . This result indicates that CYP15C1 is a strong candidate for the mod locus . Therefore , we further characterized CYP15C1 and its gene product . The strict regulation of JH biosynthesis in CA is critical for the successful development and reproduction of insects [14] , [15] , [20] . We next examined the spatial expression pattern of CYP15C1 mRNA . We examined 12 tissues at four different developmental stages and found that CYP15C1 mRNA was highly specific to the corpus cardiacum ( CC ) -CA complex ( Figure 3A ) . A whole mount in situ hybridization experiment in the brain ( Br ) -CC-CA complex ( Figure 3B and Figure S2 ) showed that the signal for CYP15C1 was strictly limited to CA , where JH is synthesized , and could not be detected in the brain or CC . These results showed a close spatial correlation between CYP15C1 expression and JH biosynthesis . Next , we carried out a detailed analysis of the temporal expression pattern of CYP15C1 in the CC-CA complex and compared it to that of the gene for JHA methyltransferase ( JHAMT ) , a key enzyme that acts in the final step of the JH biosynthetic pathway in CA [21] . CYP15C1 mRNA was constitutively expressed in CA from the first instar larval to adult stages ( Figure 3D ) , even when JH is not synthesized ( Figure 3C ) [20]; no apparent differences in levels of CYP15C1 mRNA were observed between males and females during pupal and adult stages ( Figure 3D ) . In contrast , the temporal expression pattern of JHAMT correlates well with the JH synthetic activity of CA ( Figure 3D and Figure S2 ) . JHAMT transcript completely disappeared by day 4 of the fifth instar when CA ceased production of JH ( see Figure 3C ) . It reappeared from the mid-pupal stage and increased to a very high level in the female CA . This was consistent with the temporal profile of JH biosynthesis activity in CA as this occurs only in females during the pupal and adult stages [20] . Taken together , our results strongly indicate that CYP15C1 is involved in JH biosynthesis in CA , but does not appear to act as a rate-limiting factor for JH biosynthesis . The cockroach CYP15A1 , the ortholog of B . mori CYP15C1 , catalyzes the epoxidation of ( 2E , 6E ) -methyl farnesoate ( MF ) to JH III [18] . Although biochemical studies predicted the presence of FA epoxidase in the CA of the lepidopteran insect Manduca sexta [22] , [23] , the corresponding gene has not been identified to date . Therefore we examined the enzymatic activity of B . mori CYP15C1 against two plausible substrates , FA and MF . First , we employed a transient expression system using Drosophila S2 cells . When S2 cells expressing CYP15C1 were incubated with medium containing FA , a major HPLC peak was generated that had the same retention time ( 15 . 1 min ) as standard JH III acid ( JHA III ) ( Figure 4A , middle ) . This peak did not appear when S2 cells expressing GFP were used ( Figure 4A , bottom ) . The ESI-MS spectrum of this peak gave an [M-H]− at m/z 251 , consistent with the C15H24O3 formula of JHA III , confirming that CYP15C1 catalyzes the conversion of FA to JHA III . The enzymatic properties of CYP15C1 was further examined in a stable Sf9 cell line ( Sf9/BmCYP15C1 ) that constitutively expresses CYP15C1 . When the Sf9/BmCYP15C1 cells were cultured in medium containing FA , significant levels of JHA III were detected; in contrast , JHA III production was difficult to detect when original Sf9 cells were used ( Table S2 , Exp . 1 ) . When Sf9/BmCYP15C1 cells were cultured in medium containing MF , JH III generation was detected at low levels . However , a similar level of JH III production was also detected in the original Sf9 cells when they were cultured in the same medium ( Table S2 , Exp . 1 ) . These results suggest that JH III production observed in Sf9/BmCYP15C1 was might be due to the presence of endogenous P450 epoxidases in Sf9 cells , which have been reported previously to have lower substrate specificity and stereospecificity [18] , [24] . The addition of the JH esterase inhibitor 3-octylthio-1 , 1 , 1- trifluropropan-2-one ( OTFP ) did not increase production of JHA III ( Table S2 , Exp . 2 ) , indicating that the degradation of JH III by intrinsic JH esterases in the cells was negligible . Therefore , we were able to estimate the conversion ratio of FA and MF to JH III by CYP15C1 . This showed that CYP15C1 exhibited at least 18-fold higher activity for FA than MF ( Table S2 , Exp . 1 ) , a result that is consistent with previous biochemical studies on lepidopteran FA epoxidase in CA . To further examine the stereospecificity of CYP15C1 , the JHA III generated by Sf9/CYP15C1 was chemically methylated and analyzed by a Chiral-HPLC . The methylated product had a major ( R ) -JH III and a minor ( S ) -JH III peak ( R∶S = 97∶3 ) ( Figure 4B ) . These results show that B . mori CYP15C1 encodes a functional P450 epoxidase that preferentially converts FA to JHA III rather than MF to JH III , and does so in a highly ( R ) -enantioselective manner ( Figure 4C ) . To obtain direct evidence that CYP15C1 is responsible for the mod mutation , we performed transgenic rescue experiments using the GAL4/UAS system [25] . We generated transgenic silkworm lines carrying the UAS-CYP15C1 transgene with the eye-specific 3xP3-EGFP marker [26] . The UAS-CYP15C1 transgene was driven using a silkworm enhancer trap line ET14 in which GAL4 was strongly expressed in CA ( Figure 5A ) , although weak expression was also detected in peripheral tissues including fat bodies and the midgut [9] , [27] . As these lines were generated using the standard Shiro-C ( w-1; +mod ) strain , we changed the genetic background to w-1/w-1; mod/mod by crossing to the mod strain . The resultant w-1; mod; ET14/+ females were then crossed with w-1; mod; UAS-CYP15C1/+ males to determine whether the mod phenotype could be rescued by CYP15C1 overexpression . We used two independent UAS-CYP15C1 lines with ET-14 ( Figure 5B ) . In both UAS-CYP15C1 lines , CYP15C1 overexpression efficiently prevented precocious metamorphosis and 97 . 1% of the larvae ( 34/35 in total ) underwent the fourth larval molt to become fifth instar larvae ( Figure 5B and 5C ) . Only one larva ( 1/35 ) became a late-maturing trimolter , but neither dimolters nor early-maturing trimolters appeared . This result was in contrast to what was observed in control larvae or larvae carrying either the GAL4 or UAS construct alone: approximately half of the larvae became dimolters and the remainder became trimolters , while no larvae became tetramolters . We also measured the JH titer in the hemolymph ( Figure 5D ) . As expected , the JH titers in control , ET14 , and UAS larvae were below the detectable limit . In contrast , we were able to detect JH I and JH II in the hemolymph of mod larvae carrying both ET14 and UAS-CYP15C1 constructs . Taken together , these results provide direct evidence that CYP15C1 is responsible for the mod mutation and is essential for JH biosynthesis .
JH III is the most common JH in many insect orders , although its ethyl-branched homologs ( JH I and II ) are the major JHs in the order Lepidoptera [22] , [28] . Biochemical studies have shown that in the late steps of JH biosynthesis in many insect species , including cockroaches and locusts , FA is first methylated to MF and then epoxidized to JH III in CA [22] . However , the final two steps of JH biosynthesis are reversed in Lepidoptera: ethyl-branched homologs of FA ( homo-FAs ) are first epoxidized and the resultant JHAs ( i . e . , JHA I and II ) are then methylated to the authentic JHs ( i . e . , JH I and II ) [22] . This study showed that B . mori CYP15C1 epoxidizes FA to JHA III in a highly stereospecific manner . CYP15C1 might also epoxidize MF to JH III , but in a far less efficient manner ( Table S2 ) . Given that B . mori JHAMT can methylate both FA and JHAs with similar efficiencies [21] , our data clearly demonstrate the major JH biosynthetic pathway in B . mori: homo-FAs are first epoxidized to JHAs by CYP15C1 , and then methylated to JHs by JHAMT ( Figure 4C and Figure 6A ) . Interestingly , D . punctata CYP15A1 does not convert FA to JHA III [18] . Thus , the difference in specificity of CYP15 to the substrates FA and MF may determine the order of the final steps of JH biosynthesis in insects . The expression of most early JH biosynthetic enzyme genes and JHAMT in B . mori is limited to the CA and shows dynamic developmental fluctuations [20] , [21] , [29] . In particular , the temporal expression profile of JHAMT correlates well with JH biosynthetic activity in B . mori [20] , [21] , [30] , [31] and in the Eri silkworm Samia cynthia ricini [32] , indicating that JHAMT is a key regulatory gene whose transcriptional control is critical for the regulation of JH biosynthesis in Lepidoptera . Here , we found that expression of CYP15C1 was also limited in CA but in a different pattern to other JH biosynthesis genes in that it was constitutively expressed from larval to adult stages . This result suggests that the transcriptional regulation of CYP15C1 is less important than JHAMT for the temporal regulation of JH production in B . mori . CA of the silkworm ceases JH biosynthesis by day 3 of the last ( fifth ) instar [20]; however , it is speculated that CA synthesizes and secretes JHAs during the following prepupal period . Our data indicate that this endocrine switch can be explained by constitutive CYP15C1 expression and the shut-off of JHAMT expression in CA ( Figure 6A ) . During the larval-pupal transition , homo-FAs are constantly converted to JHA I and II by CYP15C1 , and the resultant JHAs are secreted from the gland without further conversion because of the absence of JHAMT . CYP15 P450 family members are found in both hemimetabolous and holometabolous insects [33] . In a similar manner as CYP15C1 expression in B . mori , CA-specific CYP15 expression has also been observed in two cockroach species , D . punctata and Blattella germanica [18] , [34] , in the locust Schistocerca gregaria [35] , and in the mosquito Aedes aegypti [36] , suggesting a conserved function in JH biosynthesis . However , the enzymatic properties of CYP15 products , with the exception of those of D . punctata [18] and B . mori ( this study ) , have not been studied and the physiological role of CYP15s in the development of other insects remains unknown . By characterizing the CYP15C1-null mutant silkworm , we have demonstrated here that CYP15C1 plays an essential role in JH biosynthesis and for the maintenance of the proper timing of metamorphosis . Accumulating data have suggested that CYP15 genes are evolutionarily diversified in terms of their gene regulation and nature . For example , unlike B . mori CYP15C1 , A . aegypti CYP15 shows developmentally and dynamically regulated changes of expression , which appear to correlate well with the JH synthetic activity in the CA [36] . In addition , CYP15 is not present in the genome of D . melanogaster , but a P450 gene ( Cyp6g2 ) is expressed in CA in a highly tissue-specific manner [37] . More extensive research on the transcriptional controls and enzymatic properties of JH epoxidases across a broader range of insect taxa will shed light on the roles of these enzymes . Our results consistently indicate that the mod strain is a JH-deficient mutant that is unable to synthesize JHs in CA . One unique characteristic of the precocious pupation in the mod strain is the variation in the timing of the onset of spinning ( Figure 1 ) . The feeding period in early-maturing trimolters was unusually short ( 50 h after molting ) compared with that observed in surgical allatectomy of newly molted fourth instar larvae . In the latter larvae , the feeding period was comparable in length to that of the late-maturing trimolters [e . g . ∼130 h [17]] and no timing segregation was observed [17] . In addition , most of the early-maturing trimolters displayed a larval-pupal intermediate phenotype and eventually died , unlike allatectomized larvae , most of which successfully developed into small but normal pupae [17] . One explanation for this phenomenon is that the early-maturing trimolters were destined to undergo larval molting to the fifth instar on day 2 , while the late-maturing trimolters were destined for pupation after a prolonged fourth instar , similar to allatectomized larvae [17] ( Figure 1D ) . Molting in early-maturing trimolters on day 2 usually resulted in the formation of larval-pupal intermediates . One possible explanation for this mixed phenotype is that metamorphosis in the mod strain is induced in the presence of homo-MFs ( unepoxidized JH I and II ) , presumed products instead of epoxidized JH I and II in CA of the mod strain ( see Figure 6B ) . MF is known as a crustacean JH and has recently been reported to have JH activity in D . melanogaster [38] , [39] . Therefore , MF and its homologs might have JH-like activity but not able to fully substitute for authentic ( epoxidized ) JHs in the physiology of the silkworm . Alternatively , other P450 epoxidases in B . mori that have low substrate specificity and stereospecificity , like CYP9E1 [18] and CYP6A1 [24] in other insects , might substitute for the absence of CYP15C1 in peripheral tissues of mod larvae , and such locally-synthesized JHs may prevent precocious metamorphosis in the first and second instar larvae carrying the mod mutation . Further studies are needed to elucidate the mechanism for this unique characteristic of the mod strain . We found that the precocious phenotype was more severe in the w-1; mod strain compared to that in t011 , a genetic stock of the mod strain . We rarely observed dimolter larvae in the t011 stock ( Figure 1B ) . However , in the original manuscript in 1956 , it was reported that 28–92% of mod larvae became dimolters [11] . This difference might have developed as a consequence of unintended artificial selection during stock maintenance that favored broods producing trimolters in higher proportions , as it is difficult to obtain sufficient number of eggs using dimolter moths [11] , [12] . Thus , we speculate that the present t011 stock may be genetically fixed to produce mostly trimolters , and that this attribute can be varied by outcrossing to other strains . In the silkworm , premature metamorphosis can be induced by surgical removal of JH-producing CA ( allatectomy ) [17] , by application of an imidazole-based insect growth regulator KK-42 [40] or an anti-juvenile hormone agent KF-13S [41] , [42] , or by continuous overexpression of the JH-degrading enzyme , JH esterase [16] . In any case , however , premature pupation is not induced in larvae younger than the third instar . In agreement with these studies , we did not observe precocious pupation in first or second instar mod larvae , nor did we observe apparent developmental abnormalities during these early instars . Therefore , our data support the hypothesis that there are two physiological phases in the life of silkworm larvae [16]: the JH-independent phase ( first and second instar ) in which JH does not have a morphogenetic function; and , the JH-dependent phase ( third instar and thereafter ) in which the morphostatic action of JH is required to prolong the larval stage until the attainment of the appropriate body size for metamorphosis . Given that most generally the minimum number of the larval instar in insects is three [1] , [2] , our data further imply that insect larvae need to experience at least one [e . g . , L2 pupae in D . melanogaster [43]] or two ( e . g . , B . mori ) larval-larval molts and/or require a certain length of time of postembryonic development in order to acquire competence for metamorphosis . The silkworm is a classic model organism that has been used for pioneering studies in genetics , physiology , and biochemistry [5] . The availability of whole genome data [8] , post-genomic tools [10] , and unique mutant resources [6] , together with the classic “status quo” responses to JHs in this insect [14] , [15] , [17] , makes the silkworm well-suited for study of hormonal control of growth and development . Indeed , these advantages have greatly contributed to the identification of essential components in the biosynthesis of ecdysteroids , the insect molting hormones [44] . Moreover , recent success in targeted gene disruption using a zinc-finger nuclease [45] increases the utility of this model organism . We are hopeful that our present study will encourage further studies on other “moltinism” strains in the silkworm , and consequently pave the way for a greater understanding of physiological control , developmental plasticity , and evolutionary history of the number of larval molting in insects , which may reflect adaptive strategies of insects to diverse environmental conditions . It is also noteworthy that the late step of the JH biosynthetic pathway is insect-specific and is therefore a potential target for biorational insecticides [46] .
Silkworms were reared on an artificial diet or mulberry leaves at 25–27°C under standard conditions as described previously [47] . The silkworm strain t011 ( mod/mod ) was obtained from Kyushu University [6] . The Spodoptera frugiperda Sf9 and Drosophila melanogaster S2 cells were maintained as described previously [48] . To determine the developmental profile of mod , larvae from two batches of t011 were individually reared in plastic dishes , and their developmental stages were recorded at ∼8-h intervals . The JH analog , methoprene ( a kind gift from S . Sakurai ) was applied to newly molted third or fourth instar larvae ( ∼8–12 h after molting ) . Methoprene was diluted with acetone and the selected doses ( 0 . 01–10 µg/larva ) were topically applied to the dorsum using a 10-µl Hamilton microsyringe . The same volume of acetone was applied as a control . Positional cloning of the mod locus was performed as described previously [49] . Codominant PCR markers and p50T-specific PCR markers were generated for each position of the scaffold Bm_scaf16 ( chromosome 11 ) [9] , and used for genetic analysis ( Figure 2A and Figure S1 ) . Homozygotes of the mod locus were collected from the BC1 population [t011× ( p50×t011 ) ] based on the phenotype of precocious pupation . Total RNAs were collected from CA of day 0 fifth instar larvae of p50T and Kinshu×Showa strains and used for 5′- and 3′-rapid amplification of cDNA ends ( RACE ) using the GeneRacer Kit ( Invitrogen ) . PCR was performed using the primers listed in Table S1 . The PCR products were subcloned and sequenced as described previously [47] . The obtained cDNA sequence was deposited in the GeneBank ( accession number: AB124839 ) . qRT-PCR was performed essentially as described previously [21] . The primers used for the quantification of the CYP15C1 transcript are listed in Table S1 . In situ hybridization was performed essentially as described previously [50] . A CYP15C1 cDNA fragment ( ∼1 . 1 kb ) was amplified by PCR listed in Table S1 and subcloned into a pDrive plasmid vector ( QIAGEN ) . ( 2E , 6E ) -farnesoic acid ( FA ) and ( 2E , 6E ) -methyl farnesoate ( MF ) were purchased from Echelon Research Laboratories ( Salt Lake City ) and racemic JH III from Sigma . JH III acid was prepared from the racemic JH III as described previously [21] . ( R ) -JH III was a kind gift from W . G . Goodman . CYP15C1 overexpression in S2 cells was achieved using a GAL4/UAS system [51] . To generate a vector for expressing CYP15C1 under the control of the UAS promoter ( UAS-CYP15C1-HA ) , a cDNA fragment coding the entire CYP15C1 ORF was ligated into the pUAST vector . UAS-GFP . RN3 [52] was used as a negative control . UAS-CYP15C1-HA or UAS-GFP . RN3 was transfected with the Actin5C-GAL4 construct ( a gift from Yasushi Hiromi , National Institute of Genetics , Japan ) . Forty-eight hours after transfection of S2 cells in a 60-mm dish , the old medium was replaced with 2 ml of fresh medium . S2 cells were detached from the bottom of the dish by pipetting , and 1 ml of the cell suspension was transferred to a siliconized glass test tube . FA or MF ( 100 µM at final concentration ) was then added to the tube . After incubation at 25°C for 16 h , 500 µl of medium was collected and mixed with 500 µl of acetonitrile . Samples were centrifuged for 10 min at 15 , 000 rpm , followed by incubation at 25°C for 10 min . After filtration using a 0 . 2 µm filter , 10–20 µl of each sample was subjected to HPLC analysis as described below . A cDNA with the full ORF of CYP15C1 cDNA was subcloned into the pIZT/V5-His vector ( Invitrogen ) . The plasmid was transfected into Sf9 cells with Cellfectin reagent ( Invitrogen ) , then cells transiently expressing CYP15C1 were selected successively with Zeocin according to the manufacture's instruction and a cell line ( Sf9/CYP15C1 ) stably expressing CYP15C1 was established . Sf9/CYP15C1 cells were placed in a glass tube ( 12×75 mm ) coated with PEG20 , 000 and cultured in SF900-II SFM medium containing FA or MF ( 10 µg/ml ) for either 2 or 6 h at 26°C . In some experiments , the specific JH esterase-specific inhibitor OTFP ( 6 µM ) was added to the medium to prevent possible degradation of the generated JH III by intrinsic JH esterase present in the cells . After incubation , an equal volume of CH3CN was added to the medium , vortexed vigorously and centrifuged for 4 , 800 rpm for 10 min to remove cell debris . The supernatant was directly subjected to an HPLC analysis as described below for JH III acid or JH III , which were expected to be generated from FA and MF , respectively . JH III was analyzed by reversed-phase HPLC as described previously [21] . JH III acid was analyzed by reversed-phase HPLC ( column , Shiseido ODS UG80 , 150 mm×3 . 0 mm ID; solvent , CH3CN-20 mM CH3COONH4 , pH 5 . 5 , 35∶65 , flow rate , 0 . 5 ml/min; detection , UV 219 nm ) . ESI-MS spectrum of JH III acid was obtained by TSQ system ( Thermo Quest Finnigan , USA ) . The stereospecificity of the epoxide group of JH III acid formed by CYP15C1 was analyzed as follows under semi-dark conditions . Sf9/CYP15 cells were cultured in medium containing 10 µg/ml FA for 48 hrs . An equal volume of CH3CN was added to the medium ( 2 ml ) , vortexed vigorously and centrifuged at 4 , 800 rpm for 10 min . One ml of 1 M CH3COONH , ( pH 5 . 5 ) was added to the supernatant and extracted with 5 ml of CH2CH2; this step was performed 5 times . The extract was dehydrated with anhydrous Na2SO4 and evaporated to dryness in vacuo at 40°C , then the residue was dissolved in 200 µl of CH2Cl2 , 50 µl of MeOH and 100 µl of TMS-diazomethane were then added and the solution was incubated at room temperature for 30 min . The reaction was dried with an N2 gas stream , the residue dissolved in 100 µl of hexane , and subjected to a normal-phase HPLC ( column , Shiseido SG80 , 250×4 . 6 mm ID; solvent , hexane-EtOH , 99∶1; flow rate , 0 . 5 ml/min; detection , UV 211 nm ) . The peak corresponding to JH III ( r . t . = 9 . 8 min ) was collected . The stereospecificity of the epoxide group of the JH III was analyzed by a chiral-HPLC ( column , Chiralapack IA , 250×4 . 6 mm ID , DAICEL; solvent , hexane-EtOH , 99∶1; flow rate , 0 . 5 ml/min; detection UV 219 nm ) as described previously [31] . Ten microliters of deuterium-substituted JH III ( d3-JH III ) [53] in toluene ( 67 . 1 pg/ml ) was transferred to a clean glass tube to which 0 . 5 ml of methanol was added . The hemolymph sample ( 100 µl ) was then added and mixed vigorously , and 1 . 5 ml of 2% NaCl was added to the JH sample . JH was extracted by partition with 0 . 5 ml hexane; this step was performed three times . The combined solvent containing JH ( 1 . 5 ml ) was evaporated under a stream of nitrogen . One hundred microliters of methanol and 2 µl trifluoroacetic acid were added to the crude JH extract and mixture heated at 60°C for 30 min . After removal of the methanol , methoxyhydrin derivatives of JH ( JH-MHs ) were purified using a Pasture pipette packed with 1 . 0 g of aluminum oxide ( activity grade III , ICN Ecochrom ) prewashed with hexane . After loading the extract and washing with 2 ml of 30% ether in hexane , JH-MHs were eluted with 2 ml of 50% ethyl acetate in hexane and then dried under a stream of nitrogen . The residue was dissolved in 25 ml of 80% acetonitrile containing 5 µM sodium acetate . The HP1100MSD system ( Agilent ) was equipped with a 150×3 mm C18 reversed phase column ( UG80 , Shiseido ) protected by a guard column with 70% acetonitrile containing 5 µM sodium acetate at a flow rate of 0 . 4 ml/min . For MS analysis , electrospray ionization in the positive mode was used under the conditions of drying gas temperature at 320°C with 10 l/min flow rate , ionization voltage of 70 V . Under these conditions , selected ion masses for each JH-MH were monitored as [M+Na]+ , i . e . , m/z 321 , 324 , 335 , and 349 for JH III , d3-JH III , JH II , and JH I , respectively . Overexpression of CYP15C1 was performed in transgenic silkworms using the GAL4/UAS system as described previously [25] , [27] , [54] . A coding sequence of CYP15C1 was introduced into a silkworm UAS vector carrying the marker gene 3xP3-EGFP . B . mori transformants were established using standard protocols [10] . To overexpress CYP15C1 on the mod/mod background , established UAS lines and an enhancer trap line ET14 [27] were crossed with the t011 strain , and the resultant F1 animals were sib mated to obtain the F2 generation . In the F2 generation , we collected animals showing premature pupation with white eyes ( i . e . , mod/mod; w-1/w-1 ) and confirmed the presence of the fluorescent marker gene using a fluorescent microscope ( SZX12 , Olympus ) . The established w-1; mod lines carrying UAS-CYP15C1 or ET14 were crossed , and their offspring were examined to determine whether precocious metamorphosis was blocked by CYP15C1 overexpression . | The number of larval instars in insects varies greatly across insect taxa and can even vary at the intraspecific level . However , little is known about how the number of larval instars is fixed in each species or modified by the environment . The silkworm , Bombyx mori , provides a unique bioresource for investigating this question , as there are several “moltinism” strains that exhibit variations in the number of larval molts . The present study describes the first positional cloning of a moltinism gene . We performed genetic and biochemical analyses on the dimolting ( mod ) mutant , which shows precocious metamorphosis with fewer larval–larval molts . We found that mod is a juvenile hormone ( JH ) –deficient mutant that is unable to synthesize JH , a hormone that prevents precocious metamorphosis and allows the larvae to undergo multiple rounds of larval–larval molts . This JH–deficient mutation is the first described to date in any insect species and , therefore , the mod strain will serve as a useful model for elucidating the molecular mechanism of JH action . Remarkably , precocious larval–pupal transition in mod larvae does not occur in the first or second instar , suggesting that morphostatic action of JH is not necessary for young larvae of B . mori . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry",
"developmental",
"biology",
"genetics",
"biology",
"zoology",
"genetics",
"and",
"genomics"
] | 2012 | Precocious Metamorphosis in the Juvenile Hormone–Deficient Mutant of the Silkworm, Bombyx mori |
In mammalian auditory cortex , sound source position is represented by a population of broadly tuned neurons whose firing is modulated by sounds located at all positions surrounding the animal . Peaks of their tuning curves are concentrated at lateral position , while their slopes are steepest at the interaural midline , allowing for the maximum localization accuracy in that area . These experimental observations contradict initial assumptions that the auditory space is represented as a topographic cortical map . It has been suggested that a “panoramic” code has evolved to match specific demands of the sound localization task . This work provides evidence suggesting that properties of spatial auditory neurons identified experimentally follow from a general design principle- learning a sparse , efficient representation of natural stimuli . Natural binaural sounds were recorded and served as input to a hierarchical sparse-coding model . In the first layer , left and right ear sounds were separately encoded by a population of complex-valued basis functions which separated phase and amplitude . Both parameters are known to carry information relevant for spatial hearing . Monaural input converged in the second layer , which learned a joint representation of amplitude and interaural phase difference . Spatial selectivity of each second-layer unit was measured by exposing the model to natural sound sources recorded at different positions . Obtained tuning curves match well tuning characteristics of neurons in the mammalian auditory cortex . This study connects neuronal coding of the auditory space with natural stimulus statistics and generates new experimental predictions . Moreover , results presented here suggest that cortical regions with seemingly different functions may implement the same computational strategy-efficient coding .
The precise role played by the auditory cortex in hearing remains unclear . Before reaching cortical areas , raw sounds undergo numerous transformations in the brainstem and the thalamus . This subcortical processing seems more substantial than in other senses and is a specific property of the auditory system . What computations are performed by the cortex on the output generated by lower auditory regions is a question far from being answered . One of the issues in functional characterization of the auditory cortex is an apparent lack of specificity . Spiking activity of cortical auditory neurons is modulated by sound features such as pitch , timbre and spatial location [1 , 2] . Responses invariant to any of those features seem rare . This interdependence is especially puzzling in the context of extracting spatial information . A number of studies attempted to identify “what” and “where” streams in the auditory system ( e . g . [3 , 4] ) . Despite those efforts the existence of a sharp separation of spatial and identity information in the auditory cortex is still not definitely confirmed [5 , 6] . Neurons reveal sensitivity to sound position in most parts of the mammalian auditory cortex [7] . Their spatial tuning is quite broad — neural firing can be modulated by sounds located anywhere on the azimuthal plane . While activity of single units does not carry information sufficient to accurately localize sounds , larger numbers of neurons seem to form a population code for sound location [8–11] . These observations strongly differ from assumptions made early in the field that the auditory space is represented by a topographic cortical map , where neighboring units would encode the presence of a sound source at proximal positions [12] . Results described above constitute a conceptual challenge for theoretical models of the auditory cortex and understanding its role in spatial hearing in particular . Nevertheless , a number of candidate roles for this region have been proposed . It has been suggested , for instance , that the main function of the primary auditory cortex ( A1 ) is to process sound features extracted by subcortical structures [13] on multiple time scales . Another theory proposes that the auditory cortex in the cat represents abstract entities ( such as a bird song or wind ) rather than low-level spectrotemporal features of the stimulus [14] . It is also debated whether auditory cortical areas bear similarities to visual areas , and if yes , what sort of understanding can be gained by combining knowledge about those brain regions [15] . From a theoretical perspective one question seems to be particularly important — is there any general principle behind the functioning of auditory cortex , or does it carry out computations which are task- or modality-specific and therefore not performed in other parts of the brain ? A particularly succesful theoretical framework attempting to explain general mechanisms behind the functioning of the nervous system is provided by the Efficient Coding Hypothesis [16 , 17] . It proposes that sensory systems maximize the amount of transmitted stimulus information . Under the additional assumption that a typical stimulus activates only a small fraction of a neuronal population , the hypothesis is known as sparse coding[18 , 19] . Perhaps the strongest prediction of the efficient coding hypothesis is that the neuronal activity at consecutive stages of sensory processing should become less and less redundant , hence more independent . This prediction has been experimentally tested in the auditory system of the cat . Chechik and colleagues [20] recorded neuronal responses to natural sounds at three levels of the auditory hierarchy — Inferior Colliculus ( IC ) , Medial Genniculate Body ( MGB ) and A1 . They observed that spiking activity was progressively less redundant between single neurons , as quantified using information theoretic measures . This result suggests that audition can be understood using concepts provided by the efficient coding hypothesis . In order to form an efficient stimulus representation , neuronal codes should reflect regularities present in the sensory environment [21] . This implies that by studying statistics of natural input , one should be able to predict neuronal tuning properties . In audition , this idea has been followed by a number of researchers . Starting at the lowest level of the auditory system , Lewicki and Smith [22 , 23] demonstrated that learning a sparse representation of natural sound chunks reproduces shapes of cochlear filters of the cat . A recent extension of this work has suggested that while the auditory nerve may be optimally encoding all sounds it encounters , neurons in the cochlear nucleus may be tuned to efficiently represent particular sound classes [24] . Climbing higher in the auditory hierarchy — Carlson et al [25] have reproduced shapes of spectrotemporal receptive fields ( STRFs ) in the inferior colliculus by learning sparse codes of speech sounds . The relationship between spectrotemporal tuning of cortical neurons and sparse representation of speech spectrograms were explored by Klein , Koerding and Koenig [26 , 27] . More recently , some aspects of the topographic structure of the auditory cortex were shown to emerge from statistics of speech sounds by Terashima and Okada [28] . Terashima and colleagues have also studied the connection between sparse codes of natural sound spectra and harmonic relationships revealed by receptive fields of macaque A1 neurons [29] . Maximizing coding efficiency by learning sparse codes has also lead to emergence of signal representations useful in spatial hearing tasks . Asari et al [30] learned overcomplete dictionaries of monaural spectrograms and demonstrated that this representation allows for the segregation of acoustically overlapping and yet spatially separate sound sources ( the “cocktail party problem” ) . A recent study found that sparse coding of a spectrotemporal representation of binaural sound extracts spatial features invariant to sound identity [31] . As discussed above , a growing body of evidence seems to point to efficient coding as a computational process implemented by the auditory system . To date however , the connection between natural stimulus statistics and auditory spatial receptive fields remains unexplained . It is therefore unclear if spatial computations performed by the auditory cortex are unique to this brain area or whether they can be also predicted in a principled way from a broader theoretical perspective . The present work attempts to connect spatial computations carried out by the auditory cortex with statistics of the natural stimulus . Here , a hierarchical statistical model of stereo sounds recorded in a real auditory environment is proposed . Based on principles of sparse coding the model learns the spectrotemporal and interaural structure of the stimulus . In the next step , it is demonstrated that when probed with spatially localized sounds , higher level units reveal spatial tuning which strongly resembles spatial tuning of neurons in the mammalian auditory cortex . Additionally , the learned code forms an interdependent representation of spatial information and spectrotemporal quality of a sound . Activity of higher units is therefore modulated by sound’s position and identity , as observed in the auditory system . This study provides computational evidence that spatial tuning of auditory cortical neurons may be a manifestation of an underlying general design principle — efficient coding . Present results suggest that the role of the auditory cortex is to reduce redundancy of the stimulus representation preprocessed by the brainstem . Representation obtained in this way may facilitate tasks performed by higher brain areas , such as sound localization .
Binaural sound used to train the model was recorded by a human subject walking freely in a wooded area , in the presence of another speaker . The obtained recording included ambient ( wind , flowing stream ) and transient environmental sounds ( wood cracking , steps ) as well as human speech . The auditory scene changed over time due to the motion of the the recorder , the speaker , and changes in the environment . It consisted of multiple sound sources emanating from a diverse set of locations , creating together a complex , natural auditory environment . Please refer to the Methods section for details of the recording . The present study proposes a hierarchical statistical model of binaural sounds , which captures binaural and spectrotemporal structure present in natural stimuli . The architecture of the model is shown in Fig 1 . It consists of the input layer and two hidden layers . The input to the model was N sample-long epochs of binaural sound: from the left ear—xL and from the right ear—xR . The role of the first layer was to extract and separate phase and amplitude information from each ear by encoding them in an efficient manner . Monaural sounds were transformed into phase ( ϕL , ϕR ) and amplitude ( aL , aR ) vectors . This layer can be thought of as a statistical analogy to cochlear filtering . Phase vectors were further modified by computing interaural phase differences ( IPDs ) — a major sound localization cue [32] . This tranformation may be considered an attempt to mimic functioning of the medial superior olive ( MSO ) — the brainstem nucleus where phase differences are extracted [32] . The second layer of the model learned a joint sparse representation of monaural amplitudes ( aL , aR ) and phase differences ( Δϕ ) . Level ( amplitude ) and temporal ( phase ) information from each ear was jointly encoded by a population of M units . Each of the units was therefore capturing higher-order spectrotemporal patterns of sound in each ear . Additionally , by combining monaural information into single units higher level representation achieved spatial tuning not present in the first layer . The second hidden layer can be interpreted as a model of auditory neurons which receive converging monaural input and jointly operate on phase and amplitude — two kinds of information known to be important for spatial hearing . As demonstrated in previous work , filtering properties of the auditory nerve can be explained by sparse coding models of natural sounds [22] . There , short epochs of natural sounds are modelled as a linear combination of real-valued basis functions multiplied by sparse , independent coefficients ( i . e . having highly curtotic marginal distributions ) . Adapted to sets of natural sound chunks , basis functions become localized in time and/or frequency matching properties of cochlear filters . While being capable of capturing interesting properties of the data , real valued representations are not well suited for modeling binaural sounds . This is because binaural hearing mechanisms utilize interaural level and time differences ( ILDs and ITDs respectively ) . In narrowband channels , differences in time correspond to phase displacements known as interaural phase differences ( IPDs ) . Therefore a desired representation should both be adapted to the data ( i . e . non-redundant ) and separate amplitude from phase ( where phase is understood as a temporal shift smaller than the oscillatory cycle of a particular frequency ) . The present work addresses this twofold constraint with complex-valued sparse coding . Each data vector x ∈ ℝN is represented in the following way: x t = ∑ i = 1 N R { z i * A i , t } + η ( 1 ) where zi ∈ ℂ are complex coefficients , * denotes a complex conjugation , Ai ∈ ℂT are complex basis functions and η ∼ 𝓝 ( 0 , σ ) is additive gaussian noise . Complex coefficients in Euler’s form become z i = a i e j ϕ i ( where j = − 1 ) therefore Eq ( 1 ) can be rewritten to explicitely represent phase ϕ and amplitude a as separate variables: x t = ∑ i = 1 N a i ( cos ϕ i A i , t R + sin ϕ i A i , t I ) + η ( 2 ) Real and imaginary parts A i R and A i I of basis functions { Ai }i=1N span a subspace within which the position of a data sample is determined by amplitude ai and phase ϕi . Depending on number of basis functions N ( each of them is formed by a pair of vectors ) , the representation can be complete ( N/2 = T ) or overcomplete ( N/2 > T ) . In a probabilistic formulation , Eqs ( 1 ) and ( 2 ) can be understood as a likelihood model of the data , given coefficients z and basis functions A: p ( x | z , A ) = [ 1 σ 2 π ] T ∏ t = 1 T e - ( x t - x ^ t ) 2 2 σ 2 ( 3 ) where x ^ t = ∑ i = 1 N R { z i * A i , t } is the reconstruction of the t−th dimension of the data vector x . A prior over complex coefficients applied here assumes independence between subspaces and promotes sparse solutions i . e . solutions with most amplitudes close to 0: p ( z ) = 1 Z ∏ i = 1 N e - λ S ( a i ) ( 4 ) where Z is a normalizing constant . Function S ( ai ) promotes sparsity by penalizing large amplitude values . Here , a Cauchy prior on amplitudes is assumed i . e . S ( a i ) = l o g ( 1 + a i 2 ) . One should note however that amplitudes are always non-negative and that in general the Cauchy distribution is defined over the entire real domain . The model attempts to form a data representation keeping complex amplitudes maximally independent across subspaces , while still allowing dependence between coordinates z𝕽 , z𝕴 which determine position within each subspace . Inference of coefficients z which represent data vector x in the basis A is performed by minimizing the following energy function E 1 ( z , x , A ) ∝ 1 2 σ 2 ∑ t = 1 T ( x t ^ - x t ) 2 + λ ∑ i = 1 N S ( a i ) ( 5 ) which corresponds to the negative log-posterior p ( z∣x , A ) . This model was introduced in [33] and used to learn motion and form invariances from short chunks of natural movies . Assuming N = T/2 and σ = 0 , it is equivalent to 2-dimensional Independent Subspace Analysis ( ISA ) [34] . When trained on natural image patches , real and imaginary parts of basis functions A form pairs of Gabor-like filters , which have the same frequency , position , scale and orientation . The only differing factor is phase—real and imaginary vectors are typically in a quadrature-phase relationship ( shifted by π 2 ) . By extension , one might expect that the same model trained on natural sounds should form a set of frequency localized phase-invariant subspaces , where imaginary vector is equal to the real one shifted a quarter of a cycle in time . Somewhat surprisingly , such representation does not emerge , and learned subspaces capture different aspects of the data — bandwidth , frequency or time invariance [35 , 36] . In order to learn a representation from the statistics of the data that preserves a desired property such as phase invariance , one could select a parametric form of basis functions and adapt the parameter set [37] . Such a parametric approach has the disadvantage that the assumed family of solutions might not be flexible enough to efficiently span the data space . Another , more flexible alternative to learn a structured representation is to regularize basis functions by imposing temporal-coherence promoting priors [36] . This , however , requires determining the strength of regularizing priors . To overcome these problems , a different approach was taken here . The first-layer representation was created in two steps . Firstly a real-valued sparse code was trained ( see Methods ) . Learned basis functions were well localized in time or frequency and tiled the time-frequency plane in a uniform and non-overlapping manner ( Fig 2B ) . They were taken as real vectors Aℜ of complex basis functions A . In the second step , imaginary parts were created by performing the Hilbert transform of real vectors . The Hilbert transform of a time varying signal y ( t ) is defined as follows: H ( y ( t ) ) = 1 π p . v . ∫ - ∞ ∞ y ( τ ) t - τ d τ ( 6 ) Where p . v . stands for Cauchy principal value . In such a way every real vector A i ℜ was paired with its Hilbert transform A i ℑ = H ( A i ℜ ) i . e . a vector which complex Fourier’s coefficients are all shifted by π 4 in phase . The obtained dictionary is adapted to the stimulus ensemble , hence providing a non-redundant data representation , yet makes phase clearly interpretable as a temporal displacement . The model was trained using T = 128 sample-long chunks of sound sampled at 8 kHz , which corresponds to 16 ms duration . The complete representation of 128 real basis functions was trained , and each of them was paired with its Hilbert transform , resulting in the total number of 256 basis vectors . Selected basis functions are displayed in Fig 2A . Real vectors are plotted in black together with associated imaginary ones plotted in gray . Panel B of the same figure displays isoprobability contours of Wigner-Ville distributions associated with the 256 basis functions . This form of representation localizes each temporal feature on a time-frequency plane [38] ( one should note that real and imaginary vectors within each pair are represented by the same contour on that plot ) . A clear separation into two classes is visible . Low frequency basis functions ( below 1 kHz ) are non-localized in time ( spanning the entire 16 ms interval ) , while in higher frequency regions their temporal precision increases . An interesting bandwidth reversal is visible around 3 kHz , where temporal accuracy is traded for frequency precision . Interestingly , the sharp separation into frequency and time localized basis functions , which emerged in this study was not clearly visible in other studies which performed sparse coding of sound [22 , 38] . Time-frequency properties observed here reflect the statistical structure of the recorded auditory scene , which mostly consisted of non-harmonic environmental sounds sparsely interspersed with human speech . Fig 3 depicts a typical distribution of binaural phase . Phases of the same basis function in each ear reveal dependence in their difference . This means that joint probability of monaural phases depends solely on the IPD: p ( ϕ i , L , ϕ i , R ) ∝ p ( Δ ϕ i ) ( 7 ) where Δϕi = ϕi , L−ϕi , R is the IPD . This property is a straightforward consequence of physics of sound — sounds arrive to each ear with a varying delay giving rise to positive and negative phase shifts . From a statistical point of view this means that monaural phases become conditionally independent given their difference and a phase offset ϕi , O: ϕ i , L ⊥ ϕ i , R | Δ ϕ i , ϕ i , O ( 8 ) The phase offset ϕi , O is the absolute phase value — indicating the time from the beginning of the oscillatory cycle . It can be therefore said that: ϕ i , L = ϕ i , O + Δ ϕ i 2 ( 9 ) ϕ i , R = ϕ i , O - Δ ϕ i 2 ( 10 ) This particular statistical property allows us to understand IPDs not as an ad-hoc computed feature but as an inherent property of the probability distribution underlying the data . It is reflected in the structure of the graphical model ( see Fig 1 ) . Since the phase offset ϕi , O does not carry spatial information for the purposes of current study it is treated as an auxiliary variable and therefore marked in gray . In an approach to model the cochlear coding of sound , monaural sound epochs xL and xR were encoded independently using the same dictionary of complex basis functions A described in the previous section . Signal from both ears converged in the second hidden layer , which role was to form a joint , higher-order representation of the entire stimulus processed by the auditory system . The celebrated Duplex Theory of spatial hearing specifies two kinds of cues used to solve the sound-localization task: interaural level and time ( or phase ) differences [39] . While IPDs are supposed to be mostly used in localizing low-frequency sounds , ILDs are a cue , which ( at least in the laboratory conditions ) can be used to identify the position of high frequency sources . Phase and level cues are known to be computed in lateral and medial superior olive ( LSO and MSO respectively ) — separated anatomical regions in the brainstem [32] . However , an assumption made here was that neurons in the auditory cortex receive converging input from subcortical structures . This would enable them to form their spatial sensitivity using both fine structure phase and amplitude information . One can take also the inverse perspective: a single object ( a “cause” ) in the environment generates level and phase cues at the same time . Its identification therefore has to rely on observing dependencies between those features of the stimulus . The second layer formed a joint representation of monaural amplitudes and interaural phase differences . However , not all IPDs were modelled in that stage . Humans stop utilizing fine structure IPDs in higher frequency regimes ( roughly above 1 . 3 kHz ) , since this cue becomes ambiguous [32] . Aditionally , cues above around 700 Hz become ambiguous ( a single cue value does not correspond to a unique source position ) . For those reasons , and in order to reduce the number of data dimensions , 20 out of 128 IPD values were selected . The selection criteria were the following: ( i ) an associated basis function should have the peak of the Fourier spectrum below 0 . 75 kHz ( which provided the upper frequency bound ) , and ( ii ) it should have at least one full cycle ( which provided the lower bound ) . All basis functions fulfilling these criteria were non-localized in time ( they spanned entire 16 ms interval ) . As a result , the second layer of the model was jointly encoding T = 128 log-amplitude values from each ear and P = 20 phase differences . Monaural log-amplitude vectors aL , aR ∈ ℝT were concatenated into a single vector a ∈ ℝ2×T , and encoded using a dictionary of amplitude basis functions B . Representation of IPDs ( Δϕ ) was formed using a separate feature dictionary ξ . Both — phase and amplitude basis functions ( B and ξ ) , were coupled by associated sparse coefficients s . The overall generative model of phases and amplitudes was defined in the following way: a n = ∑ i = 1 M s i B i , n + η ( 11 ) Δ ϕ n = | w | ∑ i = 1 M s i ξ i , n + ϵ ( 12 ) The amplitude noise was assumed to be gaussian ( η ∼ 𝓝 ( 0 , σ2 ) ) with σ2 variance . Since phase is a circular variable its noise ε was modelled by the von Mises distribution with concentration parameter κ . The second layer was encoding two different physical quantities — phases , which are circular values , and log-amplitudes , which are real numbers . The goal was to form a joint representation of both parameters and learn their dependencies from the data . A simple , linear sparse coding model could be in principle used to achieve this task . However , if a single set of sparse coefficients si was used to model both quantities , scaling problems could arise , namely a coefficient value which explains well the amplitude vector may be too large or too small to explain the concomittant IPD vector . For this reason an additional phase multiplier w was introduced . It enters Eq 11 as a scaling factor , which gives the model additional flexibility required to learn joint probability distribution of amplitudes and IPDs . Fig 1 depicts it in gray as an auxiliary variable . In this way , amplitude values and phase differences were modelled by variables sharing a common , sparse support ( coefficients s ) , with a sufficient flexibility . Pairs of basis functions Bi , ξi represent binaural spectrotemporal stimulus and IPD patterns respectively , while sparse coefficients s signal their joint presence in the encoded sound epoch . An i−th second-layer unit was activated ( si ≠ 0 ) whenever a pattern of IPDs represented by the basis function ξi or a pattern of amplitudes represented by Bi was present in its receptive field . The activity was maximized , when both features were present at the same time . For this reason , when seeking analogies between the higher-level representation and auditory neurons , coefficients s can be interpreted as neuronal activity ( e . g . firing rate ) and basis function pairs Bi , ξi as receptive fields ( i . e . stimulus preferred by a neuron ) . The likelihood of amplitudes and phase differences defined by the second layer was given by: p ( a , Δ ϕ | s , w , B , ξ ) = [ 1 σ 2 2 π ] 2 T ∏ n = 1 2 T e - ( a n - a ^ n ) 2 2 σ 2 2[ 1 2 π I 0 ( κ ) ] P ∏ m = 1 P e κ cos ( Δ ϕ m - Δ ϕ ^ m ) ( 13 ) where a ^ n = ∑ i = 1 M s i B i , n , Δ ϕ ̂ m = ∣ w ∣ ∑ i = 1 M s i ξ i , m are amplitude and phase reconstructions repsectively and I0 is the modified Bessel function of order 0 . The joint distribution of coefficients s was assumed to be equal to the product of marginals: p ( s ) = 1 Z ∏ i = 1 M e - λ 2 S ( s i ) ( 14 ) where λ2 is a sparsity controlling parameter . A Cauchy distribution was assumed as a prior over marginal coefficients ( i . e . S ( s i ) = log ( 1 + s i 2 ) ) . To prevent degenerate solutions , where sparse coefficients s are very small and the scaling coefficient w grows undbounded , a prior p ( w ) constraining it from above and from below was placed . A generalized Gaussian distribution of the following form was used: p ( w ) = β 2 α Γ ( 1 β ) e - ( | w - μ | α ) β ( 15 ) Γ denotes tha gamma function , α , β and μ denote the scale , shape and location parameters respectively . When the shape parameter β is set to a large value ( here β = 8 ) , the distribution approximates a uniform distribution . Varying the scale parameter α changes the upper and the lower limit of the interval . Taken together the negative log-posterior over the second layer coefficients was defined by the energy function: E 2 ( s , w , B , ξ ) ∝ 1 σ 2 2 ∑ n = 1 2 × T ( a n - a ^ n ) 2 + κ ∑ m = 1 P cos ( Δ ϕ m - Δ ϕ ^ m ) + λ 2 ∑ i = 1 M S ( s i ) + λ w ( | w - μ | α ) β ( 16 ) the λw coefficient was introduced to control the strength of the prior on the scaling coefficient w . Similarly as in the first model layer , learning of basis functions and inference of coefficients was performed using gradient descent ( see Methods ) . The total number M of basis function pairs was set to 256 .
The second layer of the model learned a distributed representation of sound features accesible to neurons in the auditory cortex . Assuming that the cortical auditory code indeed develops driven by principles of efficiency and sparsity , one can interpret second layer basis functions as neuronal receptive fields and sparse coefficients s as a measure of neuronal activity ( e . g . firing rates ) . The model can be then probed using spatial auditory stimuli . If it indeed provides an approximation to real neuronal computations , its responses should be comparable with spatial tuning properties of the auditory cortex . In order to verify whether this was true , a test recording was performed . As a test sound the hiss of two pieces of paper rubbed against each other was used . It was a broadband signal , reminiscent of white noise used in physiological experiments , yet posessing natural structure . Recording was performed in an anechoic chamber , where a person walked around the recording subject while rubbing two pieces of paper ( see Methods for a detailed description ) . The recording was divided into 18 windows , each corresponding to a 20 degree part of a full circle . The number of windows was selected to match experimental parameters in [8 , 10] . From each window 3000 epochs were drawn and each of them was encoded using the model . Computing histograms of coefficients s at each angular position θ , provided an estimate of conditional distributions p ( si∣θ ) . Panel A in Fig 7 displays a conditional histogram of coefficient s corresponding to the basis function pair depicted in Fig 4A . Distributions of sparse coefficients revealed a strong dependence in the position of the sound source . As visible in the figure , the conditional mean of the distribution p ( si∣θ ) traced by the red line varied in a pronounced way across all positions . By analogy to averaged firing rates of neurons , average unit responses at each position were further studied to understand the spatial sensitivity of basis functions . Mean vectors μi , θ were constructed for each second-layer unit by taking its average response at the sound source position θ . Each mean vector was shifted and scaled such that its minimum value was equal to 0 and the maximum to 1 . Such transformation was analogical to physiological studies [8] and allowed for comparison with experimetally measured spatial tuning curves of auditory neurons , and for this reason scaled vectors μi will be referred to as model tuning curves in the remainder of the paper . In order to identify spatial tuning preferences , the population of model tuning curves was grouped into two clusters using the k-means algorithm . Obtained clusters consisted of 118 and 138 similar vectors . Tuning curves belonging to both clusters and revealing a strong correlation ( ∣ρ∣ > 0 . 75 ) with sound position are plotted in Fig 7C as gray lines . Cluster centroids ( averages of all tuning curves belonging to a cluster ) are plotted in black . Second layer units were tuned broadly—most of them were modulated by sound located at all positions surrounding the subject’s head . A clear spatial preference is visible—members of cluster 1 were most highly activated ( on average ) by sounds localized close to the left ear ( θ ≈ −90° ) , while cluster 2 consisted of units tuned to the right ear ( θ ≈ 90° ) . Very similar tuning properties of auditory neurons were identified in the cat’s auditory cortex [8] . Data from this study is plotted for comparison in the subfigure B of Fig 7 . Neuronal recordings were performed in the right hemisphere and two panels depict two subpopulations of neurons . The larger contra- and the smaller ipsi-lateral one . It is important to note , that the notion of ipsi , and contra laterality is not meaningful in the proposed model , therefore one should compare shapes of model and experimental tuning curves , not the numerosity of units in each population or cluster . Two major features of cortical auditory neurons responsive to sound position were observed experimentally: ( i ) tuning curve peaks were localized mostly at extremely lateral positions ( opposite to each ear ) and ( ii ) slopes of tuning curves were steepest close to the auditory midline . Both properties are visible in model tuning curves in Fig 7 . However , in order to perform a more direct comparison between the model and experimental data , analysis analogous to the one described in [8] was performed . First , tuning curve centroids were computed . A centroid was defined as an average position , where the unit activation was equal to 0 . 75 or larger ( see Methods ) . In the following step , the position of maximal slope towards midline was identified for each unit . This meant that for units tuned to the left hemifield ( cluster 1 ) the position of the minimal slope value was taken , while the position of the maximal one was taken for units tuned to the right hemifield ( cluster 2 ) . In this way , the position of maximal sensitivity to changes in sound location was identified . Distributions of model centroids and maximal slope positions are depicted in Fig 8B . Centroids were distributed close to lateral positions , opposite in each cluster ( −90° cluster 1 , +90° cluster 2 ) . Distribution peaks were located at positions close to each ear . No uniform tiling of the space by centroid values was present . At the same time , maximal slope values were tightly packed around the midline—peaks of their distributions were located precisely at , or very close to 0 degrees . This means that while the maximal response was on average triggered by lateral stimuli , the largest changes were triggered by sounds located close to the midline . Both properties were in good agreement with the experimental data reported in [8] . Fig 8A depicts in three panels centroid and slopes distributions measured in three different regions of cat’s auditory cortex—Primary Auditory Field ( A1 ) , Posterior Auditory Field ( PAF ) and Dorsal Zone ( DZ ) . A close resemblance between the model and physiological data was visible . It has been argued that while single neurons in the auditory cortex provide coarse spatial information , their populations form a distributed code for sound localization [8 , 9 , 9 , 10] . Here , a decoding analysis was performed to verify whether similar statement can be made about the proposed model . A gaussian mixture model ( GMM ) was utilized as a decoder . The GMM modelled the marginal distribution of sparse coefficients as a linear combination of 18 gaussian components , each corresponding to a particular position of a sound source ( i . e . the θ value ) . In the first part of the decoding analysis , single coefficients were used to identify the sound position . The GMM was fitted using the training dataset consisting of coefficient values si and associated position labels θ . In the testing stage , position estimates θ ^ were estimated ( decoded ) using unlabeled coefficients from the test dataset ( see Methods section for a detailed description of the decoding procedure ) . For each of the coefficients , a confusion matrix was computed . A confusion matrix is a two-dimensional histogram of θ and θ ^ and can be understood as an estimate of the joint probability distribution of these two variables . Using a confusion matrix , an estimate of mutual information ( i . e . , the number of bits shared between the position estimate θ ^ and its actual value θ ) was obtained . Fig 9A depicts histograms of information carried by each coefficient si about the sound source position , estimated as described above . A general observation is that single coefficients carried very little information about the sound location . The histogram peaks at a value close to 0 . 1 bits . Only few units coded approximately 1 bit of positional information . Even 1 bit , however , suffices merely to identify a hemifield , not to mention the precise sound position . As can be predicted from the broad shapes of the tuning curves , single second-layer units carried little spatial information . A similar result was obtained for neurons in different areas of the cats auditory cortex [12] . The amount of information about the sound position encoded by spike count of neurons in A1 and PAF regions has a distribution closely similar to that of model units ( compare with the left panel of figure 11 in [46] ) . Spike count ( which essentially corresponds to a firing rate ) is a feature of a neuronal response most directly corresponding to coefficients s in the model described here . The median of mutual information estimated from model coefficients ( marked by a diamond symbol in panel A ) aligns well with the same quantity estimated from neuronal data , and is close to 0 . 2 bits [46] . Overall , physiological measurements and the behavior of the model were highly similar . While single neurons did not carry much spatial information , the joint population activity was sufficient to decode the sound position [8–10 , 46] . Therefore in the second step of the decoding analysis , multiple coefficients s were used to train and test the GMM decoder . Results of the population decoding are plotted in Fig 9B . The decoder was trained with a progressively larger number of second-layer units ( from 1 to 256 ) and the mutual information was estimated from obtained confusion matrices . Each line in the plot depicts the number of bits as a function of the number of units used to perform decoding . Line colors correspond to the number of samples over which the average activity was computed . Broadly speaking , larger populations of second-layer units allowed for a more precise position decoding . As in the case of single units , averages over larger amounts of samples were also more informative—population activity averaged over 32 samples saturated amount of bits required to perform errorless decoding ( 4 . 17 ) . Two confusion matrices obtained from raw population activity and an average over 16 samples are displayed in subfigures Fig 9C and 9D . In the former case , the decoder was mostly misclassifying sound positions within each hemifield . Averaging over 16 sound samples yielded an almost diagonal ( errorless ) confusion matrix . The decoding analysis allowed us to draw the conclusion that while single units carried very little spatial information , their population encoded source location accurately , consistent with experimental data . Second layer units achieved spatial tuning by assigning different weights to amplitudes in each ear , and to IPD values in different frequency channels . At the same time they encoded spectrotemporal features of sound , as depicted in Fig 4 . Their activity should therefore be modulated by both sound position as well as its quality . Such comodulation is a prominent feature of the majority of cortical auditory neurons [1 , 7] . In order to verify this , model spatial tuning curves were estimated with a second sound source , very different from a hiss created by rubbing paper—human speech ( see Materials and Methods for details ) . Frequency spectra of both test stimuli are depicted in Fig 10D . Test sounds distributed their energy over non-overlapping parts of the frequency spectrum . While speech consisted mostly of harmonic peaks below 1 . 5 kHz , the paper sound was much more broadband and its energy was uniformely distributed between 1 . 5 and 4 kHz . Panels A-C of Fig 10 depict three amplitude/IPD basis function pairs together with their spatial tuning curves estimated using different sounds . The spatial preference of depicted units ( left or right hemifield ) was predictable from their binaural composition . Each of them , however , was activated stronger by a stimulus , which spectrum matched better amplitude basis functions . Basis functions visible in panels A and C had a lot of energy accumulated in higher frequencies , therefore the paper sound activated them stronger ( on average ) . Basis function B ) was spectrally better corresponding to speech sounds , therefore speech was a preferred class of stimuli . This observation suggests that tuning curves i . e . position-conditional means μi , θ should be understood not as averages of coefficient ensembles conditioned only on the sound position θ but also on spectral properties of sound . When interpreting coefficients s as neuronal activity this means that spatial tuning curves would alter their shapes when the neuron is tested with two different sound sources . Taken together , one can state that the second-layer representation encoded position and identity of the stimulus in an interdependent fashion .
In mammals , the location of a sound is encoded by two populations of broadly tuned , spatially non-specific units [32] . This finding challenges initial expectations of finding a “labelled-line code” ( i . e . a topographic map of neurons narrowly tuned to small areas of space ) . The “spatiotopic map” was expected by analogy to the tonotopic structure of the cortex , as well as the high localisation accuracy of humans and animals . Instead , it has been found that auditory cortical neurons within each hemisphere are predominantly tuned to far , contralateral positions . Peaks of observed tuning curves did not tile the auditory space uniformly , rather they were clustered around the two lateral positions . A prominent observed feature of cortical representation of sound location were slopes of the tuning curves . Regardless of the position of the tuning curve peak , slopes were steepest close to the interaural midline—the area where behavioral localisation acuity is highest [32] . From described observations , two prominent conclusions were drawn . Firstly , that the slope of tuning curves , not the distribution of their peaks determines spatial acuity [8 , 32 , 47 , 48] . Secondly that sound position is encoded by distributed patterns of population activity , not single neurons [8–10] . It has been argued that these properties are a manifestation of a coding mechanism which evolved to specifically meet the demand of binaural hearing tasks [8 , 32] . Here it is shown that crucial properties of cortical spatial tuning emerge in an unsupervised learning model , which learns a sparse representation of natural binaural sounds . The objective of the model was to code the stimulus efficiently ( i . e . with a minimal redundancy within limits of applied transformations ) , while minimizing unit activity . Properties of the learned representation are therefore a reflection of stimulus statistics , not of any task-specific coding strategy ( required for instance to localize sounds with the highest accuracy at the midline ) . The position of the sound-generating object is a latent variable for the auditory system . It means that its value is not explicitly present in the raw stimulus—it has to be estimated . This estimation , ( or inference ) is a non-trivial task in the real acoustic environment , where sounds reaching ear membranes are a reflection of intricate auditory scenes . Sensory neurons perform transformations of those sound waveforms to reconstruct the spatial configuration of the scene . Therefore , in an attempt to understand cortical representation of space , it may be helpful to think what is the statistical structure of the naturally encountered binaural stimulus that the auditory system operates on . Sounds reaching the ear contain information about their generating sources , the spatial configuration of the scene , position and motion of the organism and the geometry of its head and outer ears . Results obtained here suggest that the shapes of the model spatial tuning curves reflect regularities imposed on the sensory data by the filtering properties of the head . At lateral positions ( directly next to the left or the right ear ) there is no acoustic attenuation by the skull , hence sounds are loudest and least delayed . This in turn , elicits the strongest response in units preferring that side . When the sound is at a contralateral position , response is much weaker , due to the maximal head attenuation and largest delay . The curve connecting those two extrema is steepest in the transition area—at the midline . Since the auditory environment was uniformly sampled at both sides of the head , model units were clustered into two roughly equal subpopulations , basing on the shapes of their tuning curves . Clusters were symmetric with respect to each other—one tuned to to the left and the other to the right hemifield . This groupping is reminiscent of the “opponent-channel” representation of the auditory space , which has been postulated before [8 , 32] . Present results provide a theoretical interpretation of this tuning pattern . They suggest that neuronal population which forms a sparse , efficient representation of natural stimuli would reveal two broadly tunned channels , when probed with sounds located at different positions . It has been shown previously that IPD coding strategies in different species can be predicted from statistics of binaural sound [45] . Harper and McAlpine demonstrated that if the goal of the nervous system is to represent IPD values with the maximal possible accuracy ( quantified by Fisher information ) two populations of neurons tuned to opposite locations constitute an optimal representation of low-frequency IPDs . Their approach differs significantly from the one presented here . On the most abstract level , the authors of [45] assume that the purpose of IPD sensitive neurons is to maximize Fisher information , while here mutual information is the quantity implicitly maximized by the representation ( although interesting relationships exist between those two measures [49] ) . Secondly , Harper and McAlpine limit their analysis to IPD statistics only—here entire binaural waveforms are modelled . Finally the current study does not assume any parameteric shape of tuning curves , nor make any other assumptions about physiology as is the case in [45] . The similarity of model responses and neuronal activity emerges from data statistics . There is an ongoing debate about the presence ( or lack of thereof ) of two-separate “what” and “where” streams in the auditory cortex [5] . The streams would separate spatial information from other sound features which determine its identity . An important prediction formed by this dual-stream hypothesis is that there should exist neurons selective to sound position and invariant to other aspects in the auditory cortex . While some evidence has been found supporting this notion [3 , 4] it seems that at least in vast parts of the auditory cortex neural activity can be modulated by multiple features of sound such as pitch , timbre and location [1] . Neurons are sensitive to sound position ( i . e . changing position affects their firing patterns ) , but not selective nor invariant to it . The majority of studies analyzing spatial sensitivity in the auditory cortex use a single class of sound and the source position is the only varying parameter . Therefore , despite initial efforts , the influence jointly exerted by sound quality and position on neuronal activity is not yet well understood . The statistical model proposed here suggests that no dissociation of spatial and non-spatial information is necessary to either reconstruct the sound source or identify its position . The learned second-layer representation carries both kinds of information—about the sound quality ( contained in the spectrotemporal structure of basis functions ) and about spatial aspects ( contained in the binaural amplitude weighting and IPD vectors ) . The learned code forms a “what is where” representation of the stimulus ( i . e . , those two aspects are represented interdependently ) . A manifestation of this fact is visible in different scaling of spatial tuning curves , when probed with two different sound sources . Such comodulation of neuronal activity by sound position and quality has been observed experimentally [1] , which may suggest that recorded neurons form a sparse , efficient representation of binaural sound . An advantage of an interdependent “what is where” representation is the absence of the “feature binding problem” , which has to be solved if spatial information is processed independently . After separating the location of a source from its identity they would have to be fused at processing stages beyond the auditory cortex . A code similar to the one described here does not create such a problem . This idea goes in hand with results of a recent perceptual study [50] . Parise et al . demonstrated that the perception of sound source elevation is strongly influenced by its frequency . Furthermore they show that this relationship can be explained by adaptation to the joint distribution of natural sounds’ positions and spectra . This implies that the quality of the sound source as well as its spatial position are mutually dependent , and as such should be represented jointly , if the goal of the nervous system is to increase coding efficiency . The model proposed in this work is a statistical one—it constitutes an attempt to describe functional , not anatomical modules of the auditory system . Rather than explicitly modelling stages of the auditory pathway , its goal is to approximate the distribution of natural binaural sounds . The behaviour of units in the highest layer reveals a strong resemblance to cortical auditory neurons in an abstract , information processing domain . In the mammalian auditory system the sound is processed in at least five anatomical structures before it reaches the cortex [51] . It is therefore almost certain that the stimulus is subjected to many more complex transformations than the ones proposed here . On the other hand , the fact that similiarities between cortical and model responses emerge despite this lack of detail , imply that the model may be capturing some aspects of information processing , as it happens in the real auditory system . The relationship between abstract computational principles such as sparse coding and neurophysiology is an area of ongoing research [52–54] . An interesting extension of the present work would attempt to increase the level of biological detail , and see whether this allows formation of more refined experimental predictions . This could be done by implementing sparse coding computations using spiking neuron models , as it has been done in studies of the visual system ( e . g . [52 , 54] ) . The match between the model and biology could be also improved by including phenomena specific to the auditory system , such as the phase locking in the auditory nerve . This study focuses predominantly on explaining the broad spatial tuning of cortical auditory neurons estimated by the analysis of firing rates . With progressively larger amounts of biological detail added to the model , one could attempt to explain other aspects of spatial information encoding . For instance , the notion of spike timing does not exist in the approach proposed here , while temporal spike patterns of cortical neurons seem to carry relevant spatial information [9 , 10 , 46] . Moreover , as mentioned in the results section , the concept of contra- and ipsilaterality is spurious for high-layer model units since they are not associated with any anatomical locus ( left or right hemisphere ) . Overrepresentation of the contralateral ear is an interesting feature of panoramic population codes [8] , which is also not addressed by the present work . Further exploration of the relationship between specific biological observations and spatial information processing constitutes a possible goal for future research . It is highly likely that the main result of this study ( i . e . , spatial tuning properties of the binaural sound representation ) could be reproduced by replacing the first layer with a different sort of spectrotemporal signal representation . It would not necessarily have to be the sparse , efficient encoding of sound epochs . A spectrogram could be a candidate signal , although it has been demonstrated that a sparse code of relatively long binaural spectrogram chunks generates features of very different spatial tuning [31] . In this work , for the sake of theoretical consistency , both layers were learned using the same principles and statistical assumptions—sparse factorial coding . The data used for comparisons originated from studies of cat auditory cortex ( [8 , 46] ) . Since statistics of the binaural signal are affected by the geometry of ears and the head of the organisms , one could argue that model trained on binaural recordings performed by a human should not be compared with cat physiology . As long as detailed features of neuronal tuning to a sound position may vary across those species , tuning patterns highly similar to those of the cat have been observed in the auditory cortex of primates [55 , 56] . Overall , the cortical representation of sound position seems to be highly similar across mammals [32] . Finally , in the current study a binaural recording of only a single auditory scene was used to train the model . Even though the recording included many types of sound—ambient environmental noises , transient cracks and clicks and harmonic structures such as the human speech , it did not include many other possible sources ( for instance animal vocalizations ) . The recording included also only a narrow range of other parameters which characterize natural auditory scenes , such as reverberation . Analysis of longer recordings performed in different environmental settings may generate more diverse results and additional insights . One should note however , that certain properties of the learned representation ( such as the tradeoff in the spectrotemporal modulation ) seem to be a general proprerty of natural sounds as such and remain invariant to a specific dataset [25 , 40] . Basing on this observation one may expect that units revealing similar spatial tuning can be learned from recordings of numerous , diverse sets of natural sounds . Taken together , this paper proposes a candidate theoretical mechanism explaining how neurons in the auditory cortex represent spatial information . This model allows us to speculate they do not have to implement any task-dependent strategy . Instead , their behavior can be explained by sparse coding—a statistical model which has succesfully predicted properties of multiple other sensory systems [18 , 21] . Taking a broad perspective , ( as suggested by Barlow in his later work [57 , 58] ) this means that redundancy reduction by sparse coding can be used by the brain to identify sensory data patterns allowing sucesful interaction with the environment .
Sound recordings received approval of the Ethics Council of the Max-Planck Society . Human participants provided a written consent to participate in recordings . Sounds used to train and test the model were recorded using Soundman OKM-II binaural microphones placed in the ear channels of a human subject , whose head circumference was 60 cm . While recording training sounds , the subject walked freely in a wooded area accompanied by another person who spoke rarely . In this way , collected data included transient and ambient environmental sounds as well as harmonic speech . The binaural composition of sound was affected by spatial configuration of the environment and motion patterns of the recording subject . The recording used to train the model was 60 seconds long in total . Binaural recordings are availible in the supplementary material of [59] . Test recordings used to map the spatial tuning of second-layer units was performed in an anechoic chamber at the Department of Biology , University of Leipzig . The same recording subject was seated in the middle of the chamber . A female speaker walked at a constant pace following a circular path surrounding the recording subject . While walking she counted out loud . This was repeated four times . The second test recording was performed in a similar fashion , however instead of speaking the walking person rubbed two pieces of cardboard against each other , generating a broadband sound . To estimate the conditional distribution of sparse coefficients given the position and identity of the sound , test recordings were divided into 18 intervals , each corresponding to the same position on a circle . All recordings were registered in an uncompressed wave format at 44100 Hz sampling rate . Prior to training the model , sounds were downsampled to 8000 Hz . Test recordings are availible in the supplementary material ( S1 , S2 , S3 , S4 , S5 , S6 , S7 , S8 Files ) . The goal of the learning procedure was to estimate first- ( A ) , and second- layer basis functions ( B , ξ ) . This was done using a two-step approach . Firstly maximum a posteriori ( MAP ) estimates of model coefficients ( z in the first layer , s and w in the second ) were inferred via gradient descent [18 , 33] . Secondly , a gradient update on basis functions was perormed using current coefficient estimates . Those two steps were consecutively iterated until the model converged . A dictionary of complex-basis functions in the first layer was created by first , training a standard sparse code of sound epochs x ∈ ℝT: x t = ∑ i = 1 T c i Θ i , t + η ( 18 ) The negative log-posterior of this model was: E s ( x , c , Θ ) ∝ 1 σ 2 ∑ t = 1 T ( x t - x ^ t s ) 2 + λ ∑ i = 1 T S ( c i ) ( 19 ) where x ^ t s = ∑ i = 1 T c i Θ i , t is the reconstruction of the data vector . Corresponding gradients over linear coefficients c and basis functions Θ were given by: ∂ ∂ c i E s ∝ - 2 σ 2 ∑ j = 1 T Θ j , t ( x t - x ^ t s ) + 2 λ c i log ( 1 + c i 2 ) ( 20 ) ∂ ∂ Θ i , t E s ∝ - 2 σ 2 ∑ t = 1 T c i ( x t - x ^ t s ) ( 21 ) Learned basis functions Θi were used as real vectors A i ℜ and extended with their Hilbert transforms . Such complex basis function dictionary was used to encode monaural sound epochs . Gradients of Eq 5 over phase ϕi and amplitudes ai of complex coefficients zi were equal to: ∂ ∂ a i E 1 ∝ - 2 σ 2 ∑ t = 1 T ( cos ϕ i A i , t ℜ + sin ϕ i A i , t ℑ ) ( x t - x ^ t ) + 2 λ a i log ( 1 + a i 2 ) ( 22 ) ∂ ∂ ϕ i E 1 ∝ - 2 σ 2 ∑ t = 1 T a i ( A i , t ℑ cos ϕ i A i , t ℑ - A i , t ℜ sin ϕ i A i , t ℜ ) ( x t - x ^ t ) ( 23 ) The second layer of the model was trained after the first layer converged , and cofficient values z were inferred for all training data samples . The higher order encoding formed by coefficients s as well as the scaling factor w was inferred via gradient descent on function E2 ( Eq 13 ) : ∂ ∂ s i E 2 ∝ - 2 σ 2 2 ∑ n = 1 2 × T B i , n ( a n - a ^ n ) + κ | w | ∑ m = 1 P sin ( Δ ϕ m - Δ ϕ ^ m ) ξ i , m + 2 λ 2 s i log ( 1 + s i 2 ) ( 24 ) ∂ ∂ w i E 2 ∝ κ w | w | 2 ∑ m = 1 P Δ ϕ ^ m sin ( Δ ϕ m - Δ ϕ ^ m ) + λ w [ ( 1 α ) β β w | w | β - 2 ] ( 25 ) The gradients steered sparse coefficients s to explain amplitude and phase vectors a and Δϕ while preserving maximal sparsity . Simultaneously the multiplicative factor w was adjusted to appropriately scale the estimated vector Δ ϕ ̂ . Finally , learning rules for second-layer dictionaries were given by: ∂ ∂ B i , k E 2 ∝ - 2 σ 2 2 s i ( a k - a ^ k ) ( 26 ) ∂ ∂ ξ i , k E 2 ∝ s i κ | w | sin ( Δ ϕ k - Δ ϕ ^ k ) ( 27 ) Altogether 75000 epochs of binaural sound were used to train the model . Each of them was T = 128 samples long , which corresponded to 16 ms . Both layers were trained separately . Before training the first layer , Principal Component Analyis was perfomed and 18 out of 128 principal components were rejected , which corresponded to low pass filtering the data . Left and right ear sound epochs were shuffled together to create a 150000 sample training dataset for the first layer . The first layer sparsity coefficient λ was set to 0 . 2 . Noise variance σ2 was equal to 2 . The sparse coding algorithm converged after 200000 iterations . A complex-valued dictionary was created by extending the real valued one with Hilbert-transformed basis functions . Amplitude and phase vectors a and ϕ were inferred for each sample using 20 gradient steps . Amplitude vectors were concatenated and transformed with a logarithmic function , and IPD vectors Δϕ were computed by substracting left ear phase vectors ϕL from right ear ones ϕR . The second layer was trained by performing 250000 gradient updates on basis functions B and ξ . The amplitude sparsity coefficient λ2 was set to 1 . The λw parameter was set to 0 . 01 and the noise variance σ 2 2 as well as the von Mises concentration parameter κ were set to 2 . Numerical values of the prior-controlling parameters λ , λ2 , λw as well as noise parameters σ , σ2 , κ were set empirically in this study . By running simulations with multiple parameter settings it has been found that due to the presence of a strong environmental noise in the training recording , noise variances σ , σ2 and the von Mises concentration parameter κ should be relatively large in order to achieve convergence . Sparsity of the high layer representation was set to be larger than that of the first layer in order to mimic the biological intuition that neural responses in the ascending auditory pathway become progressively less redundant and sparser [20 , 60] . It has been found however , that the exact value of sparsity paramaters did not affect the spectrotemporal properties , nor the spatial tuning of the second layer units strongly . The λw parameter which controls the strength of the prior over the multiplicative factor w was set to be relatively small . Otherwise the w prior term in the Eq 16 became too strong and dominated learning , preventing the convergence . More principled and theoretically sound ways of parameter selection are possible . One could ask what are the natural noise levels and sparsity values of the training data by specifying them as hyperparameters of the model and learning the appropriate values . Also the number of basis functions at each level could be treated as a parameter and estimated from the data , not chosen ad-hoc . After extending the model in this way , the choice of the correct parameter setting could be performed by cross-validation or Bayesian model selection ( as in [61] ) . Spectrograms of amplitude basis functions Bi were computed by combining spectrograms of real , first layer basis functions A n ℜ , linearly weighted by a corresponding weight exp ( Bi , n ) . First layer spectrograms were computed using T = 29 windows , each 16 samples ( 0 . 002 second ) long , with a 12 sample overlap . Altogether , F = 128 logarithmically-spaced frequencies were sampled . A two-dimensional fourier transform of each spectrogram was computed using the matlab built-in function fft2 . The amplitude spectrum of obtained transform is called the Modulation Transfer Function ( MTF ) of each second layer feature [40] . The center of mass i . e . the point ( C S , i f , C S , i t ) of each monaural part ( S ∈ {L , R} ) of basis functions Bi was computed in the following way: C S i t = ∑ t t ∑ f M T F ( B S , i ) ( 28 ) C S i f = ∑ f f ∑ t M T F ( B S , i ) ( 29 ) where t and f are time and frequency respectively . To estimate conditional distribution of sparse coefficients given the position and identity of the sound , test recordings of a sound source ( either speech , or rubbed paper ) moving around the recording subject were used . Each source circled the recording person 4 times resulting in 4 recordings . Each of them was divided into 18 intervals . Intervals corresponding to the same area on the circle were joined together across all recordings . For each out of 18 sound positions 3000 random sound chunks were drawn and encoded by the model . Position-conditional ensembles were then used to compute conditional histograms . Conditional mean vectors μi , θ were computed by averaging all values of coefficient si at position θ . Mean vectors were mapped to a [0 , 1] interval by adding the absolute value of a minimal entry and dividing it by the value of the maximum . For plotting purposes in Fig 10 , endings of tuning curves were connected if values at −180° and 180° were not exactly equal . The decoding analysis was performed using K second-layer sparse coefficients s averaged over D of samples . The response vectors d ∈ ℝK were therefore formed as: d = 1 D ∑ i = 1 D s { 1 , … , K } ( 30 ) Such averaging procedure can be interpreted as an analogy to computation of firing rates in real neurons . The marginal distribution of response coefficients d over all 18 sound positions θ ∈ {−180° , −160° , … , 160° , 180°} was equal to: p ( d ) = ∑ θ p ( d | θ ) p ( θ ) ( 31 ) where each conditional p ( d∣θ ) was a K-dimensional Gaussian distribution with class specific mean vector μθ and covariance matrix Cθ: p ( d | θ ) = 𝓝 ( μ θ , C θ ) ( 32 ) The prior over class labels p ( θ ) was uniformly distributed i . e . p ( θ i ) = 1 18 for each i . The decoding procedure iterated over all class labels and returned the one , which maximized the likelihood of the observed data vector . Out of the entire dataset , 80% was used to train the model and remaining 20% to test and estimate the confusion matrix . Confusion matrix M was a joint histogram of a decoded and true sound position θ ^ and θ . After normalization , it was an estimate of a joint probability mass function p ( θ ^ , θ ) . Mutual information was estimated from each confusion matrix as: M I ( θ ^ θ ) = ∑ θ ^ ∑ θ p ( θ ^ , θ ) log 2 ( p ( θ ^ , θ ) p ( θ ^ ) p ( θ ) ) ( 33 ) | Ability to localize the position of a sound source is vital to many organisms , since audition provides information about areas which are not accessible visually . While its importance is undisputed , its neuronal mechanisms are not well understood . It has been observed in experimental studies that despite the crucial role of sound localization , single neurons in the auditory cortex of mammals carry very little information about the sound position . The joint activity of multiple neurons is required to accurately localize sound , and it is an open question how this computation is performed by auditory cortical circuits . In this work I propose a statistical model of natural stereo sounds . The model is based on the theoretical concept of sparse , efficient coding which has provided candidate explanations of how different sensory systems may work . When adapted to binaural sounds recorded in a natural environment , the model reveals properties highly similar to those of neurons in the mammalian auditory cortex , suggesting that mechanisms of neuronal auditory coding can be understood in terms of general , theoretical principles . | [
"Abstract",
"Introduction",
"Results",
"Properties",
"of",
"the",
"second",
"layer",
"representation",
"Discussion",
"Methods"
] | [] | 2015 | The Opponent Channel Population Code of Sound Location Is an Efficient Representation of Natural Binaural Sounds |
To assess if a probabilistic model could be used to estimate the combined prevalence of infection with any species of intestinal nematode worm when only the separate prevalence of each species is reported , and to estimate the extent to which simply taking the highest individual species prevalence underestimates the combined prevalence . Data were extracted from community surveys that reported both the proportion infected with individual species and the combined proportion infected , for a minimum sample of 100 individuals . The predicted combined proportion infected was calculated based on the assumption that the probability of infection with one species was independent of infection with another species , so the probability of combined infections was multiplicative . Thirty-three reports describing 63 data sets from surveys conducted in 20 countries were identified . A strong correlation was found between the observed and predicted combined proportion infected ( r = 0 . 996 , P<0 . 001 ) . When the observed and predicted values were plotted against each other , a small correction of the predicted combined prevalence by dividing by a factor of 1 . 06 achieved a near perfect correlation between the two sets of values . The difference between the single highest species prevalence and the observed combined prevalence was on average 7% or smaller at a prevalence of ≤40% , but at prevalences of 40–80% , the difference was about 12% . A simple probabilistic model of combined infection with a small correction factor is proposed as a novel method to estimate the number of individuals that would benefit from mass deworming when data are reported only for separate species .
The World Health Organization ( WHO ) estimates that intestinal nematode worms , also known as soil-transmitted helminths , are currently endemic in 130 countries in the world [1] . These worms include the common roundworm Ascaris lumbricoides , the whipworm Trichuris trichiura and the hookworms Ancylostoma duodenale and Necator americanus , which are usually treated as a single type as the eggs are indistinguishable under a microscope . In places where these worms are endemic , infections with two or three types , are commonly observed . Such mixed infections may occur randomly , as a simple probabilistic function of the prevalence of each individual species , or there may be factors that result in non-random association between species . The latter is possible , particularly because these worms are all transmitted on soil that has been contaminated with faeces from infected people . A probabilistic model to predict the prevalence of multi-species worm infections in human communities was proposed by Booth & Bundy in 1995 [2] . In testing this model against field data using log-linear analysis , it was found that combined infections with A . lumbricoides and T . trichiura occurred more frequently than expected by chance [2] . The authors concluded that their model was more effective in predicting the numbers of multiple infections involving hookworms than those involving only A . lumbricoides and T . trichiura . As all these worms can be treated using a single dose of an inexpensive anthelmintic drug , the WHO recommends a strategy called “preventive chemotherapy” [3] . This involves annual mass treatment in all communities in which the prevalence of infection with any type of intestinal nematode worm among school-aged children is 20% or more , and twice yearly mass treatment if the prevalence is 50% or higher [3] . When mapping the prevalence of all intestinal nematode infections in order to determine the frequency of treatment , the WHO Global Databank on Schistosomiasis and Soil-Transmitted Helminths simply uses the highest prevalence [4] when surveys do not report the combined prevalence , and give only the separate prevalence of each species . This is done perhaps because the extent to which concurrent infections affect the accuracy of predictions made by the probabilistic model is not known . With the resurgence of interest in controlling soil-transmitted helminth infections , much more field survey data are now available than when the probabilistic model was first proposed and tested [2] . The principal aim of the analysis reported here was to examine the accuracy of the probabilistic model in estimating the combined prevalence of intestinal nematode worm infection using data from a wide range of countries in all regions of the world , but using a simpler mathematical approach that could be easily applied . The subsidiary aims were to estimate the extent to which taking the highest individual prevalence underestimates the combined prevalence and to assess the degree of correlation between the proportions infected with each species of worm .
A database of 230 publications in peer-reviewed journals , grey and unpublished literature that had been compiled in 2003 to estimate the global prevalence of intestinal nematode worms ( described in ref . [5] ) was searched for data that reported both the proportion infected with each species and the combined proportion infected . This was updated with a PubMed search limited to papers published in the last 10 years in English with free online access to the full text , using the terms ‘Soil transmitted helminths prevalence’ and ‘Ascaris AND Trichuris AND hookworm AND prevalence’ . Only data from community-based studies with a sample size of >100 and published after 1990 were included . Where surveys had been carried out in several areas within a country and the results were presented in a geographically disaggregated manner , they were treated separately , rather than as a single data set . Data on prevalence is usually presented in the form of a percentage , with values ranging from 0 to 100 . For purposes of this analysis , the percentage prevalences were converted into the proportion infected , with values ranging from 0 to 1 . Figure 1 represents the seven possible permutations of infections with A . lumbricoides , T . trichiura and the hookworms . The data from each survey were extracted as follows: The proportions infected with each permutation of infection were then calculated as: The combined proportion infected with Ascaris , Trichuris and hookworms ( path ) is thus the sum of all seven equations above: This can be simplified by cancellation to: ( 1 ) If only two species were present , such as Ascaris and Trichuris , then the proportion of double and single infections is calculated in a similar way: So the combined proportion infected with Ascaris and Trichuris ( pat ) is the sum of the three equations above: . This can be simplified by cancellation to: ( 2 ) The same simplified equations for infections with Ascaris and hookworm ( pah ) and Trichuris and hookworm ( pth ) can be written as: ( 3 ) ( 4 ) Equation 1 was applied in an Excel spreadsheet to calculate the predicted combined proportion infected from the data from each survey and the values were plotted against the observed combined proportion infected in the same survey . When only two worms were identified in a survey if a value of zero is entered for the missing type then the spreadsheet calculates the correct proportion infected with either or both species and it is not necessary to apply Equations 2 to 4 . To investigate the degree to which the highest single species prevalence may underestimate the combined prevalence , the differences between the highest individual species value and the observed combined proportion infected were plotted against the observed combined proportion infected . To investigate the degree to which individual species were associated , correlation coefficients ( r ) were calculated for data derived from all surveys for the proportions infected with Ascaris and Trichuris , Ascaris and hookworm , and Trichuris and hookworm .
Thirty-three papers describing surveys conducted in 20 different countries were identified for this analysis: eight in Asia , six in Africa , five in Latin America and the Caribbean , and one in Oceania . Together they contained 63 sets of data: 30 from Asia , 23 from Africa , nine from Latin America & the Caribbean and one from Oceania ( see Annex S1 ) . The observed combined prevalences included in the analysis ranged from 1 . 9% , recorded in the Southern Highlands of Malawi , to 96 . 8% , recorded in Feni District in Bangladesh . Fifteen data sets ( 24% ) had only two worm infections; the rest had three . Figure 2 shows a scatter plot of the observed proportion infected against the predicted combined proportion infected ( r = 0 . 996 , P<0 . 001 ) with the line of equivalence . As the predicted combined proportion infected in Figure 2 tends to be above the line of equivalence , Figure 3 shows the observed combined proportion infected plotted on the x-axis , plotted against the difference between the observed and predicted proportions on the y-axis . The slope of the equation for the line in Figure 3 is 0 . 0596 rounded to 0 . 06 , which indicates that the overestimation shown in Figure 2 increases by 0 . 06 for every 10% increase in prevalence . This provides a factor by which to correct the over-estimation of the predicted combined proportion infected ( path ) so that: ( 5 ) A plot of the observed combined proportion infected against this adjusted predicted combined proportion infected ( not shown since it is almost identical to Figure 2 ) gave an equation for the line of , indicating a intercept of almost zero , a slope of almost 1 and correlation coefficient of r = 0 . 996 . Figure 4 shows a scatter plot of the observed combined proportion infected against the proportion infected with the single most common species , with the line of equivalence . It shows that the prevalence of the single most common species usually underestimates the combined prevalence , with the smallest differences occurring at the lowest and highest prevalences . The correlation coefficient of r = 0 . 973 was less than that between the observed and adjusted predicted combined proportion infected ( 0 . 996 ) . To assess the magnitude of these underestimates in relation to the prevalence , Figure 5 shows a plot of the average difference between the observed combined proportion infected in the 63 data sets and the proportion infected with the single most common species for ten centiles of combined proportions infected . Between 3 and 11 data points were available to calculate the average for each centile . Figure 5 shows that when the observed combined proportion infected is 0 . 4–0 . 8 the observed combined prevalence is about 12% higher than the highest prevalence of any one species , with 95% CI ranging from about 6–18% . The data were also analysed for correlations between proportions infected with Ascaris and Trichuris , Ascaris and hookworm , and Trichuris and hookworm in the 63 pairs of data points . The correlation coefficients were 0 . 544 , 0 . 191 and 0 . 180 respectively , indicating a much stronger correlation between Ascaris and Trichuris than other pairs of infections . Seven of the 63 data sets also presented disaggregated data on the observed number of single , double and triple infections [6]–[11] . Of the seven , six were from communities that had both Ascaris and Trichuris infections; and in five of these , the observed prevalence of co-infection was higher than that expected to occur by chance .
This paper presents a simple equation ( Eqn 1 ) to estimate the combined prevalence of infection with A . lumbricoides , T . trichiura and the hookworms from data on the separate prevalence of infection with each type . The combined prevalence can then be corrected ( Eqn 5 ) to allow for the estimated degree of association between types , probably between A . lumbricoides and T . trichiura . The strong correlation reported here between the observed and predicted combined prevalences supports the hypothesis proposed by Booth & Bundy [2] that when the three main species of intestinal nematode worms co-occur , the probability of infection with one species is largely independent of infection with another . This is despite the use of two different forms of analysis in testing the probabilistic model against field data . The results presented here also confirm the findings of Booth & Bundy that concurrent infections of A . lumbricoides and T . trichiura are more common than expected by chance . Several other studies have also noted this association [12]–[19] , which probably arises from their common mode of transmission . One hypothesis argues that both these worms are transmitted in a “domestic” domain , within and around the house , while hookworm is transmitted in a public domain [20] . However , the present analysis also shows that a small downward correction of the predicted combined proportion infected is enough to achieve a very high correlation between predicted values and values reported by field surveys . The apparent over-estimation of combined prevalence probably results from the association between A . lumbricoides and T . trichiura . This over-estimation does not seem to be very large however , and is easily corrected . Equation 1 , to estimate the combined proportion infected , and Equation 5 , which provides a correction factor , could thus provide a novel and relatively simple and practical method to estimate the combined prevalence of infection with any intestinal nematode worm from data published on the prevalence of separate species . This analysis does not take into account potential errors in parasitological diagnosis , particularly false negatives leading to an underestimated prevalence . The sensitivity and specificity of diagnosis are likely to be related to the concentration of eggs in faeces , which is related to fecundity of worms , the dispersal and dilution of eggs in the faecal mass , and to the amount of faeces examined under a microscope [21] . As these factors affect all three types of worms if present , they should not affect the analysis presented here , only that any combined prevalence could be an underestimate of the true prevalence of infection . The difference between the prevalence of the single most common species of intestinal nematode , which is currently used by the WHO in the absence of data on combined prevalence , and the observed combined prevalence , seems to vary depending on the prevalence . At a combined prevalence of ≤40% , the difference is on average 7% or smaller , but when the combined prevalence is higher , the difference is about 12% . The difference is less also when the combined prevalence is very high ( >90% ) . This has implications for mass treatment , especially at low prevalence rates . For example , if the proportion infected with Ascaris is 0 . 15 , and the proportion infected with Trichuris is also 0 . 15 , then the combined proportion infected is 0 . 2775 ( 0 . 15+0 . 15−0 . 0225 ) . This prevalence of 28% is above the threshold at which the WHO currently recommends mass treatment , but the highest single species prevalence of 15% is below the threshold value of 20% . At higher prevalences the underestimation due to using of the highest single species prevalence is of less importance . This analysis includes a modest number of data sets from all major geographical areas where intestinal nematode infections are endemic . It suggests that a simple probabilistic model with a small correction could be used to estimate the proportion of people infected with any intestinal nematode worm . This could help with the global mapping of disease and is likely to increase the estimated number of individuals that would benefit from mass deworming in the world today . | Mixed infections with roundworm , whipworm and hookworm are common , but survey reports often give only the separate prevalence of each type . However , the combined prevalence is important to estimate accurately the number of individuals who would benefit from control programmes and to make decisions about the frequency of treatment . Previous work suggests that mixed infections involving hookworm occur randomly , but that roundworm and whipworm infections are found together more frequently than would be expected by chance . We used 63 data sets from community surveys that reported both the proportions infected with individual types of worms and the combined proportion infected with any worm . We then calculated the proportion that would be infected with any type of worm if infections had occurred randomly and compared it with the observed combined proportion infected . We found a strong correlation between the observed and predicted combined proportions infected . A small downward correction of the predicted proportion infected by dividing by a factor of 1 . 06 brought it to a value that nearly equalled the observed proportion infected almost all the time . This simple model could be applied to published survey data to estimate accurately the number of individuals that would benefit from mass deworming . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"public",
"health",
"and",
"epidemiology/infectious",
"diseases"
] | 2010 | Using the Prevalence of Individual Species of Intestinal Nematode Worms to Estimate the Combined Prevalence of Any Species |
Reducing social distance between hospital staff and patients and establishing clear lines of communication is a major challenge when providing in-patient care for people afflicted by Buruli ulcer ( BU ) and chronic ulcers . Research on hospitals as therapeutic communities is virtually non-existent in Africa and is currently being called for by medical anthropologists working in the field of health service and policy planning . This paper describes a pioneering attempt to establish a therapeutic community for patients suffering from BU and other chronic ulcers requiring long term hospital care in Benin . A six-month pilot project was undertaken with the objectives of establishing a therapeutic community and evaluating its impact on practitioner and patient relations . The project was designed and implemented by a team of social scientists working in concert with the current and previous director of a hospital serving patients suffering from advanced stage BU and other chronic ulcers . Qualitative research initially investigated patients’ understanding of their illness and its treatment , identified questions patients had about their hospitalization , and ascertained their level of social support . Newly designed question–answer health education sessions were developed . Following these hospital wide education sessions , open forums were held each week to provide an opportunity for patients and hospital staff to express concerns and render sources of discontent transparent . Patient group representatives then met with hospital staff to problem solve issues in a non-confrontational manner . Psychosocial support for individual patients was provided in a second intervention which took the form of drop-in counseling sessions with social scientists trained to serve as therapy facilitators and culture brokers . Interviews with patients revealed that most patients had very little information about the identity of their illness and the duration of their treatment . This knowledge gap surprised clinic staff members , who assumed someone had provided this information . Individual counseling and weekly education sessions corrected this information gap and reduced patient concerns about their treatment and the status of their healing process . This led to positive changes in staff–patient interactions . There was widespread consensus among both patients and staff that the quality of communication had increased significantly . Open forums providing an opportunity for patients and staff to air grievances were likewise popular and patient representative meetings resulted in productive problem solving supported by the hospital administration . Some systemic problems , however , remained persistent challenges . Patients with ulcers unrelated to BU questioned why BU patients were receiving preferential treatment , given special medicines , and charged less for their care . The idea of subsidized treatment for one disease and not another was hard to justify , especially given that BU is not contagious . This pilot project illustrates the basic principles necessary for transforming long term residential hospitals into therapeutic communities . Although the focus of this case study was patients suffering from chronic ulcers , the model presented is relevant for other types of patients with cultural adaptation .
Providing in-patient care for people afflicted with diseases requiring long-term hospitalization is a major challenge in low-income countries . In these countries , health staff must manage patients with limited resources . At the same time , patients struggle to maintain a positive attitude while far from their families and burdened by concerns about both the progress of their treatment and the welfare of their households during their absence . Patients and hospital staff live and work in close quarters , yet they are often socially distant , their interactions cordial yet primarily focused on disease management tasks . While considerable literature exists in developed countries on the hospital as a social system and on formation of therapeutic communities to care for long-term patients ( primarily mental health and substance abuse patients ) [1 , 2] , hospital based research on other types of therapeutic communities is sparse , and virtually non-existent for Africa . This paper describes a pioneering attempt to establish a therapeutic community for patients suffering from Buruli ulcer ( BU ) and other chronic ulcers requiring long-term care in Benin , West Africa . The hallmark of a hospital-based therapeutic community , as we define it in this paper , is a communication process that invites open dialogue between patients and health staff , patient participation in problem-solving associated with everyday living , ways and means of resolving conflicts that arise , and information exchange that fosters adherence , as distinct from one-sided directives demanding compliance . Our definition of therapeutic community is based on the principle of mutual respect and recognition that respect is only forthcoming when patients and staff better understand the works , responsibilities , challenges , and constraints each faces . Buruli ulcer ( BU ) is the third most common mycobacterial disease in the world . A majority of cases are found in West Africa [3] . It is a neglected tropical disease affecting poor rural villagers in several West African countries . Cases diagnosed early can be cured with 56 doses of a combined regimen of intramuscular streptomycin and oral rifampicin . Treatment of advanced cases of BU often requires surgery and long-term residential treatment . During their stay in hospital , a patient’s dressings must be changed daily or at least three times a week , and the patient must undergo physical therapy to prevent disabilities and joint contractures [4 , 5] . The Allada Buruli Ulcer Treatment Center ( CDTUB ) is one of the four primary reference centers for BU care in Benin and a recognized center of excellence for clinician training in BU management . The hospital also treats patients suffering from other types of chronic ulcers of various etiologies such as sickle-cell disease , necrotizing fasciitis , and phagedenic or vascular ulcers . Since BU treatment is the primary vocation of the center , BU patients receive subsidized treatment thanks to the government and international NGOs . Patients with advanced BU residing at the hospital require extensive post-operative care . Other chronic ulcer patients have to pay for much of their therapy out of pocket . When patients suffering from more advanced stages of BU and other chronic ulcers come to hospitals like Allada , they have to adapt to a new way of life in unfamiliar surroundings . They have to learn to get along with other patients who are members of groups they have had little contact with in the past . They then have to cope with the uncertainty of their illness trajectory , the demands of treatment , and the physical discomfort associated with the frequent changing of bandages and physical therapy sessions . For more advanced cases requiring skin grafts , the duration of treatment is uncertain and difficult to predict due to individual variability in wound healing . Given that the duration of BU treatment is long , and patients are unable to care for themselves , family caretakers are asked to accompany patients and attend to their daily needs such as cooking , washing clothes , and daily assistance . One of the main conditions for being admitted to the hospital is identifying a suitable caretaker from one’s extended kin network . This is often difficult , as removing household members responsible for agricultural operations or child care at home can place the wellbeing of an entire household in peril [6] . In some cases , caretakers come and go , and in other cases they are not able to remain at the hospital and the patient is abandoned [7] . Food is partly provided free of charge for BU patients , but not for their caretakers , and not for patients suffering from other types of chronic ulcers . Although treatment is subsidized for BU patients , there are indirect costs related to hospitalization that can prove burdensome .
The CDTUB is located in Allada , a small city of 127 , 493 inhabitants located in Benin ( West Africa ) [8] . It is staffed by four doctors , 18 nurses , eight laboratory technicians , six support staff , three maintenance workers and three drivers . The director of the hospital is a doctor actively engaged in the care of BU patients as well as BU-related research . He is assisted by an administrative staff composed of five secretaries and accountants . The hospital receives approximately 200 new patients a year , out of which around 40 are BU cases . BU patients typically remain in the hospital for 8–18 months , but some remain much longer . Patients in the hospital range from 2 to 70 years of age , with 60% being children . There is an even split between male and female residents in the hospital , with residents divided into nine wards segregated by gender . Caretakers range in age from 9 to 50 years of age , and an overwhelming majority ( over 90% ) are female . At the CDTUB , all patients are required to obey rules put in place by the hospital administration to assure a sense of order as well as quality of care . Compliance with hospital policies is mandatory . At the time of the pilot project , patients were treated as passive recipients of care and not provided much knowledge about their disease beyond being told what medications , if any , they were required to take and how to assist in the cleaning and bandaging of their wounds . For patients , their stay at the hospital was a highly liminal experience marked with much apprehension and uncertainty . Ethical approval was obtained from Benin’s National Ethical Committee of Health Research before the start of the research ( IRB00006860 N° 148 /MS/DC/SGM/DFRS/CNPERS/SA ) . Informed consent procedures already in place at Allada hospital were strictly adhered to over the course of the project . All patients and staff interviewed were assured that their opinions would be kept confidential . Patients were assured that information volunteered would in no way affect the quality of their care at the hospital . Patients were also reassured that issues discussed at open forum meetings would not result in negative actions by the staff , and a grievance process was put into place to make sure this did not occur . Oral consent was documented by the presence of witness . The use of oral consent is approved by the ethical review board because many study participants were illiterate . When a participant was under 18 years of age , both the child/adolescent and his/her caretaker were informed about the nature and aim of study before being asked to give consent .
Three core challenges to establishing a therapeutic community were identified during the pilot project . The first challenge is how to establish an open forum where patients and staff feel comfortable enough to speak their minds without fear of reprisal . If staff feels they are being criticized and that this will have negative impact on their job performance , they will assume a defensive posture . This challenge requires the active support of the hospital director and hospital administration . In the present case , the hospital director let it be known that he viewed the airing of discontent as the first step of a problem solving process that was valued at the hospital . Establishing trust in this process took time and required change on the part of all members of the therapeutic community . By the end of the seven -month pilot project , all stakeholders interviewed had enough trust in the process to feel they could communicate their problems without compromising their position or the quality of care they received . The second challenge faces social scientists attempting to establish a therapy facilitator/cultural broker role . It is important that they not be seen as the handmaiden of the hospital administration or an advocate for either health staff or patients . Trust demands a neutral position where the charge of the social scientist is to identify , investigate and present all sides of a dispute and to provide in depth understanding of issues affecting administration–staff–patient relations . During the project , there were times when various parties attempted to gain the support of a social scientist in opposition to another . It became important for the social scientist to be clear about what they can and cannot do as part of a process of problem solving . For example , when a patient became destitute because they lost a caretaker or the resources needed for treatment , the social scientists assisted the patient in presenting a case to the administration , but could not be seen as directly solving the resource problem themselves . During the community outreach program that preceded the therapeutic community intervention , the social science team created a resource assessment screening tool to facilitate patient referral to the hospital . The same assessment tool was used in the hospital when an economic crisis was revealed to a social scientist . The screener enabled the case to be systematically presented to the administration after all data necessary to make a decision had been acquired . A third challenge is sustainability and cost-effectiveness of the social scientist role . The therapeutic community model presented in the study requires the presence of a social scientist and justification for the resources needed to support the position . Based on the results of the pilot study , the Allada hospital administration has decided to employ a social scientist to assist in therapy facilitation and community- based outreach activities , and to secure the services of a psychologist in cases where patients need to be treated for mental health problems requiring medication . In this paper , we have described a pioneering attempt to transform an African hospital serving long-term residential patients into a therapeutic community . Although the focus of this case study is BU patients , the model and experience presented here are relevant for many other types of patients . It requires a rethinking of hospital staff–patient relations in concert with the tenets of patient-centered and humanized patient care [19–22] and people-centered health policy [23 , 24] . For patients , it addresses their concerns , enhances their sense of well-being , and provides a sense of support and compassion during their long hospital experience . For staff , it leads to greater patient adherence and the resolution of conflicts that can compromise care . In addition , it provides staff as well as patients a forum to articulate their grievances . And for administrators , it provides them with a finger on the pulse of everyday life in the hospital such that tensions can be identified and resolved , policies revisited , and greater transparency provided when necessary . The pilot project proved to be highly successful as assessed by patients , staff , and administrators . Communication patterns improved , patient uncertainty about the status of wound healing decreased , and patients became far more knowledgeable about their illness . Socially , petty disputes were resolved in a far more amicable fashion , and both patients and staff felt vindicated by expressing discontentment and being heard by others , who could then better understand their position . The pilot project made use of two distinct but complementary forms of problem solving as a means to establish a therapeutic community in keeping with culturally meaningful modes of conflict resolution in Africa . Much has been written in the anthropological literature about the value of both collective and individual forms of conflict resolution in settings ranging from the settling of social disputes between factions in villages , to processes of divination used to air grievances both past and present [25 , 26 , 27] . An open forum both facilitated collective problem solving and enrolled public support for one’s position , serving to establish their moral identity [28] . Individual counseling provided a patient the complementary opportunity to speak to an empathetic witness [29] about difficulties that one would not like to share in public , for reasons ranging from embarrassment to spiritual danger . Is it feasible to transform African hospitals serving long-term patients into therapeutic communities ? We would argue that it is feasible given two conditions . First , hospital administrators need to recognize the utility of building a therapeutic community and be willing to engage in the problem-solving processes outlined in this paper . Second , health social scientists need to receive basic training in health systems analysis and conflict resolution as well as hospital ethnography [30 , 31] and an anthropological approach to patients’ illness experience attentive to their many “works of illness” . Treating patients as active agents in the hospital will serve as a corrective to paternalistic approaches to patient care that treat them as passive recipients of treatment whose only work is compliance with medical advice [32 , 33] . Life is far more complicated , and when both patient and staff needs are not met , discontent undermines quality of care . We would end with one last observation . There is another important way establishing a therapeutic community benefits the hospital . Former satisfied patents are positive sources of information about both the hospital and the community based outreach program it has promoted to identify early category one BU cases . As the old adage goes: the best advertisement is a satisfied customer . This is particularly important in a disease like BU , where the reputation of the hospital is essential to the success of community outreach and the entire BU program . Patients educated in wound care as well as BU re-enter the community as a valuable resource and “go to” person for information about the disease and wound management . In Benin , former patients already play an active role in identifying cases of BU in some communities [34] . Increased patient education and a more positive experience in the hospital increases the likelihood that they will refer chronic ulcer patients to health staff they know and trust . | Little is known about communication patterns and social relations between health staff and long -term patients in African hospitals . An ethnography of a reference hospital treating patients afflicted with Buruli Ulcer ( BU ) and other chronic ulcers in Benin was conducted . Sources of psychosocial distress and communication patterns compromising quality of care were documented . Based on this research , an intervention was mounted to transform the hospital into a higher functioning therapeutic community . Question: answer education sessions were introduced to provide patients the opportunity to inquire about their illness , it’s treatment and trajectory; weekly open- forums were established to give patients and hospital staff a chance to air grievances; patient representatives met with hospital staff to resolve problems in a non-confrontational manner , and psychosocial support for individual patients was provided through drop-in counseling sessions with social scientists in residence . Patients reported positive changes in the quality of their care and interactions with care providers , care providers reported that the problem solving process instituted was productive , and hospital administrators actively supported efforts to improve social relations and lines of communication . Systemic problems related to perceptions of preferential treatment for BU patients provided subsidized treatment supported by a national program remained contentious . | [
"Abstract",
"Introduction",
"Methods",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"health",
"services",
"research",
"sociology",
"tropical",
"diseases",
"social",
"sciences",
"health",
"care",
"bacterial",
"diseases",
"research",
"design",
"scientists",
"signs",
... | 2016 | Steps Toward Creating A Therapeutic Community for Inpatients Suffering from Chronic Ulcers: Lessons from Allada Buruli Ulcer Treatment Hospital in Benin |
Recently , a molecular pathway linking inflammation to cell transformation has been discovered . This molecular pathway rests on a positive inflammatory feedback loop between NF-κB , Lin28 , Let-7 microRNA and IL6 , which leads to an epigenetic switch allowing cell transformation . A transient activation of an inflammatory signal , mediated by the oncoprotein Src , activates NF-κB , which elicits the expression of Lin28 . Lin28 decreases the expression of Let-7 microRNA , which results in higher level of IL6 than achieved directly by NF-κB . In turn , IL6 can promote NF-κB activation . Finally , IL6 also elicits the synthesis of STAT3 , which is a crucial activator for cell transformation . Here , we propose a computational model to account for the dynamical behavior of this positive inflammatory feedback loop . By means of a deterministic model , we show that an irreversible bistable switch between a transformed and a non-transformed state of the cell is at the core of the dynamical behavior of the positive feedback loop linking inflammation to cell transformation . The model indicates that inhibitors ( tumor suppressors ) or activators ( oncogenes ) of this positive feedback loop regulate the occurrence of the epigenetic switch by modulating the threshold of inflammatory signal ( Src ) needed to promote cell transformation . Both stochastic simulations and deterministic simulations of a heterogeneous cell population suggest that random fluctuations ( due to molecular noise or cell-to-cell variability ) are able to trigger cell transformation . Moreover , the model predicts that oncogenes/tumor suppressors respectively decrease/increase the robustness of the non-transformed state of the cell towards random fluctuations . Finally , the model accounts for the potential effect of competing endogenous RNAs , ceRNAs , on the dynamics of the epigenetic switch . Depending on their microRNA targets , the model predicts that ceRNAs could act as oncogenes or tumor suppressors by regulating the occurrence of cell transformation .
The characteristics of cancer rest on many biological capabilities acquired during the multistep of the development of tumors [1] . These biological properties include sustaining proliferative signaling , evading growth suppressors , resisting cell death , allowing replicative immortality , promoting angiogenesis , and eliciting formation of metastasis [1] . The progression from normal cells to cancer could be strongly influenced by the tumor microenvironment . In that context , many studies have shown close relations between inflammation and different types of cancer [2]–[4] . Inflammatory molecules , such as the interleukin-6 ( IL6 ) or the transcription factor NF-κB , could provide growth signals , which elicit the proliferation of malignant cells [5] , [6] . However , until recently , the molecular regulatory network linking inflammation to cell transformation was poorly understood . To study the molecular link between inflammation and cancer , Iliopoulos and coworkers used an experimental model of oncogenesis , which involves a derivative of MCF10A , a spontaneous immortalized cell line derived from normal mammary epithelial cells containing ER-Src , a fusion of the oncoprotein Src with the ligand binding domain of estrogen receptor [7] . They demonstrated that transient treatment with tamoxifen results in stable cell transformation , defined by their invasive capabilities , their increased motility , as well as their ability to form mammospheres ( multicellular structure enriched in cancer stem cells ) . This stable cell transformation can be defined as an epigenetic switch , which corresponds to a stable cell change to another phenotype without any change in DNA sequence . The triggering event of the epigenetic switch is mediated by a transient inflammatory signal driven by the Src oncoprotein [7] . After the triggering event , the transformed state of the cell mediated by the epigenetic switch is stable for many generations . A positive inflammatory feedback loop driven by the transcription factor NF-κB , the microRNA binding protein Lin28 , the Let-7 microRNA , and IL6 is responsible for the maintenance of this transformed state [7] . Besides the critical role of microRNAs in the regulation of protein expression [8]–[10] or in conferring robustness of biological processes [11]–[14]; microRNAs , such as miR-122 , are at the core of molecular regulatory feedback loop driving hepatocyte differentiation [15] , or microRNAs , such as Let-7 microRNA family , are critical regulators in cancer development and progression [16]–[19] . The positive inflammatory feedback loop discovered by Iliopoulos and coworkers brings to light the importance of microRNAs in the molecular regulatory circuit linking inflammation to cell transformation [7] , [20] . Other studies showed that epigenetic regulations could play critical role in cancer [21]–[24] , and open novel perspectives in understanding developmental processes and diseases [25] . Here , we propose a computational model of a large network involving kinases , transcription factors , messenger RNAs and miRNAs to account for the qualitative dynamics of the epigenetic switch linking inflammation to cell transformation [7] , [20] . By means of a deterministic model for the epigenetic switch , we will study the dynamical nature of the positive inflammatory feedback loop leading to cell transformation . We will see how tumor suppressors ( inhibitors of this positive feedback loop ) or oncogenes ( activators of the inflammatory feedback loop ) regulate the occurrence of cell transformation . By resorting to stochastic simulations as well as deterministic simulations in a heterogeneous cell population , we will assess the role of random fluctuations ( resulting from molecular noise or from cell-to-cell variability ) on the dynamics of the epigenetic switch . Furthermore , based on recent experiments describing the potential crucial role of competing endogenous RNA ( ceRNA ) as microRNA sponge on the control of molecular regulatory network [26]–[29] , we will study the effect of ceRNAs on the dynamical behavior of the model for the epigenetic switch linking inflammation to cell transformation .
We propose a computational model for the epigenetic switch linking inflammation to cell transformation ( see Fig . 1 as well as [7] , [20] ) . The model is based on a positive inflammatory feedback loop between NF-κB , Lin28 , Let-7 microRNA and IL6 . An inflammatory signal mediated by the oncoprotein Src activates NF-κB , which promotes the synthesis of Lin28 . The microRNA-binding protein Lin28 rapidly reduces the synthesis of mature Let-7 microRNA , which ensures a higher production of IL6 than achieved directly by NF-κB . In turn , IL6 triggers the activation of NF-κB and promotes the synthesis of the transcription factor STAT3 , whose activity is crucial to induce cell transformation [30]–[35] . As shown by Iliopoulos and coworkers [20] , STAT3 is not only a downstream output of the inflammatory regulatory signal , but is part of the positive feedback loop linking inflammation to cancer . Indeed , STAT3 triggers the synthesis of miR-21 and miR-181b-1 microRNAs [20] . For simplicity , we only consider miR-21 in the model ( see Fig . 1 ) . The latter microRNA down-regulates the translation of the tumor suppressor PTEN , which is responsible for the inhibition of the activity of NF-κB ( see Fig . 1 as well as [20] , [36] , [37] ) . Furthermore , we also consider the positive feedback loop between NF-κB , Lin28 , Let-7 and the oncoprotein Ras . Indeed , NF-κB activates the synthesis of Lin28 , which inhibits the synthesis of Let-7 . The down-regulation of Let-7 prevents it to repress the synthesis of Ras , which promotes the activation of NF-κB ( see Fig . 1 ) . In healthy cells , negative feedback loops between NF-κB and IκB protein family control NF-κB oscillations and ensure the occurrence of a transient inflammatory response within the cell [38]–[42] . However in the immortalized cell line [7] , it seems that NF-κB is rapidly activated through I-κBα phosphorylation , but its activity remains elevated during all the cell transformation process . For that reason and in order to confine our model to the dynamics of the epigenetic switch , we do not consider negative feedback regulation between NF-κB and IκB . The model proposed for the molecular mechanisms of the epigenetic switch , linking inflammation to a stable cell transformation , counts 14 kinetic equations describing the time evolution of the concentration of the different variables considered ( see Section “Methods” here below , as well as Table 1 for a definition of the different variables of the model ) . As observed in the experiments [7] , a transient inflammatory signal mediated by Src is sufficient to trigger the switch from an immortalized cell line to a stable transformed state ( see Fig . 2 ) . Time evolution of Let-7 , Lin28 , IL6 , NF-κB and STAT3 is shown in the absence of Src ( Fig . 2A ) , in the presence of a constant , low level of Src ( Fig . 2B ) , in the presence of higher level of Src for only 5 min ( Fig . 2C ) , or in the presence of a constant higher level of Src ( Fig . 2D ) . In the absence of Src , a high level of Let-7 together with low levels of Lin28 , IL6 , NF-κB and STAT3 are observed ( Fig . 2A ) . We define those expression levels as a normal , non-transformed state of the cell . However , with a constant level of Src ( Fig . 2B , D ) or with a transient level of Src for only 5 min ( Fig . 2C ) , the model exhibits a switch-like behavior in the expression levels of Let-7 , Lin28 , IL6 , NF-κB and STAT3 . Indeed , the level of Let-7 is reduced while levels of Lin28 , IL6 , NF-κB and STAT3 are highly increased at about t = 35 h ( Fig . 2B ) , t = 65 h ( Fig . 2C ) , and t = 10 h ( Fig . 2D ) . In the model , we define this switch as the passage to cell transformation . As in the experiments [7] , the model shows that cell transformation can occur with an inflammatory signal of only 5 min . In the latter case , the occurrence of cell transformation is slow , and happens at t = 65 h instead of t = 36 h with a constant level of Src ( compare Fig . 2B and 2C as well as [7] ) . The model is characterized by a biphasic regulation of the different variables . For instance , we can observe a first slow increase of IL6 followed by a boost of IL6 expression . This dynamical behavior may correspond to the experimental observation showing also a biphasic regulation of IL6 expression [7] . Furthermore , the model also accounts for the experimental observations showing that the positive feedback loop involving NF-κB , Lin28 , Let-7 , and IL6 is required for maintenance of the transformed state of the cell [7] . Indeed , from a stable transformed state , a transient inhibition of either NF-κB , Lin28 , or IL6 ( as in the experiments ) abolishes the transformed state and brings back the normal , non-transformed state of the cell ( see Fig . S1 ) . Let-7 microRNA and PTEN are tumor suppressors and negative regulators of the epigenetic switch leading to cell transformation ( see [7] as well as Fig . 1 ) . By resorting to the computational model , we can assess the role of these tumor suppressors on the dynamics of the switch linking inflammation to cell transformation . To this end , we analyze the dynamical behavior of the model by means of bifurcation diagrams , which bring to light the steady-state levels of the system , i . e . equilibrium levels of the different variables of the model , as a function of a parameter value . In Fig . 3 , the steady-state levels of NF-κB , Let-7 , IL6 and STAT3 are shown as a function of the inflammatory signal , Src , for different rates of synthesis of Let-7 , VSLET7 . The model indicates that the epigenetic switch linking inflammation to cell transformation behaves as an irreversible bistable switch towards the inflammatory signal , Src . Indeed , in the presence of low levels of Src , the model predicts that a non-transformed state defined by low levels of NF-κB , IL6 and STAT3 together with a high level of Let-7 coexists with a transformed state characterized by high levels of NF-κB , IL6 and STAT3 together with a low level of Let-7 . Depending on the value of initial conditions , the system reaches the non-transformed or the transformed state . In the presence of high level of Src , only the transformed state remains . The model suggests that Let-7 microRNA acts as a tumor suppressor because it moves the threshold leading to cell transformation to higher level of Src ( see Fig . 3A–D when VSLET7 increases from 3 to 8 ) . The model also predicts that , when the rate of synthesis of Let-7 is too small , only the transformed state is present ( see curves in Fig . 3 when VSLET7 = 1 ) . Steady-state levels of Let-7 and IL6 vs Src illustrated for different rates of transcription of PTEN , VSMPTEN , indicate that the tumor suppressor PTEN exhibits a similar effect as Let-7 on the dynamics of the epigenetic switch leading to cell transformation ( see Fig . S2 ) . Indeed , an increase in VSMPTEN moves the threshold defining cell transformation to higher levels of Src . Contrary to Let-7 and PTEN , Ras is a positive regulator of the epigenetic switch leading to cell transformation ( see Fig . 1 as well as [7] , [43] ) . Bifurcation diagrams of Let-7 , IL6 and Ras vs Src are shown for different rates of transcription of Ras , VSMRAS in Fig . 4A–C , respectively . In the framework of the epigenetic switch linking inflammation to cancer , the model indicates that Ras acts as an oncogene by moving the threshold leading to cell transformation to smaller levels of the inflammatory signal , Src ( see Fig . 4 when VSMRAS increases from 0 to 0 . 027 ) . For high level of Ras , i . e . VSMRAS = 0 . 03 , the model predicts that the cell is only present in a transformed state ( low level of Let-7 together with high levels of IL6 and Ras ) regardless of the level of Src . It has been shown in many biological systems that stochastic transitions may represent a driving force in development and in cell fate decisions [44] . Here , by resorting to a stochastic version of the model for the epigenetic switch linking inflammation to cancer ( see Table S1 in Supporting Information ) , we assess the effect of stochastic fluctuations on the dynamics of the epigenetic switch . Deterministic ( Fig . 5A ) and the corresponding stochastic time evolution ( Fig . 5C , E ) of NF-κB , Lin28 , Let-7 , IL6 and STAT3 indicate that while deterministic transition to a transformed state of the cell occurs at about t = 20 h ( Fig . 5A ) , stochastic fluctuations may prevent the occurrence of such transition ( compare Fig . 5C and 5E ) . The opposite is also observed: stochastic fluctuations driving the occurrence of cell transformation while the deterministic version of the model predicts a non-transformed state ( compare Fig . 5B with Fig . 5D and 5F ) . Could the tumor suppressors , Let-7 microRNA and PTEN , or the oncogene Ras influence the occurrence of stochastic transitions in the model linking inflammation to cell transformation ? To answer that question , we constructed deterministic as well as stochastic bifurcation diagrams of IL6 vs Src for different rates of synthesis of Let-7 , PTEN , or Ras . Deterministic bifurcation diagrams of IL6 vs Src are shown when the rate of synthesis of Let-7 , VSLET7 is equal to 3 , 3 . 5 and 4 in Fig . 6A , C , E , respectively . The corresponding stochastic bifurcation diagrams are illustrated in Fig . 6B , D , F , respectively . With the initial conditions used ( see section: “Methods” ) , deterministic simulations indicate that the cell is always in a transformed state when VSLET7 = 3 or 3 . 5 ( red dots in Fig . 6A , C ) ; while , when VSLET7 = 4 , the cell is in a non-transformed state for low levels of Src and in a transformed state for higher levels of Src ( see red dots in Fig . 6E ) . The corresponding stochastic simulations for the same conditions indicate that , for low levels of Src , the proportion of cells present in a non-transformed state is larger when VSLET7 increases ( compare Fig . 6B , D and F ) . Thus , the model indicates that high levels of Let-7 may enhance the robustness of the non-transformed state to stochastic fluctuations by reducing the occurrence of stochastic switches leading to cell transformation . The fact that the deterministic unstable steady state is closer to the stable steady state with a low level of IL6 ( non-transformed state ) when VSLET7 is small is already an indication that the latter stable steady state may be less robust to stochastic fluctuations in the presence of low levels of Let-7 ( compare dashed curves in Fig . 6A , C , E ) . Similarly to Let-7 microRNA , the model shows that the tumor suppressor PTEN is also able to increase to robustness of the cell towards stochastic cell transformation ( Fig . S3 ) . Deterministic bifurcation diagrams of IL6 steady-state levels vs Src indicate that the stable steady state corresponding to a non-transformed state ( low level of IL6 ) is larger when the rate of transcription of PTEN , VSMPTEN is high ( compare Fig . S3A , C ) . In the stochastic approach , the model shows that increasing the level of PTEN reduces the occurrence of cell transformation for low level of Src ( compare Fig . S3B and S3D ) . Deterministic and stochastic simulations performed for different rates of transcription of the oncogene Ras , VSMRAS suggest that , contrary to Let-7 or PTEN , an elevated level of Ras increases the sensitivity of the switch to stochastic fluctuations for the occurrence of cell transformation ( see Fig . 7 ) . Indeed , deterministic bifurcation diagrams indicate that the stable steady state with a low level of IL6 ( corresponding to a non-transformed state ) gets smaller when VSMRAS increases from 0 . 005 to 0 . 027 ( see Fig . 7A , C , E ) . The corresponding stochastic simulations suggest that a robust switch without transformed cells , for low level of Src , is present when VSMRAS is small ( Fig . 7B ) . By increasing the level of Ras , the switch leading to cell transformation becomes less robust to stochastic fluctuations and a significant proportion of transformed cells appear even for a low level of inflammatory signal , Src ( Fig . 7D , F ) . Besides ‘intrinsic noise’ coming from molecular noise ( stochastic fluctuations ) , the dynamics of the inflammatory network may be also influenced by ‘extrinsic noise’ originating from to cell-to-cell variability in a population . In that framework , recent studies have shown that the dynamics of activation of NF-κB is heterogeneous within a cell population [40] , [45] , [46] . Those studies suggested that the dynamics of cell responding to inflammatory signaling is partly driven by stochastic processes , but seems mostly controlled by pre-existing variation in internal variables of the cells ( extrinsic noise ) [40] , [45] , [46] . To address the issue of cell-to-cell heterogeneity on the dynamics of the network , we will study , in the next section , the dynamics of the epigenetic switch in a heterogeneous cell population . In order to see if this epigenetic switch is relevant for human cancers , Iliopoulos and coworkers examined the expression levels of IL6 and Let-7 in cancer and normal breast , prostate , hepatocellular , and lung tissues [7] . They showed that cancer tissues have higher levels of IL6 and lower levels of Let-7 as compared to normal tissues . A negative correlation is also observed in the expression levels of Let-7 and IL6 for breast , prostate , and hepatocellular tissues [7] . By resorting to the deterministic model proposed for the switch leading to cell transformation , we are able to reproduce qualitatively these experimental observations . To do so , we analyze the expression pattern of IL6 and Let-7 in a model for a heterogeneous cell population . For each cell in the population , which counts 100 cells , all the parameters values are chosen randomly with 10% of variation from their default value ( see Fig . 8 ) . The model for the cell population shows that a normal , non-transformed state can be achieved with high levels of Let-7 together with low levels of IL6 ( Fig . 8A ) . From that state , the model predicts that an increase in the level of the oncogene Ras triggers cell transformation in a significant proportion of cells in the population ( see Fig . 8B ) , which results in a mixed population of non-transformed and transformed cells . A further increase in Ras almost completely switches the cell population to a transformed state , defined by high levels of IL6 and low levels of Let-7 ( Fig . 8C ) . The model indicates that by starting from a non-transformed state of the cell population ( Fig . 8A ) , the switch of the cell population to a transformed state can be also triggered by reducing the rate of synthesis of Let-7 , VSLET7 ( Fig . 8D ) , or by increasing the level of Lin28 ( Fig . 8E ) . Finally , the model predicts that a transformed state of the cell population ( Fig . 8C ) could switch back to a non-transformed state by increasing the level of the tumor suppressor PTEN ( Fig . 8F ) . This latter result supports the experimental observations showing that elevated levels of PTEN induce a tumor-suppressive metabolic state [47] . In all cases , the model indicates , as in the experiments [7] , that the expression pattern of Let-7 and IL6 is negatively correlated . To assess the robustness of the model with respect to random variation of parameters in a heterogeneous cell population , we further illustrate the expression levels of Let-7 and IL6 in a heterogeneous cell population for increasing levels of random variation on parameters: 5% , 10% , 25% and 50% ( see Fig . S4 ) . Simulations show that even with large random variation on every parameter of the model ( 25% or 50% ) , most cells in the population are characterized by a non-transformed state in normal conditions ( see Fig . S4G , S4J ) as well as in the presence of high levels of both Ras and PTEN ( Fig . S4I , S4L ) . In contrast , in each case , most cells in the population are defined by a transformed state in the presence of high levels of Ras ( Fig . S4H , S4K ) . This result indicates that the model is quite robust to random variations on parameter values . By resorting to the deterministic model for a heterogeneous cell population , we are also able to reproduce the dynamical behavior of NF-κB activation . Indeed , single-cell analysis revealed that the activation of NF-κB is heterogeneous and is a digital on-off process with fewer cells responding at lower doses of inflammatory signal [45] , [46] . The model accounts for this observation by showing that the expression levels of Let-7 and NF-κB are characterized by two main states: ( 1 ) high levels of Let-7 and low levels of NF-κB , or ( 2 ) low levels of Let-7 and high levels of NF-κB ( see Fig . S5 ) . The number of responding cells , defined by high levels of NF-κB and low levels of Let-7 increases by increasing the level of inflammatory signal , Src ( compare panels A to C in Fig . S5 ) . The microRNA miR-21 is overexpressed and promotes invasion in pancreatic ductal adenocarcinoma [48] . Moreover , it was shown experimentally that miR-21 is an activator in the regulatory feedback loop linking inflammation to cell transformation [20] . Indeed , STAT3 promotes the expression of miR-21 , which results in the down-regulation of the tumor suppressor PTEN leading to an activation of NF-κB ( see Fig . 1 ) . A transient expression of miR-21 can induce the epigenetic switch , which leads to cell transformation and formation of mammospheres in mouse xenografts [20] . The positive feedback loop between NF-κB , Lin28 , Let-7 , IL6 is crucial for this process of cellular transformation because the concomitant overexpression of miR-21 together with the inhibition of Lin28 considerably reduce cell transformation and mammospheres formation [20] . In the model , from a non-transformed state of the cell characterized by a high level of Let-7 and low levels of NF-κB , Lin28 , IL6 , and STAT3 , the overexpression of miR-21 triggers the switch to cell transformation , which results in a low level of Let-7 and high levels of NF-κB , Lin28 , IL6 , and STAT3 ( see Fig . 9A ) . The model qualitatively reproduces the experimental observation by showing that the concomitant overexpression of miR-21 together with a partial inhibition of Lin28 is not able to trigger the switch to cell transformation ( Fig . 9B ) . Indeed , only a small increase in the level of NF-κB , Lin28 , IL6 , and STAT3 and a small decrease in Let-7 are observed ( see Fig . 9B for t>100 h ) . Iliopoulos and coworkers also showed , in colon adenocarcinomas , a positive correlation in the expression pattern of miR-21 and STAT3 as well as a negative correlation in the expression pattern of miR-21 and PTEN [20] . By resorting to the deterministic model for a heterogeneous cell population , we can reproduce qualitatively these expression patterns ( see Fig . S6 ) . A low rate of synthesis of miR-21 generates a mixed population of non-transformed and transformed cells ( Fig . S6A , B ) . From that condition , an increase in the rate of synthesis of miR-21 switches the cell population to a transformed state characterized by high levels of miR-21 and STAT3 together with low levels of PTEN ( Fig . S6C , D ) . From the latter condition , a reduction in the level of Lin28 brings back a large proportion of cells to a non-transformed state defined by low levels of miR-21 and STAT3 together with high levels of PTEN ( Fig . S6E , F ) . This result supports the experimental observations showing the reduced number of mammospheres formation in the presence of Lin28 inhibition [20] . Recent hypothesis suggests that some messenger RNAs , competing endogenous RNAs ( ceRNA ) , could possess a regulatory role , independently of their protein-coding function , by their ability to compete for microRNA binding [26] , [27] , [29] . ceRNA could act as natural microRNA sponge [49] . In that context , by sharing the same microRNAs , it was shown that the expression of the tumor suppressor PTEN and its pseudogene PTENP1 are positively correlated . Indeed , even if the pseudogene PTENP1 does not encode a functional protein , PTENP1 mRNA may regulate the expression of the tumor suppressor PTEN by competing for their common microRNA [28] . It was also shown that the pseudogene PTENP1 is mutated in some cancers [28] . By resorting to our computational model linking inflammation to cell transformation , we will analyze here the regulatory role of a generic ceRNA , which could compete for the binding to Let-7 microRNA as well as the role of the pseudogene PTEN1 , which competes with PTEN mRNA for the binding to miR-21 ( see Methods for more details as well as the wiring diagram in Fig . 10 ) . Bifurcation analyzes of the steady-state levels of NF-κB , Lin28 , Let-7 , IL6 , STAT3 , and ceRNA vs Src for different rates of synthesis of ceRNA , VSCERNA , suggest that a competing-endogenous RNA for the binding to Let-7 microRNA could act as an oncogene by promoting the switch to cell transformation ( see Fig . S7 ) . Indeed , the model predicts that an increase in VSCERNA from 0 to 1 moves the switch from a non-transformed to a transformed state to smaller levels of the inflammatory signal , Src . Furthermore , the model also shows that if the level of ceRNA is too high , i . e . VSCERNA = 2 , the cell is always present in a transformed state defined by high levels of NF-κB , Lin28 , IL6 , STAT3 , and ceRNA together with a low level of Let-7 . By means of stochastic simulations , the model indicates that an elevated level of ceRNA decreases the robustness of the non-transformed state of the cell towards stochastic fluctuations ( see Fig . S8 ) . Indeed , by increasing VSCERNA , the proportion of transformed cells at low levels of Src is larger ( compare Fig . S8B , D , F ) . The deterministic model for a heterogeneous cell population shows that , in the absence of ceRNA , the expression levels of Let-7 and IL6 are negatively correlated with high levels of Let-7 together with low levels of IL6 , which corresponds to a non-transformed state of the cell population ( Fig . 11A ) . An increase in ceRNA progressively promotes the switch of the cell population to a transformed state characterized by low levels of Let-7 together with high levels of IL6 ( see Fig . 11B , C ) . Thus , the model predicts that the effect of a ceRNA for Let-7 binding is very similar to the effect of Ras oncogene on the dynamics of the epigenetic switch leading to cell transformation . Finally , the model indicates that high levels of PTEN1 mRNA delay or eventually suppress the occurrence of cell transformation ( see Fig . S9 where the time evolution of NF-κB , IL6 , miR-21 , and PTEN is represented for different rates of transcription of the pseudogene PTEN1 , VSMPTEN1 ) . This supports the experimental observations showing that PTEN ceRNAs exhibit tumor-suppressive properties [50] .
Rudolf Virchow made the first connection between chronic inflammation and cancer in 1863 [5] . Nowadays , molecular links between inflammation and oncogenic transformation have been established [6] . However , until recently , the molecular pathways linking inflammation to cellular transformation were unknown [7] . In the latter study , the authors showed that ER-Src oncoprotein by treatment with tamoxifen could convert a non-transformed cell line to a transformed state within 24–36 h . They also demonstrated that this cell transformation is mediated by an inflammatory positive feedback loop driven by NF-κB , Lin28 , Let-7 , and IL6 ( see [7] and Fig . 1 ) . Based on these experiments [7] , [20] , we proposed here a computational model to account for the dynamics of the positive inflammatory feedback loop leading to cell transformation . The model includes the regulations between NF-κB , Lin28 , Let-7 , IL6 , Ras , STAT3 , miR-21 , and PTEN ( see Fig . 1 ) . First the model accounts for the experimental observations showing that the non-transformed state , characterized by low levels of activators of the switch ( NF-κB , Lin28 , IL6 , and STAT3 ) together with high levels of inhibitors of the switch ( Let-7 and PTEN ) , is stable without the inflammatory signal , Src ( Fig . 2 ) . As in the experiments , a transient activation of Src for only 5 min triggers cell transformation within 60 h , while a constant level of Src may elicit the switch to a transformed state within 35 h ( see Fig . 2 ) . The model also accounts for the fact that maintenance of the transformed state needs the presence of the positive feedback loop between NF-κB , Lin28 , Let-7 , and IL6 . Indeed , a transient inhibition of NF-κB , Lin28 , or IL6 abolishes the transformed state of the cell ( Fig . S1 ) . In the model , each component is embedded at least in one positive feedback ( PF ) loop ( see wiring diagram in Fig . 1 ) . The main PF loop is based on the regulations between NF-κB , Lin28 , Let-7 and IL6 . The second PF loop rests on the interactions between NF-κB , Lin28 , Let-7 and Ras . The third PF loop is the mutual activation between NF-κB and IL6 , while the fourth PF loop is driven by the regulatory interactions between NF-κB , IL6 , STAT3 , miR-21 , and PTEN . Each component in the network can be clustered into one of the following two groups: activators ( oncogenes such as NF-κB , Lin28 , Ras , IL6 , STAT3 and miR-21 ) or inhibitors of the switch leading to cell transformation ( tumor suppressors such as Let-7 and PTEN ) . By resorting to bifurcation analyzes , we show that an irreversible bistable switch between a transformed and a non-transformed state is at the core of the dynamics of the epigenetic switch linking inflammation to cancer ( see Figs . 3 , 4 , S2 ) . The model suggests that Let-7 and PTEN act as tumor suppressors by increasing the threshold of inflammatory signal , Src , at which the switch to cell transformation occurs ( see Figs . 3 and S2 , respectively ) . On the opposite , the model shows that Ras exhibits oncogenic properties by reducing the threshold of Src at which cell transformation happens ( see Fig . 4 ) . Thus , at any moment , there is a sensitive balance in the relative levels of oncogenes and tumor suppressors that defines the state of the cell: non-transformed versus transformed . The importance of stochastic gene expression and stochastic transitions has been highlighted is many different biological contexts [51] , [52] . A study showed that a reduction of stochastic transitions could enhance cellular memory [53] . Stochastic mechanisms have been involved in the differentiation of mature subsets of T lymphocytes [54] . Stochastic gene fluctuations could drive the phenotype diversity in HIV-1 [55] , and stochastic epigenetic variation has been proposed to be a major force of development , evolutionary adaptation and disease [44] . Stochastic phenomena could also play an important role in cancer . Indeed , stochastic appearance of mammary tumors has been observed experimentally [56] . A model showed that the stochastic effect due to the finite size of active stem cell population could greatly influence the dynamics of cancer evolution [57] . Stochastic fluctuations may also control nonheritable cell variability , which could drive the evolutionary rate of cancer progression [58] . Finally , it was shown both theoretically and experimentally that stochastic state transitions generate phenotypic equilibrium in populations of cancer cells [59] . Here , we show that stochastic fluctuations may be responsible for transitions between a non-transformed and a transformed state of the cell ( Fig . 5 ) . Furthermore , the model predicts that the tumor suppressors Let-7 and PTEN increase the robustness of the non-transformed state of the cell towards stochastic fluctuations ( Figs . 6 and S3 , respectively ) , while the oncogene Ras decreases this robustness ( Fig . 7 ) . Thus , besides the crucial role of multiple and successive genetic mutations in the process of cell transformation and cancer development [1] , the model points out , based on experimental observations [7] , [20] , the important effect of epigenetic mechanisms and stochastic transitions on the dynamics of cell transformation . The positive inflammatory feedback loop described here seems relevant in human cancers . Indeed , cancer tissues have lower levels of Let-7 and higher levels of IL6 as compared to normal tissues , and an inverse relationship is also observed in the expression pattern of Let-7 and IL6 in prostate , breast , and hepatocellular tissues [7] . Moreover , it was shown that a related microRNA inflammatory feedback loop controls hepatocellular oncogenesis [60] . Here , deterministic simulations of the model in a heterogeneous cell population allow accounting qualitatively for these observations by showing this negative correlation in the expression of Let-7 and IL6 ( Fig . 8 ) . The model exhibits high levels of Let-7 and low levels of IL6 in normal conditions ( Fig . 8A ) . As suggested by the model , an increase in the level of the oncogene Ras , in the level of Lin28 or a decrease in Let-7 triggers the switch of the cell population to a transformed state ( Fig . 8B–E ) ; while , from a transformed state , an increase in the tumor suppressor PTEN allows recovering a normal , non-transformed , state of the cell population ( Fig . 8F ) . Here , we consider separately the two kinds of noises that can emerge in such dynamical system: ‘intrinsic’ , due to molecular noise , and ‘extrinsic’ , originating from cell-to-cell variability . Of course , in real , physiological conditions , these two kinds of noises combine within a cell population . Thus , a combination of both noises ( intrinsic and extrinsic ) may be a driving force to trigger , in a random manner , the occurrence of the switch leading to a transformed state of the cell . The latter source of noise resulting from cell-to-cell variability seems to have an important role leading to a heterogeneous activation of NF-κB under inflammatory signaling [45] . The model also accounts for the fact that miR-21 can be viewed as an oncogene . Moreover , the positive inflammatory feedback loop between NF-κB , Lin28 , Let-7 , and IL6 is important for the oncogenic property of miR-21 [20] . Indeed , we show , as in the experiments , that an overexpression of miR-21 elicits cell transformation , while a concomitant overexpression of miR-21 together with an inhibition of Lin28 greatly impede this transformation ( Figs . 9 and S6 ) . It was recently hypothesized that messenger RNAs , transcribed pseudogenes , and long non-coding RNAs may interact with each other using microRNA response element , MREs [27] . It is suggested that those competing endogenous RNAs ( ceRNAs ) form a large regulatory network , which could play important roles in normal and pathological conditions , such as cancer or cell differentiation [27] , [61] , [62] . By using our computational model for the epigenetic switch linking inflammation to cancer , we analyzed the dynamical consequences on cell transformation of the addition of a ceRNA competing for the binding to Let-7 microRNA as well as the addition of the transcribed pseudogene PTEN1 ( see wiring diagram in Fig . 10 ) . The model shows that increasing the level of a ceRNA competing for Let-7 binding reduces the threshold of Src at which the switch to cell transformation occurs ( Fig . S7 ) . In the presence of ceRNA , the non-transformed state of the cell is also less robust towards stochastic fluctuations ( Fig . S8 ) . The dynamics of a heterogeneous cell population predicts that an elevated level of ceRNA leads to the switch of the cell population to a transformed state ( Fig . 11 ) . Thus , the model suggests that a ceRNA competing for Let-7 binding will sponge the available Let-7 microRNA , which promotes cell transformation . Such ceRNA may thus be viewed as an oncogene . On the opposite , the model suggests that the addition of a ceRNA such as PTEN1 mRNA , competing with PTEN mRNA for binding to miR-21 microRNA , could act as a tumor suppressor inhibiting the occurrence of cell transformation ( Fig . S9 ) . This result holds with experimental observations showing the potential role of PTEN1 mRNA , or of other ceRNAs for PTEN , as tumor suppressors [26] , [28] , [50] . In summary , the model proposed here brings to light a comprehensive qualitative picture of the dynamics of the epigenetic switch linking inflammation to cell transformation and cancer . The model predicts that bistability is at the core of the underlying mechanism driving the switch between a non-transformed and a transformed state of the cell . Activators of the switch ( oncogenes ) and inhibitors of the switch ( tumor suppressors ) regulate the occurrence of cell transformation by modulating the threshold of inflammatory signal ( Src ) at which the switch occurs . The model also suggests that stochastic fluctuations could be a driving force for cell transformation , and predicts that tumors suppressors and oncogenes render the non-transformed state of the cell respectively more or less robust towards stochastic fluctuations . Finally , depending on their microRNA targets , the model shows that ceRNAs may act as oncogenes or as tumor suppressors , which points out the potential role of ceRNAs in the regulation of cell transformation .
In the model , we consider for simplicity a constant total concentration of NF-κB , NFKBT . Activation and inactivation reactions of NF-κB behave as Goldbeter-Koshland switches [64] . All other processes of the model rest on mass-action kinetics . ( 1 ) ( 2 ) ( 3 ) ( 4 ) ( 5 ) ( 6 ) ( 7 ) ( 8 ) ( 9 ) ( 10 ) ( 11 ) ( 12 ) ( 13 ) ( 14 ) With the following conservation equation:The ‘ode’ code of the deterministic model can be found in Supporting Information ( Model S1 ) . Such code can be run quite easily with the program XPP/XPPAUT ( http://www . math . pitt . edu/~bard/xpp/xpp . html ) . To account for the effect of a competing endogenous RNA , ceRNA , binding to Let-7 microRNA on the dynamics of the epigenetic switch , we add two kinetic equations describing the time evolution of ceRNA ( see Eq . ( 15 ) ) as well as the time evolution of the inhibitory complex between ceRNA and Let-7 , ceRNAlet7 ( see Eq . ( 16 ) ) . Moreover , we include the terms of association and dissociation of Let-7 to ceRNA in the kinetic equation describing the time evolution of Let-7 in Eq . ( 3 ) ( see new Eq . ( 3′ ) ) . ( 15 ) ( 16 ) ( 3’ ) In a similar manner , to account for the effect of PTEN1 mRNA , mPTEN1 , on the dynamics of the epigenetic switch , we add two kinetic equations describing the time evolution of mPTEN1 ( see Eq . ( 17 ) ) as well as the time evolution of the inhibitory complex between mPTEN1 and the microRNA miR-21 , miRmPTEN1 ( see Eq . ( 18 ) . Moreover , we include the terms of association and dissociation of miR-21 to mPTEN1 in the kinetic equation describing the time evolution of miR-21 in Eq . ( 11 ) ( see new Eq . ( 11′ ) ) . ( 17 ) ( 18 ) ( 11’ ) In the simulations of Figs . 2 and S1 , with the parameter values of Table 2 and with Src = 0 , the following initial conditions have been chosen so as to reach a non-transformed state of the cell characterized by high levels of Let7 and PTEN together with low levels of NF-κB , Lin28 , IL6 and STAT3 . NFKB = 0 . 01; Lin28 = 0 . 01; Let7 = 1; mIL6 = 0 . 01; IL6 = 0 . 01; mRas = 0 . 01; Ras = 0 . 01; STAT3 = 0 . 01; mPTEN = 0 . 01; mRasLet7 = 0; mIL6Let7 = 0; miR21 = 0; PTEN = 0; MiRmPTEN = 0 . For the same conditions on the parameters , the following initial conditions have been chosen in all the other figures , giving rise to a switch at about t = 20 h . NFKB = 0 . 5; Lin28 = 0 . 5; Let7 = 1; mIL6 = 0 . 5; IL6 = 0 . 5; mRas = 0 . 5; Ras = 0 . 5; STAT3 = 0 . 5; mPTEN = 0 . 01; mRasLet7 = 0; mIL6Let7 = 0; miR21 = 0; PTEN = 0; MiRmPTEN = 0; CeRNA = 0; CeRNALet7 = 0; mPTEN1 = 0; MiRmPTEN1 = 0 . Because the dynamics of the model for the epigenetic switch linking inflammation to cell transformation rests on an irreversible bistable switch , such dynamics will be sensitive to the value of the initial conditions of the different variables defining the system . For the same parameter values , depending on the values of the initial conditions , the system may tend to a transformed or a non-transformed state of the cell ( not shown ) . | An increasing amount of evidence demonstrates a close relation between inflammation and cancer development , which reveals the importance of the tumor microenvironment for the development of cancers . Recently , a molecular pathway linking inflammation to cell transformation , which is a prerequisite to cancer development , has been discovered . This molecular pathway is based on a positive inflammatory feedback loop between NF-κB , Lin28 , Let-7 microRNA and IL6 , allowing the occurrence of an epigenetic switch leading to cell transformation . Here , we propose a computational model to account for the dynamics of this epigenetic switch . We show that an irreversible bistable switch is at the core of the dynamics of the system . The model further indicates that oncogenes ( activators of the switch ) and tumor suppressors ( inhibitors of the switch ) regulate the occurrence of cell transformation by modulating the threshold of inflammatory signal needed to induce the switch . Stochastic simulations of the model suggest that molecular fluctuations are able to trigger cell transformation , highlighting possible links between stochasticity and cancer development . Finally , the model predicts a crucial role of competing endogenous RNAs ( ceRNAs ) for the dynamics of the epigenetic switch and the occurrence of cell transformation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"complex",
"systems",
"systems",
"biology",
"nonlinear",
"dynamics",
"signal",
"transduction",
"stat",
"signaling",
"family",
"signaling",
"in",
"cellular",
"processes",
"mathematics",
"theoretical",
"biology",
"oncogenic",
"signaling",
"applied",
"mathematics",
"ras",
... | 2014 | A Model for the Epigenetic Switch Linking Inflammation to Cell Transformation: Deterministic and Stochastic Approaches |
Since leprosy is both treated and controlled by multidrug therapy ( MDT ) it is important to monitor recurrent cases for drug resistance and to distinguish between relapse and reinfection as a means of assessing therapeutic efficacy . All three objectives can be reached with single nucleotide resolution using next generation sequencing and bioinformatics analysis of Mycobacterium leprae DNA present in human skin . DNA was isolated by means of optimized extraction and enrichment methods from samples from three recurrent cases in leprosy patients participating in an open-label , randomized , controlled clinical trial of uniform MDT in Brazil ( U-MDT/CT-BR ) . Genome-wide sequencing of M . leprae was performed and the resultant sequence assemblies analyzed in silico . In all three cases , no mutations responsible for resistance to rifampicin , dapsone and ofloxacin were found , thus eliminating drug resistance as a possible cause of disease recurrence . However , sequence differences were detected between the strains from the first and second disease episodes in all three patients . In one case , clear evidence was obtained for reinfection with an unrelated strain whereas in the other two cases , relapse appeared more probable . This is the first report of using M . leprae whole genome sequencing to reveal that treated and cured leprosy patients who remain in endemic areas can be reinfected by another strain . Next generation sequencing can be applied reliably to M . leprae DNA extracted from biopsies to discriminate between cases of relapse and reinfection , thereby providing a powerful tool for evaluating different outcomes of therapeutic regimens and for following disease transmission .
Leprosy is a complex dermato-neurologic and systemic disease[1] primarily caused by Mycobacterium leprae or to a much lesser extent by Mycobacterium lepromatosis . [2] Despite a strong decrease in leprosy prevalence since the systematic implementation of multidrug therapy ( MDT ) in the 1980’s , the incidence of disease , the major indicator of active transmission , remains high in many countries , especially in India and Brazil , showing that transmission continues unabated . [3] Overall , more than 200 , 000 new leprosy cases are reported each year worldwide . [3] The MDT regimen for leprosy consists of different antibiotic combinations that are prescribed based on the number of skin lesions: a six-month regimen of rifampicin and dapsone for paucibacillary ( PB ) patients ( <5 skin lesions ) and a twelve month regimen of rifampicin , dapsone and clofazimine for multibacillary ( MB ) patients ( >5 skin lesions ) . [4] In 2002 , WHO proposed that a uniform MDT regimen ( U-MDT ) should be considered to treat all types of leprosy in order to facilitate leprosy control . In 2007 , an open-label randomized and controlled clinical trial ( uniform multidrug therapy for leprosy patients in Brazil , U-MDT/CT-BR ) was initiated to compare U-MDT with the regular MDT for PB and MB patients . [5 , 6] Clinical monitoring is still taking place with special emphasis on disease recurrence and leprosy type 1 and type 2 reactions ( T1R/T2R ) . An increased relapse rate and the possible emergence of drug resistance are major concerns for the shortened MDT proposal for MB patients . It is therefore important to address this issue by analyzing in depth all recurrent cases from the U-MDT/CT-BR trial . Molecular genotyping techniques , such as typing selected single nucleotide polymorphisms ( SNP ) or counting variable number tandem repeats ( VNTR ) have been used to differentiate reinfection from relapse . [7–11] However , the resolution of such techniques is often limited because of the exceptional level of genome conservation in M . leprae and the limited sequence diversity between strains from the same geographical area in particular . [12] In contrast , genome-wide approaches provide higher resolution and accuracy compared to genotyping based on a predefined set of loci , but are technically more complex . High throughput sequencing is becoming increasingly efficient and cost-effective with purified DNA but is more challenging with clinical specimens such as DNA extracted directly from skin biopsies , especially from formalin-fixed paraffin-embedded ( FFPE ) samples . In this study , we investigated three recurrent cases of leprosy from the U-MDT/CT-BR trial to determine whether recurrence was due to drug resistance , bacterial persistence or to reinfection . To achieve this , we compared whole genome sequencing analysis of M . leprae collected from skin lesions at the initial diagnosis and during the recurrence of the disease and correlated the sequence data with the clinical , microbiologic and serologic findings .
This study was approved by the regional research ethical committees , by the National Committee for Ethics in Research ( CONEP , National Health Council/ Ministry of Health , Brazil , protocol # 001/06 ) and by the human and animal research ethics committee from the Federal University of Goiás ( CEMHA/HC/UFG protocol # 166/2011 ) . Written informed consent was obtained from all adult subjects and a parent or guardian of participants under the age of 18 years , provided informed consent on their behalf prior to inclusion in the study ( ClinicalTrials . gov identifier: NCT00669643 ) . Three recurrent cases of leprosy identified in the U-MDT/CT-BR trial were investigated ( Table 1 ) . Clinical diagnosis and monitoring were carried out at the National Reference Canter in Ceará state , Northeast Brazil . Leprosy diagnosis was confirmed by bacteriological analysis of slit skin smears and by histopathological examination of biopsies taken from active skin lesions . [6] At the first visit , patients had a complete dermato-neurological examination by a dermatologist with expertise in leprosy diagnosis , when the number and the body distribution of skin lesions and affected nerves were registered . Biopsy of skin lesion , venous blood and skin smear material from six sites for bacilloscopy were collected . During the clinical monitoring , patients attended the established schedule for clinical/laboratory monitoring ( monthly appointment during the first year and thereafter , yearly ) . All patients were advised to return to an urgent appointment at the reference center in case any discomfort or new clinical manifestation appeared . In this study , the following case definitions for leprosy reactions were employed: T1R was defined as an acute clinical manifestation , usually characterized by the exacerbation of pre-existing lesions , or the appearance of new lesions . T2R was characterized by the sudden appearance of tender erythematous skin nodules ( erythema nodosum leprosum/ENL ) mainly accompanied by fever and other systemic symptoms such as joint pain , bone tenderness , neuritis , edema , malaise , anorexia with or without lymphadenopathy . In the clinical diagnosis of reactions , skin signs were obligatory , nerve and systemic signs were non-compulsory while neuritis , malaise , and fever could be present in both types of reaction . Treatment for leprosy reactions followed the guidelines from the Brazilian Ministry of Health . Patients with clinical manifestations not fulfilling these previously described criteria were considered suspected cases of relapses and were clinically examined by the assistant dermatologist , by the PI ( GOP ) and by an expert member of the independent steering committee ( Dr . Sinesio Talhari ) . Additionally , in these patients skin smears and biopsies were collected from new lesions and used to investigate drug susceptibility ( inoculation in BALB/c mice , sequencing of the rpoB , folP1 , gyrA and gyrB genes and whole genome sequencing ) . As part of the U-MDT/CT-BR trial , a well-prepared biobank of biopsies from leprosy skin lesions and serum samples , collected at various time points during treatment and monitoring , was assembled and has been properly maintained at recruitment sites and an extra back-up has been kept at the coordination center . For this study , we used skin biopsies from the first episode that were formalin-fixed and paraffin-embedded to allow long-term storage and serum samples collected at diagnosis and at various time-points during and after treatment ( Table 1 and S1 Table ) . Serum IgM antibodies to M . leprae-specific PGL-1 antigen ( 0 . 01μg/mL NT-P-BSA ) and serum IgG antibodies to the synthetic LID-1 ( 1μg/mL LID-1 ) antigen were detected by enzyme-linked immunosorbent assay ( ELISA ) . [13 , 14] Patients showing recurrent symptoms after treatment had biopsies taken from new lesions ( Table 1 and S1 Table ) , which were used as the source of M . leprae for drug susceptibility testing in BALB/c mice [15] ( treated with dapsone , rifampicin or no drug ) and for partial [16] and whole genome sequencing . A truXTRACTM FFPE DNA kit ( Covaris ) was used following the manufacturer's recommendation with some optimization . Briefly , ten 20μm FFPE tissue sections for each sample were pooled in a screw-cap microTUBE in duplicate or triplicate . Paraffin was removed and the tissue rehydrated with 100μl of tissue SDS buffer using a focused-ultrasonicator series S2 with the following settings: intensity = 5 , cycles per burst = 200 , time = 300s , temperature = 20°C . Digestion was done using a 40μl mixture of proteinase K ( 20 mg/ml ) and lysozyme ( 10 mg/ml ) using a focused-ultrasonicator with the same settings as above except for the time set at 10s . Digestion occurred at 56°C overnight followed by 1 h at 80°C to reverse the formaldehyde crosslinks . Finally , DNA was isolated from lysates using the columns of the truXTRAC FFPE DNA kit and eluted in 50μl of Covaris BE buffer . DNA was quantified using a Qubit fluorometer ( ThermoFisher ) . For samples 1126–2011 and 2188–2014 , which had been passaged in mice , DNA was extracted from mouse footpad suspensions then sheared to ~600 bp by ultrasonication and purified with AMPure beads , before library preparation . The quantity of DNA was assessed after each critical step i . e . DNA extraction , library preparation and amplification post-array capture ( S2 Table ) . Since the quality of DNA is known to be low after FFPE extraction , we did not fragment the DNA with the Covaris method as it was already fragmented nor did we size select our libraries to avoid losing too much DNA . DNA from each extract was used to prepare Illumina libraries using a Kapa Hyper Prep kit ( Kapa Biosystem ) as described elsewhere . [17] To remove host DNA from the libraries , we used a custom-synthesized oligonucleotide array ( Agilent ) spanning the entire M . leprae genome . [18] Quality of the captured and re-amplified library was assessed using the Fragment Analyzer system ( Advances Analytical technologies , Inc ) . The size of the captured library was 180bp and the concentration 52ng/μl . Sequencing was performed on an Illumina Hi-Seq 2500 instrument . Raw reads from the same sample were merged and processed as described elsewhere [17] by adapter- and quality-trimming and alignment with the M . leprae TN reference genome ( NCBI a . n . AL450380 . 1 ) . To avoid false positive SNP calls the following cutoffs were applied: minimum overall coverage of 5 non-duplicated reads , minimum of 3 non-duplicated reads supporting the SNP , mapping quality score greater than 8 , base quality score greater than 15 and a SNP frequency above 80% . SNPs and short insertions and deletions ( InDels ) were compared between index and second episodes for each recurrent case . Unique sets of SNPs for each genome were established by comparison with the list of SNPs from 20 M . leprae genomes published elsewhere ( S3 Table ) . [9 , 18 , 19] All unique and/or discriminatory variants were manually visualized using the IGV browser [20] to check for possible alignment inconsistencies . We additionally genotyped all samples using the SNP model described in Monot et al . and inferred in silico the VNTR copy number for 33 out of 44 known VNTR loci ( 11 loci were too large to be spanned with Illumina reads ) . [9 , 11 , 21]
The U-MDT/CT-BR study initially enrolled 858 patients of whom 78 . 4% were classified as MB . During follow-up , four of the treated patients presented with new symptoms between four and eight years after completion of U-MDT and three of these were re-investigated in this study . These participants were three young male leprosy patients ( # 1126 , 2188 and 3208 ) from Fortaleza , Ceará , Northeast Brazil , an endemic city for leprosy . The main clinical and laboratory characteristics of these three patients with recurrent signs of leprosy after U-MDT are shown in Table 1 . In all three cases , leprosy was first diagnosed in 2007 but the patients displayed new clinical signs , which were not associated with leprosy reactions , between 2011 and 2015 . In these three patients , original leprosy skin lesions detected at diagnosis , disappeared after specific treatment and upon suspicion of relapse/reinfection , new skin lesions were observed in previously unaffected body areas . The timelines of clinical events presented by these patients during follow up ( S1 Fig ) highlight their high propensity to develop leprosy reactions , especially T2R , although all of them also developed T1R . These records also demonstrate that leprosy reactions and relapse/reinfection occurred at different time points . The timelines also illustrate the evolution of bacilloscopic index ( BI ) during follow up . In one case , the BI at the second episode was higher than the BI at the first episode . In addition , the first diagnosis revealed that the three MB patients showed high IgM and IgG antibody levels to PGL-1 and LID-1 antigens , respectively ( S1 Fig ) . Since these biomarkers have been used to monitor the disease state , we measured antibody levels by ELISA before , during and after U-MDT . Overall , the antibody titers gradually declined but remained above the threshold for positivity for at least one of the antigens during the study period except for patient 1126 . This patient showed an antibody titer below the threshold just before the recurrence of the disease ( 39 months after U-MDT ) and then both antibody titers increased by the time of recurrent disease . By contrast , despite oscillating levels of PGL-1 antibody for 3208 , antibody titers , especially to LID-1 , remained high for 3208 and 2188 during the entire study period . M . leprae from the recurrent lesions ( 1126–2011 , 2188–2014 ) was inoculated into mice and only multiplied in the untreated animals , indicating that the bacilli were viable but susceptible to dapsone and rifampicin . It was not possible to inoculate mice with the sample from 3208–2015 . Analysis of the rpoB , folP1 , gyrA and gyrB genes revealed a wild-type sequence in all six strains , confirming susceptibility to rifampicin , dapsone , and fluoroquinolones , respectively , in all cases . Sufficient whole genome read coverage was obtained from the six M . leprae samples for genotyping and comparative genomic analyses ( S4 Table ) . The recurrent strain 1126–2011 was clearly distinct from the primary strain 1126–2007 , and differed in 44 SNPs , 4 InDels and 6 VNTR loci ( Fig 1 and S5 Table ) . Furthermore , 1126–2007 and 1126–2011 share no SNPs that might indicate close relatedness or direct ancestry . Strains 3208–2007 and 3208–2015 differed in only two SNPs and one VNTR locus ( Fig 1 and S5 Table ) . Both SNPs ( T1740863C in an intergenic region and C1803024T in a pseudogene ) were present in 3208–2015 , indicating that 3208–2015 was certainly the direct progeny of 3208–2007 . In addition , eight unique variant nucleotides were restricted to these two samples ( compared to the SNPs from 20 previously published M . leprae genomes [9 , 18 , 19] and those from this study ) , confirming the identity of the strains ( S6 Table ) . Interestingly , a cluster of three SNPs leads to missense mutations , in codons 495 and 496 of asn1 , encoding an L-asparagine permease , which contributes to virulence in Mycobacterium tuberculosis [22] . Analysis of 2188–2007 and 2188–2014 revealed identical genome sequences ( Fig 1 ) . Curiously , both strains belong to a new SNP subtype intermediate between subtypes 4N and 4O . The only difference between the two genomes was found in the ( GTA ) 9 VNTR locus ( S5 Table ) , which harbored 11 repeats in 2188–2007 and 12 repeats in 2188–2014 . Genome comparisons revealed that both strains share 28 unique variant nucleotides ( S7 Table ) . Among them are two missense mutations in ML0411 , encoding a PPE protein and in ribD ( ML1340 ) , the riboflavin biosynthesis protein . An insertion of 9 nucleotides ( GGACATCTA at position 1 , 219 , 061 ) was found in ML1052 , a putative PucR-like transcriptional regulator , which leads to a modification of the protein . Interestingly this mutation was present at only 30% frequency in 2188–2007 , while it was fixed in 2188–2014 . Another frame-shift arising from a dinucleotide insertion was found in ML0825c , the ortholog of rv2358 in M . tuberculosis that codes for the protein SmtB , a zinc-sensing transcriptional regulator and member of the AsrR/SmtB family . [23 , 24] . The C-terminal part of SmtB is essential for the protein dimerization , zinc binding and DNA recognition . Furthermore , a specific histidine residue ( H138 in ML0825c and H117 in Synechococcus StmB ) is important for the allosteric coupling of the zinc and DNA binding sites in the protein . [25] Modeling of M . leprae StmB in silico ( S2 Fig ) showed that the frame-shift leads to loss of H138 and should thus impair protein function .
The relapse rate is considered to be the most important indicator of the efficacy of a given MDT . On the other hand , reinfection is an indicator of active transmission and the susceptibility of leprosy convalescents to new infections . This investigation provided a unique opportunity to apply high-resolution whole-genome tools to differentiate relapse from reinfection and to evaluate the impact of U-MDT on antibody levels to two M . leprae antigens . MDT affects both cellular and humoral M . leprae specific immunity . In MB patients , there is a decline in antibody levels during MDT and patients remain unable to mount a protective Th1 type immunity to M . leprae after treatment . [26] Levels of antibodies to PGL-1 and LID-1 were high in all three MB cases at diagnosis , then declined during and after treatment but nonetheless remained above the cut-off point for positivity , especially antibodies to LID-1 . Our data is in accordance with previous studies showing decay in antibody titers while sero-reversion is rare in leprosy patients after regular MDT [26 , 27] . By the time of recurrent disease , the antibody titers to at least one of the antigens had risen . In our study , patients were carefully monitored for treatment compliance and all completed the U-MDT treatment . In our study the three MB patients had several episodes of leprosy reactions during follow up including T1R and mainly T2R , in accordance with the reports showing increased propensity of MB patients to develop reactions . [28 , 29] In fact , several studies have shown that in some endemic areas the occurrence of T1R in BL/LL patients is higher than T2R . A study about risk factors for leprosy reactions in patients from three endemic countries ( Philippines , Nepal , Brazil ) showed that among all LL and BL patients , T1R was more frequent than T2R . [30] Another study from Thailand showed that T2R was slightly more prevalent than T1R in lepromatous patients . [31] T1R primarily affects immunologically unstable borderline patients ( BL , BT , BB ) , while although sporadic , it also occurs in LL patients . T1R is characterized by an increased inflammatory Th1-type cell-mediated immunity in pre-existing skin lesions and systemically , in serum and in circulating leukocytes . The capacity of BL/LL patients , who have a predominant Th2 response , to develop T1R was elucidated by studies showing leukocytes with a Th0 profile that produce IFNγ , IL2 and IL4 or a polarized shift to Th1 type response with IFNγ and IL-12p40 mRNA in lesional skin and in leukocytes . [32 , 33] In both leprosy and tuberculosis , host genetic factors and immunological mechanisms determine the outcome of infection so that susceptibility varies among individuals . Case 1126 was unambiguously identified as reinfection because of the extensive polymorphisms between the two strains . Reinfection has long been suspected as a cause of new leprosy episodes and it has been suggested that individuals who have already had leprosy are more likely to be reinfected after treatment due to their inherent immunogenetic susceptibility . [34–36] Around 30% of relapse cases in Recife , northeast Brazil , were reported to be in contact with other leprosy patients and more often from the same family or household . [8] Leprosy case 1126 is an example of “family disease” , because both of the patient’s parents had leprosy around five years before his diagnosis , his daughter and partner had PB leprosy and the partner’s cousin , who lives in the same household , was diagnosed with MB leprosy but failed to complete MDT due to alcohol addiction . The extremely limited genomic variability detected between strains from the same geographical origin poses a challenge in distinguishing between relapse or reinfection with a closely related strain . In a recent paper , Avanzi et al . showed that a strain infecting three patients in the same region of Guinea Conakry differed in only two SNPs [17] and four VNTRs . In our study , two SNPs and one polymorphic VNTR were found . While individual VNTRs carry virtually no ancestral information due to the risk of homoplasy and mutation reversion , the fact that only two SNPs were found strongly indicates that the recurrent strain was directly derived from the original infection . It should be recalled that in our study skin biopsies were taken from two different lesions in different body areas . Likewise , in the case of 2188 only one polymorphic VNTR locus distinguished between the first and the recurrent infection , and the absence of SNPs confirms the strain’s identity . Furthermore , there was no history of leprosy in either patient 3208’s or 2188’s households , and both patients had high antibody titers during the study period suggesting continued immunological stimulation by bacterial antigens after treatment . Therefore , based on the genomic analysis , the patients’ epidemiologic history and serological data , we consider that the recurrence of leprosy in both patients 3208 and 2188 was due to relapse . Leprosy presents a variable incubation period which can range from 2–15 years . Although more prevalent in adults , leprosy also occurs in children <15 years , with reports of cases in patients younger than 1 year of age [37] indicating at least in children , short incubation period of the disease . However , nothing is known about the incubation period of reinfection , especially in genetically susceptible individuals who remain exposed to the bacilli in endemic areas . In this study , the reinfection case was observed 4 years after the conclusion of treatment , indicating a relatively short incubation period but which is in accordance with the reported range of the incubation period of the disease . The availability and larger use of whole genome sequencing studies of M . leprae in recurrent leprosy and leprosy reinfection can clarify the duration of incubation period in such cases . Further investigations of other such cases will give us a more definitive picture of characteristics of reinfection . It is theoretically possible that the original infection in leprosy could involve more than one strain of M . leprae and , that the recurrence could be a relapse due the regrowth of one of the sub-populations of M . leprae , that had been under-treated by the first course of MDT . However , although possible , in our study this probability was implausible , since in all three patients investigated , including the reinfection case , genomic sequences of the M . leprae strains responsible for the original infections showed no mutation associated with drug resistance . Therefore , even if the original infection had involved more than one strain of M . leprae , these strains were MDT susceptible . To conclude , this study is the first to demonstrate that it is possible to differentiate reinfection from relapse in leprosy in a field setting with a follow up period extended to eight years . This provides a proof-of-concept and emphasizes the value of whole genome sequencing in clinical follow up of leprosy . Importantly , the extended observation period allowed identification of relapses/reinfection . M . leprae grows very slowly and has a relatively long incubation time , so shorter periods of monitoring would be unlikely to provide sufficient clinical evidence to suspect relapse or reinfection . Also the two relapse cases in this study exemplify the superiority of whole-genome sequencing over genotyping a limited subset of loci or VNTR typing . For instance , the current SNP genotyping scheme can only detect distinct M . leprae lineages [9] , which is not useful for analyzing closely related strains . VNTRs can distinguish such strains but do not reflect the overall genetic distance ( Fig 1 ) nor convey information about strain ancestry . Improvements in sample preparation have made whole-genome sequencing more applicable routinely and we expect that recent technological advances will culminate in sequencing platforms that can be used to deliver whole genome coverage at the point of diagnosis within days of seeing the patient . [38] All raw sequence read files have been deposited in the trace archive of the National Center for Biotechnology Information Sequence Read Archive under accession no . SRP078228 . | Leprosy , one of the most ancient human infectious diseases , affects skin and nerves and is caused by Mycobacterium leprae infection . Despite the effective use of multidrug therapy/MDT since the 80´s , over 200 , 000 new cases are reported yearly , indicating active transmission , especially in India and Brazil . Although rare , recurrent clinical manifestations after MDT can occur due to leprosy reactions , relapse by drug resistance , insufficient treatment or reinfection . Relapse and reinfection cannot be differentiated clinically and molecular genotyping of a predefined set of loci have limited resolution due to exceptional M . leprae genome conservation and low sequence diversity between strains from the same geographical area . This is the first report that has compared whole-genome sequences of M . leprae strains from original and recurrent leprosy episodes . M . leprae genome differences were detected between the strains from the first and second episodes in the three patients . In one patient , there was clear evidence for reinfection with an unrelated strain whereas the other two were considered true relapses due to minor strain differences . No known drug resistance mutations were detected , excluding drug resistance as the recurrence cause . Next generation sequencing of M . leprae DNA discriminates relapse from reinfection representing a powerful tool for evaluating different disease outcomes and transmission . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"sequencing",
"techniques",
"mycobacterium",
"leprae",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"biopsy",
"tropical",
"diseases",
"surgical",
"and",
"invasive",
"medical",
"procedures",
"genome",
"sequencing",
"bacterial",
"... | 2017 | Whole genome sequencing distinguishes between relapse and reinfection in recurrent leprosy cases |
Polycomb group ( PcG ) proteins bind to and repress genes in embryonic stem cells through lineage commitment to the terminal differentiated state . PcG repressed genes are commonly characterized by the presence of the epigenetic histone mark H3K27me3 , catalyzed by the Polycomb repressive complex 2 . Here , we present in vivo evidence for a previously unrecognized plasticity of PcG-repressed genes in terminally differentiated brain neurons of parkisonian mice . We show that acute administration of the dopamine precursor , L-DOPA , induces a remarkable increase in H3K27me3S28 phosphorylation . The induction of the H3K27me3S28p histone mark specifically occurs in medium spiny neurons expressing dopamine D1 receptors and is dependent on Msk1 kinase activity and DARPP-32-mediated inhibition of protein phosphatase-1 . Chromatin immunoprecipitation ( ChIP ) experiments showed that increased H3K27me3S28p was accompanied by reduced PcG binding to regulatory regions of genes . An analysis of the genome wide distribution of L-DOPA-induced H3K27me3S28 phosphorylation by ChIP sequencing ( ChIP-seq ) in combination with expression analysis by RNA-sequencing ( RNA-seq ) showed that the induction of H3K27me3S28p correlated with increased expression of a subset of PcG repressed genes . We found that induction of H3K27me3S28p persisted during chronic L-DOPA administration to parkisonian mice and correlated with aberrant gene expression . We propose that dopaminergic transmission can activate PcG repressed genes in the adult brain and thereby contribute to long-term maladaptive responses including the motor complications , or dyskinesia , caused by prolonged administration of L-DOPA in Parkinson's disease .
An emerging concept in neurobiology is that many mechanisms implicated in chromatin remodeling and developmental processes retain their plasticity in the adult brain . Indeed , a number of environmental stimuli are known to generate chromatin modifications that have been causally linked to synaptic plasticity and associated behavioral and pathological responses . In this context , core histone modifications [1] have been implicated in cognitive functions , as well as in psychiatric conditions [1] , [2] . Polycomb group ( PcG ) proteins maintain cell type specific gene repression that is established during early embryonic development by regulating chromatin structure [3] . The Polycomb repressive complex 1 ( PRC1 ) mediates histone H2A lysine 119 mono-ubiquitination ( H2AK119ub ) , while PRC2 di- and tri-methylates histone H3 lysine 27 ( H3K27me2/3 ) [4] , [5] . Functionally , both PRC1 and PRC2 can be recruited to genomic regions through direct binding to H3K27me3 marked chromatin . Importantly , while dysregulation of PcG binding to target genes has been implicated in serious developmental defects and diseases such as cancer [6] , [7] , aberrant derepression of PcG target genes have not been associated with pathology of terminally differentiated neurons [2] . Parkinson's disease ( PD ) is caused by the death of midbrain neurons producing dopamine . This disorder is commonly treated with L-DOPA , which upon conversion to dopamine , relieves the motor symptoms of PD [8] . However , prolonged use of L-DOPA results in the emergence of dyskinesia , involving dystonic and choreic movements [9] . Several lines of evidence indicate that L-DOPA-induced dyskinesia ( LID ) is caused by abnormal activation of dopamine D1 receptors ( D1Rs ) located on the medium spiny neurons ( MSNs ) of the striatum [10] , [11] . This leads to increased gene expression through sequential activation of PKA , dopamine- and cAMP-regulated phosphoprotein of 32 kDa ( DARPP-32 ) , extracellular signal-regulated kinases ( Erk ) , mitogen- and stress-activated kinase 1 ( Msk1 ) and eventually phosphorylation of histone H3 at serine 10 ( H3S10p ) [12]–[15] . While the regulation of histone H3S10 phosphorylation has been studied in the adult brain [16] , [17] , almost nothing is known regarding H3S28 phosphorylation in neurons . However , in non-proliferating human fibroblasts it has been shown that H3K27me3S28 phosphorylation in response to MSK activation can lead to transcription of otherwise stably repressed genes [18] . The initial derepression is caused by displacement of gene repressor complexes containing PcG proteins , followed by transcriptional activation . In this study , we describe an important link between dopamine signaling , H3K27me3S28 phosphorylation , and aberrant gene expression associated to reduced PcG binding . Using a mouse model of PD , we show that dopamine via D1Rs increases H3K27me3S28 phosphorylation in striatal MSNs via two pathways: 1 ) activation of Msk1 , leading to phosphorylation of H3K27me3S28 and 2 ) activation of DARPP-32 leading to protein phosphatase 1 ( PP1 ) inhibition and suppression of H3K27me3S28p dephosphorylation . The combined effect is an accumulation of H3K27me3S28p at gene promoters that reduces PcG binding and allows transcription of a subset of genes . The results reveal a previously unrecognized plasticity of PcG-repressed genes in the adult brain , which upon environmental changes can be aberrantly induced via Erk-Msk1 mediated H3K27me3S28 phosphorylation and PKA-DARPP-32-dependent modulation of PP1 activity towards the same histone mark .
The ability of L-DOPA to activate the Erk-Msk1 pathway in striatal MSNs [15] , [19] , [20] , led us to hypothesize that signaling through dopamine receptors would induce phosphorylation of S28 in the context of H3K27me3 marked genomic sites to generate the H3K27me3S28p double histone modification . To test this possibility we turned to an experimental mouse model of PD in which unilateral stereotaxic injection of the neurotoxin 6-OHDA results in the elimination of the dopaminergic innervation to the basal ganglia ( Figure 1A ) [20] , [21] . In the lesioned striatum , MSNs react to the loss of dopamine by developing a remarkable sensitization to D1R agonists and , upon L-DOPA treatment strongly activate Msk1 , while the MSNs in the contralateral , unlesioned striatum are unaffected [22] . This unilateral model of PD has the advantage that each mouse can serve as its own within-subject control as the dopamine sensitized MSNs in the striatum of the 6-OHDA lesioned side respond with intense D1R-mediated signaling to administration of L-DOPA , while the MSNs in the unlesioned side are not affected [15] , [19] , [20] . In the first experiment , lesioned mice were injected with L-DOPA and sacrificed 1 hour later ( acute L-DOPA ) . By Western blotting , we observed that L-DOPA caused a dramatic increase in H3K27me3S28 phosphorylation in the lesioned striatum compared to the unlesioned striatum ( Figure 1B ) . The striatum contains two main populations of MSNs that are enriched for either D1Rs or dopamine D2 receptors ( D2Rs ) . Activation of these neurons produces opposite behavioral responses , which are related to their distinct connectivity to the output stations of the basal ganglia [23]–[26] . To identify the population ( s ) of striatal MSNs that responded to L-DOPA with increased H3K27me3S28 phosphorylation we made use of transgenic mice expressing EGFP under the control of regulatory elements of the D1R or D2R ( D1R-EGFP or D2R-EGFP ) [27] . These mice were lesioned , treated with L-DOPA and perfused after 1 hour . The results showed that , in the lesioned striata of D1R-EGFP mice , H3K27me3S28p co-localized with the EGFP-labeled cell bodies ( Figure 1C ) , while in D2R-EGFP mice , H3K27me3S28p was segregated from EGFP-labeled cell bodies ( Figure 1C ) . To further confirm that D1Rs activate the signaling cascade inducing the H3K27me3S28p mark , we injected naïve mice with the specific D1R-agonist SKF81297 . Indeed , H3K27me3S28p was increased 1 hour after SKF81297 injection ( Figure S1A ) . We concluded that in hemiparkisonian mice , where the dopamine innervation was eliminated by unilateral injection of 6-OHDA , acute administration of L-DOPA produced a large increase in H3K27me3S28p specifically localized to the nuclei of striatal D1R-MSNs . Given the dramatic increase in H3K27me3S28p in D1R expressing MSNs , we examined the global distribution of this mark on chromatin . We undertook chromatin immunoprecipitations ( ChIPs ) for H3K27me3S28p from pooled lesioned or unlesioned striata , after acute administration of L-DOPA ( tissue from 45 mice were pooled for each condition ) , followed by ChIP-sequencing ( ChIP-seq ) . This was also done using antibodies for H3K4me3 and H3K27me3 , to define potentially active or PcG-repressed chromatin , respectively . In this way , we identified four genes encoding transcription factors ( TFs ) where H3K27me3S28 phosphorylation was induced near the transcription start sites ( TSS ) ( Figure 2A ) . Two of them , Atf3 and Npas4 have previously been found to be implicated in neuronal plasticity , while Klf4 and Hoxa2 are well known PcG target genes implicated in stem cell function and cellular differentiation . In the lesioned striata the peaks observed for H3K27me3S28p at the Atf3 , Klf4 and Npas4 genes coincided with the H3K27me3 peaks . Co-occurrence was also observed for H3K27me3S28p and H3K4me3 ( Figure 2A ) . In contrast , chromatin at the Hoxa2 locus was blanketed with H3K27me3 and H3K27me3S28p in the lesioned striata , without any apparent H3K4me3 signal ( Figure 2A ) . To support these observations , we performed ChIP-qPCR on chromatin from lesioned and unlesioned striata after acute L-DOPA . Antibodies against H3K27me3S28p , H3K27me3 , H3K4me3 and Rnf2 ( PRC1 subunit ) , and general IgG as a negative control were used ( Figure 2B ) . The results confirmed the induction of H3K27me3S28p in the lesioned striatum upon L-DOPA stimulation and the presence of H3K27me3 at all genes analyzed , as well as H3K4me3 enrichment in chromatin from the unlesioned and lesioned striata on Atf3 , Klf4 and Npas4 genomic loci . In line with our previous findings [18] , the induction of H3K27me3S28 phosphorylation on these loci correlated with a reduction in Rnf2 binding . Whereas L-DOPA induced H3K27me3S28 phosphorylation in the striatum was restricted to D1R-MSNs , the origins of signal in the H3K27me3 and H3K4me3 ChIPs were uncertain , due to the presence of several cell types in the tissue . We have estimated the contribution of D1R-MSNs to the bulk chromatin analyzed in our ChIPs to be approximately 43% ( Figure S3 ) . Therefore , D2R-MSNs could account for a significant part of the H3K4me3 and H3K27me3 signals detected by ChIP . The increase in H3K27me3S28p and the reduced binding of Rnf2 suggested that these genes become de-repressed upon acute L-DOPA treatment . We therefore performed RT-qPCR using RNA isolated from lesioned mice treated with L-DOPA or saline as control ( Figure 2C ) . Indeed , L-DOPA increased the expression of mRNA for Atf3 , Klf4 and Npas4 in the lesioned compared to the unlesioned striatum and to the saline controls , whereas the expression of Hoxa2 was unchanged ( Figure 2C ) . The lack of expression from the Hoxa2 locus was supported by the absence of the active H3K4me3 histone mark . Importantly , the increases of Atf3 and Npas4 were furthermore confirmed at the protein level ( Figure 2D ) , suggesting the potential involvement of these TFs in the phenotypic effects produced by L-DOPA in parkinsonian mice . In summary , our data suggest that administration of L-DOPA in a mouse model of PD promotes H3K27me3S28 phosphorylation on several PcG target genes marked by H3K27me3 . This effect occurs in the striatal D1R-MSNs of the lesioned brain hemisphere and correlates with a reduction in Rnf2 binding and increased gene expression . To estimate the genome-wide extent of H3K27me3S28 phosphorylation induced by L-DOPA , we scored regions +/−1 kb from the transcription start sites ( TSS ) of all annotated ( mm9 ) mouse transcripts ( n = 189 , 660 ) for the enrichment of H3K27me3 in chromatin from unlesioned striata ( Figure 3A ) and H3K27me3S28p in lesioned striata ( Figure 3B ) . For this correlation , the H3K27me3 mark was determined in the unlesioned striata , instead of the lesioned striata , in order to ensure that the H3K27me3 signal was not altered by epitope masking due to H3K27me3S28 phosphorylation [18] . Using a cutoff where only 5% of the regions were expected to score positive by chance ( see Materials and Methods ) , this analysis showed that approximately 20 . 7% ( n = 39 , 197 ) of all loci that can give rise to mRNA transcripts were H3K27me3 positive ( Figure 3A ) and 6 . 9% ( n = 13 , 148 ) were H3K27me3S28p positive ( Figure 3B ) . As observed in the Venn diagram presented in Figure 3C the majority ( 83% ) of H3K27me3S28 phosphorylation occurred in genomic regions that were already marked by H3K27me3 before L-DOPA administration , while approximately 17% of the genomic sites had levels of H3K27me3 below our defined cut-off for the analysis and could in principle have gained H3K27me3 upon L-DOPA administration . Having observed that an impressive 33 . 5% of all H3K27me3 positive loci that potentially could give rise to mRNA transcripts became enriched for H3K27me3S28p upon L-DOPA stimulation , we next asked if and to which extent these transcripts were actually induced . We therefore performed global RNA-sequencing ( RNA-seq ) on mRNA isolated lesioned mice treated with L-DOPA . The analysis showed that 1 hour after L-DOPA , 5 . 4% ( n = 10 , 298 ) of all transcripts had significantly changed expression in the lesioned striatum compared to the unlesioned striatum ( Figure 3D ) . Importantly , 20 . 4% of these transcripts also scored positive for H3K27me3 at the chromatin level ( Figure 3E; threshold mentioned in Figure 3A ) . We next examined to which extent the regulated transcripts that were marked by H3K27me3 at their genomic loci also scored positive for H3K27me3S28p ( according to Figure 3B ) and found that 36 . 6% matched this criterion ( Figure 3F ) . It was furthermore apparent that most transcripts with high H3K27me3S28p levels near the TSS were induced rather than repressed . Further analysis of the different populations of regulated transcripts showed that , of the 768 that originated from H3K27me3- and H3K27me3S28p-positive loci , 52 were ≥1 . 5 fold down-regulated , whereas the majority 339 were ≥1 . 5 fold up-regulated ( Figure 3G ) . This could be compared to the 1 , 329 regulated transcripts originating from H3K27me3 positive loci lacking H3K27me3S28p of which 230 were ≥1 . 5 fold down-regulated and 332 were up-regulated ( Figure 3H ) , or to all 10 , 298 regulated transcripts ( regardless of specific histone marks at their loci ) , where 844 were ≥1 . 5 fold down-regulated and 2 , 892 were ≥1 . 5 fold up-regulated ( Figure 3I ) . Altogether , regulated transcripts originating from H3K27me3 positive loci that gained the H3K27me3S28p mark were more frequently induced ( 87% , Figure 3G ) than regulated transcripts originating from H3K27me3 positive loci not gaining the H3K27me3S28p mark ( 59% , Figure 3H ) or regulated transcripts in general ( 77% , Figure 3I ) . These data highlight a clear correlation between induction of H3K27me3S28 phosphorylation at specific genomic loci and increased transcription in response to acute L-DOPA stimulation in the lesioned striata of parkisonian mice . Gene ontology ( GO ) analysis of the group of transcripts up-regulated 1 . 5 fold or more originating from H3K27me3 positive loci that gained the H3K27me3S28p mark ( roman numerical II in Figure 3G ) , suggested that up-regulation of these gene products could affect overall transcriptional activity and rate of biosynthesis in neuronal cells ( Figure 4A , Figure S4 and Table S1 ) . This was in contrast to the group of up-regulated transcripts ( ≥1 . 5 fold ) originating from H3K27me3 positive loci that did not gain H3K27me3S28 phosphorylation , which enriched for GO-terms involved in immune response ( Figure 4B , Figure S4 , subpopulation IV ) . We have previously shown that MSK1 and MSK2 are the kinases mediating H3K27me3S28 phosphorylation in human fibroblasts , as pharmacological inhibition or shRNA mediated knockdown of MSK1 and MSK2 prevented the induction of H3K27me3S28p [18] . To elucidate which molecular pathways downstream of D1R activation in MSNs are implicated in the regulation of H3K27me3S28 phosphorylation , we lesioned Msk1−/− mice with 6-OHDA and injected them with L-DOPA . In line with our previous findings in fibroblasts , induction of H3K27me3S28p in the lesioned striatum was reduced in Msk1−/− mice compared to wt mice ( Figure S5 ) . The exaggerated D1R transmission induced by L-DOPA in the MSNs of the dopamine-depleted striatum is characterized by elevated cAMP production and PKA activity [11] , [13] . PKA further relays the signal via phosphorylation of DARPP-32 at T34 [15] , [28] . This converts DARPP-32 into an inhibitor of protein phosphatase 1 ( PP1 ) , thereby suppressing dephosphorylation of downstream PP1 targets [29] . We have previously shown that a T34A mutation on DARPP-32 decreases L-DOPA-induced phosphorylation of Erk and histone H3S10 [14] . Therefore , we examined whether DARPP-32-mediated inhibition of PP1 was also involved in the regulation of H3K27me3S28p . When lesioned mice harbouring a T34A mutation in DARPP-32 were injected with L-DOPA we observed a significant less pronounced H3K27me3S28 phosphorylation in comparison to wt mice ( Figure 5A ) . This finding suggested that PP1 is involved in the dephosphorylation of H3K27me3S28p . To test this possibility , we conducted an in vitro phosphatase assay , in which H3 peptides , either unmodified or modified with S28p or K27me3S28p , were incubated with PP1 . Changes in the phosphorylation of the peptides after the reactions were detected by dot-blotting using an H3S28p antibody ( Figure 5B ) . This assay showed that PP1 could efficiently dephosphorylate H3K27me3S28p . To confirm that PP1 is the phosphatase acting on H3K27me3S28p in the striatum , we examined the effect of okadaic acid , which inhibits PP1 and PP2A [30] , on H3K27me3S28p in a slice preparation from striatum by Western blotting ( Figure 5C ) . Incubation of striatal slices with 1 µM okadaic acid , a concentration that inhibits both PP1 and PP2A , was sufficient to change the equilibrium towards increased levels of H3K27me3S28p compared to vehicle . In contrast , a concentration of 100 nM okadaic acid , which inhibits PP2A but is insufficient for PP1 inhibition [30] , did not affect H3K27me3S28p levels . This supported PP1 as a phosphatase removing S28p from the H3K27me3S28p double-marked chromatin in vivo . Next , we analyzed the outcome of the global reduction in H3K27me3S28p on regulatory regions of specific genes . Lesioned DARPP-32 T34A and wt mice were treated with L-DOPA and striatal tissue from the lesioned and unlesioned striatum was analysed by ChIP-qPCR . The induction of H3K27me3S28p at the Atf3 , Klf4 , Npas4 and Hoxa2 genes was significantly reduced in chromatin from the lesioned striatum of DARPP-32 T34A mice compared to wt mice ( Figure 5D ) . Notably , this reduction correlated with decreased expression of Atf3 , Klf4 and Npas4 mRNA in the lesioned striatum of L-DOPA injected DARPP-32 T34A mice compared to wt mice ( Figure 5E ) . Overall these data suggested that D1R stimulation induces transcription associated to H3K27me3S28 phosphorylation via two parallel pathways: 1 ) activation of Erk-Msk kinases and 2 ) concomitant PKA-mediated DARPP-32-phosphorylation , leading to inhibition of PP1 and suppression of H3K27me3S28p dephosphorylation . LID is a serious motor complication caused by prolonged administration of L-DOPA to patients affected by PD [9] . This condition has been linked to persistent hyper-activation of the cAMP/DARPP-32 signaling cascade , produced by L-DOPA acting on sensitized D1Rs [10] , [11] . Lesioned mice display dyskinetic behaviour in response to 9 sequential daily L-DOPA injections ( chronic L-DOPA ) [14] , [15] . The severity of LID after chronic L-DOPA administration has been shown to correlate to the level of H3S10 phosphorylation and the induction of specific genes , such as Fosb [12] , [15] , [19] , [31] . To investigate the contribution of H3K27me3S28 phosphorylation to the changes in gene expression associated to LID , 6-OHDA lesioned mice were treated chronically with L-DOPA and the levels of H3S28p and H3K27me3S28p were measured after 1 , 3 and 9 days of administration . As expected , L-DOPA increased H3S28p and H3K27me3S28p in the lesioned striatum . However , the induction of these histone marks was progressively reduced during the course of chronic L-DOPA administrations ( Figure 6A ) . To examine if specific changes in gene expression associated to LID occurred for PcG-repressed genes , we performed global RNA-seq on mRNA isolated from striata of lesioned mice that had been treated chronically with L-DOPA ( 9 days ) . Transcripts that were significantly changed in the lesioned striatum after chronic L-DOPA were scored for enrichment of H3K27me3 in the unlesioned striatum and then for H3K27me3S28p in the lesioned striatum after acute L-DOPA , according to Figure 3A–F . From this analysis we concluded that despite the reduced induction of H3K27me3S28p following chronic administration of L-DOPA , regulated transcripts originating from genomic loci marked by H3K27me3 and H3K27me3S28p were still largely induced ( 76% , Figure S6A , subpopulation II ) in comparison to regulated transcripts originating from genomic loci marked by H3K27me3 only ( 45% , Figure S6B , subpopulation IV ) or in comparison to all regulated transcripts ( 46% , Figure S6C , subpopulation VI ) . As the induction of the H3K27me3S28p mark was reduced after chronic L-DOPA compared to acute L-DOPA ( Figure 6A ) , we next examined whether the transcripts from genes marked by H3K27me3 and H3K27me3S28p were less induced after chronic L-DOPA compared to acute L-DOPA treatment . Amongst the regulated transcripts originating from H3K27me3 and H3K27me3S28p positive genomic loci , 53% were ≥1 . 5 fold less expressed in chronic L-DOPA compared to acute L-DOPA ( Figure 6B ) . In contrast , only 11% of regulated transcripts originating from H3K27me3 positive and H3K27me3S28p negative genomic loci ( Figure 6C ) , and 24% of all regulated transcripts were less expressed in chronic L-DOPA compared to acute L-DOPA . These data suggested that the level of transcription from H3K27me3 and H3K27me3S28p positive genes correlated with the level of induced H3K27me3S28 phosphorylation . Finally , we examined the individual genes that were ≥1 . 5 fold induced after chronic L-DOPA and scored positive for H3K27me3 and H3K27me3S28p . We found that a large number of these genes differed from those induced by acute L-DOPA ( Figure 6E , Table S1 and Table S2 ) . Thus , a total of 96 genes were induced after acute L-DOPA and 114 genes after chronic L-DOPA , but only 43 genes were commonly induced . To confirm this observation , we performed RT-qPCR on selected genes and could confirm , for instance , that Ppm1n and Galr2 were induced after acute L-DOPA , but not after chronic L-DOPA or saline ( Figure 6F; see also Figure S1B showing that the lesion alone does not induce H3K27me3S28p ( vehicle control ) ) . Atf3 , Klf4 , and Npas4 were induced by both acute L-DOPA and , albeit to a lesser extent by chronic L-DOPA ( Figure 6G ) , whereas Nr4a2 ( also known as Nurr1 ) , Ngf and Lipg were only induced after chronic L-DOPA ( Figure 6H ) . For these genes we could observe peaks in the ChIP-seq data for H3K27me3 in the lesioned and unlesioned striata , induced H3K27me3S28p in the lesioned striatum and H3K4me3 peaks in the lesioned and unlesioned striata ( Figure S7 ) . As an example of a gene that was induced at the protein level in MSNs expressing the D1-receptor , we stained for Atf3 after chronic L-DOPA stimulation ( 4 hours timepoint after the last L-DOPA administration ) as shown in Figure S6D . Overall , these data suggested that increased transcription from a subset of PcG-repressed genes was associated with the development of LID . The expression level of H3K27me3 and H3K27me3S28p marked genes was generally lower in chronic L-DOPA stimulated mice compared to acute L-DOPA stimulated mice , correlating with reduced induction of the H3K27me3S28p mark . However , repeated L-DOPA administration resulted in the induction of a unique group of PcG regulated genes that were not induced after a single L-DOPA injection . For the group of genes that gained H3K27me3S28p ( subgroup II in Figure 6B ) and had significantly higher expression in chronic L-DOPA stimulated lesioned mice compared to acute stimulated mice the most significantly enriched GO-term was “Adult behavior” . However , the term comprises only five genes from subpopulation II ( Figure 6B ) : Nr4a2 , Nr4a3 , Trh , Npy and Adra1b ( out of 93 potential genes in the genome ) , and although this was a significant ( p = 0 . 0011 ) 10 . 7 fold enrichment over the expected number , it did not remain significant when Benjamini-Hockberg corrected for multiple testing .
We have previously shown that H3K27me3S28 phosphorylation causes the displacement of PRC1- and PRC2-complexes from chromatin , leading to expression of a subset of PcG regulated genes in cultured human fibroblasts [18] . Here , we for the first time provide in vivo evidence for the relevance of this mechanism for gene regulation in adult , post-mitotic neurons utilizing a mouse model of PD . We show by genome wide analyses that , in striatal MSNs , signaling through sensitized D1Rs induces H3S28 phosphorylation in the context of H3K27me3 marked genes . Importantly , the H3K27me3S28p mark correlates with a reduction in PcG binding and increased transcription of a subset of genes , several of which have been implicated in neuronal plasticity . Notably , a systemic injection of a specific D1R agonist was also able to induce the H3K27me3S28p mark in naïve mice . This clearly indicates that the mechanisms described in this study have a general relevance with regard to D1R transmission in the adult brain . Taking advantage of the pronounced effects produced by dopamine depletion , we mapped putative downstream genomic targets regulated by L-DOPA via D1Rs . Our genome-wide analysis showed that at least 1/3 of all H3K27me3 marked gene loci , that can potentially give rise to a mRNA transcript in the non-repressed state , gained H3K27me3S28 phosphorylation upon acute L-DOPA administration . Most importantly , among regulated transcripts , phosphorylation of S28 in the context of H3K27me3 was a strong indicator of transcriptional activation . Interestingly , the combined analyses of RNA-seq and ChIP-seq data showed that the majority of transcripts originating from H3K27me3 genomic loci that gained H3K27me3S28 phosphorylation did not change expression upon L-DOPA stimulation . As previously demonstrated , H3K27me3S28 phosphorylation leads to removal of PcG complexes at H3K27me3 marked regions and is considered to derepress the promoter [18] . Therefore it appears that , in order to fully activate genes , H3K27me3S28 phosphorylation requires additional events to take place , which likely would be H3K4 methylation and histone acetylation . In line with our previous findings , induction of the H3K27me3S28p mark in response to L-DOPA was reduced in 6-OHDA lesioned Msk1−/− mice compared to wt mice ( Figure S5 ) . We found that the induction of H3K27me3S28p was less pronounced , but not abolished , in Msk1−/− mice , indicating that in striatal MSNs other histone kinases must be present and actively relay dopamine signaling to chromatin . The phosphatase responsible for the dephosphorylation of H3K27me3S28p has so far been elusive . Our results , showing that H3K27me3S28 phosphorylation induced by L-DOPA is decreased by abolishing PKA-mediated activation of DARPP-32 , pointed to PP1 as a plausible candidate . This idea was corroborated by in vitro and ex vivo experiments showing that PP1 could dephosphorylate H3K27me3S28p ( Figure 5B and C ) . Overall , our results support a model in which inhibition of PP1 via PKA/DARPP-32 works in parallel with Msk1 to promote H3K27me3S28 phosphorylation in response to activation of D1R ( see model in Figure 7 ) . Dyskinesia is a serious motor complication caused by prolonged administration of L-DOPA to parkinsonian patients [9] . It has been proposed that LID depends on the persistent and intermittent hyper-activation of the cAMP/DARPP-32 signaling cascade produced by L-DOPA through sensitized D1Rs . This , in turn , leads to hyper-phosphorylation of histone H3 and aberrant expression of specific genes implicated in dyskinetic behavior [10] , [11] . We have previously shown that a T34A mutation on DARPP-32 decreases dyskinetic behaviour in 6-OHDA lesioned mice and that this effect correlates with reduced histone H3S10 phosphorylation [14] . In this study , we found that the induction produced by L-DOPA on H3K27me3S28 phosphorylation at regulatory regions of the Atf3 and Npas4 genes , which are bound and repressed by PcG proteins in MSNs under normal conditions , correlated with increased mRNA- and protein synthesis . We also show that H3K27me3S28 phosphorylation and the associated transcriptional activation are largely reduced in DARPP-32 T34A mutant mice . These observations are particularly interesting in view of the involvement of Atf3 and Npas4 in synaptic plasticity and long-term adaptive responses . Augmented expression of Atf3 in the dorsal striatum has been observed following acute and chronic administration of amphetamine [32] , a drug that , similarly to L-DOPA , promotes dopamine transmission . Furthermore , the TF Npas4 has been proposed to promote the expression of other immediate early genes , including Arc , Egr1 and c-Fos [33] , which have all been associated to chronic administration of L-DOPA and to the development of LID [15] , [34] , [35] . Our results show that repeated administration of L-DOPA decreases its ability to induce global phosphorylation of H3K27me3S28 ( Figure 6A ) . This is not surprising since desensitization of kinase signaling pathways following persistent upstream activation and negative feedback is a common phenomenon . Accordingly , previous work showed that repeated administration of L-DOPA to 6-OHDA-lesioned mice is accompanied by a partial normalization of sensitized D1R signaling , reflected in lower levels of Erk activation and reduced global phosphorylation of H3 at S10 [15] , [36] , [37] . Importantly , despite the attenuated induction of H3K27me3S28 phosphorylation after chronic L-DOPA , our genome-wide analysis showed that there was still a good correlation between the genomic loci gaining the H3K27me3S28p mark and gene induction . This suggested that residual kinase activity was sufficient to drive transcription from PcG-repressed genes in the context of LID ( Figure S6A–C ) . The decreased ability of L-DOPA to induce H3K27me3S28p observed after chronic administration correlated with reduced expression of transcripts from H3K27me3 and H3K27me3S28p marked genes ( Figure 6B ) , but not of transcripts from H3K27me3 marked genes lacking the H3K27me3S28p mark or transcripts in general ( Figure 6C–D ) . As an exception , IE genes such as Atf3 , c-Fos and Egr2-4 , which encode TFs , were all induced to at least the same extent after chronic L-DOPA administration compared to acute administration ( Table S2 ) . This suggests that these IE genes , which are characterized by transient expression , respond with fast on/off-kinetics presumably due to rapid dephosphorylation of S28 by PP1 . This would allow PcG complexes to re-bind , when the signal leading to Msk1 activation ceases , thereby re-setting gene repression until the next activating signal , like for the repeated daily L-DOPA administration in chronically treated parkisonian mice . The IE genes , Nr4a2 and Nr4a3 , also known as Nurr1 and Nor1 , encode orphan nuclear receptors that function as TFs . These genes were more strongly induced after repeated L-DOPA administration compared to acute L-DOPA administration . It has been shown previously that the expression of Nr4a2 , a gene product involved in the development of dopaminergic neurons [38] , is increased in response to prolonged treatment with L-DOPA and that this effect occurs in the MSNs expressing D1Rs [39] . Notably , recent evidence indicates that viral vector-induced overexpression of Nr4a2 in striatal neurons increases dyskinesia in a rat model of PD [40] . Our results are in line with these observations and provide a possible mechanism accounting for the enhanced expression of Nr4a2 in response to chronic L-DOPA administration . In the striatum , Nr4a3 mRNA expression is induced by activation of D1Rs and this effect is prevented by blockade of Erk signaling [41] . Whereas the exact role of Nr4a3 in dopaminergic transmission remains to be elucidated , the present data support the idea that , in PD , this gene is de-repressed via D1R-mediated activation of Erk and increased H3K27me3S28 phosphorylation . Interestingly , we identified other PcG regulated genes such as neuropeptide encoding genes: galanin ( Gal ) , thyrotropin releasing hormone ( Trh ) and neuropeptide Y ( Npy ) , which all were only induced upon chronic treatment with L-DOPA ( Table S2 ) . In contrast to IE genes , these neuropeptide encoding genes seems to be more tightly regulated , and their activation presumably requires persistent phosphorylation of H3K27me3 marked and PcG bound regions , as well as induction of additional TFs and co-factors to lead to their transcriptional activation . The regulation of Trh is in line with previous work showing a large increase in the mRNA for this hormone occurring in the striatal MSNs of dyskinetic rats [42] . Subsequent analysis established a correlation between Trh expression in D1R-expressing MSNs and dosage of L-DOPA , which is regarded as a critical factor in the development of dyskinesia [39] . Indeed , enhanced levels of Trh may concur to the development of dyskinesia , since hyperthyroidism is typically associated to hyperkinesia . Increased levels of NPY mRNA have been found in the striata of Parkinsonian patients treated with L-DOPA [43] . The present data suggest the possibility that this increase may occur in striatal MSNs , although in naïve mice , NPY is mainly expressed in GABAergic interneurons [44] . NPY has been proposed to exert neuroprotective effects on dopaminergic neurons [45] , however further studies will be necessary to clarify its potential role in dyskinesia . Galanin ( Gal ) and galanin receptors are involved in many neuronal functions including drug addiction [46] . In the striatum , galanin receptors have been localized to cholinergic interneurons and neuronal terminals , which are critically involved in the regulation of the excitability of MSNs [47] , [48] . Galanin has also been shown to reduce dopamine release in the striatum [49] and to inhibit spontaneous locomotion by reducing the activity of midbrain dopaminergic neurons [50] . Therefore , the increase in galanin mRNA expression produced by chronic administration of L-DOPA may lead to significant modifications in basal ganglia neurotransmission . Future studies will be necessary to determine the impact of these changes on the development and manifestation of dyskinesia . The functional implications of differences in reactivation of PcG repressed genes remain to be investigated in more detail . However , it is possible that genes that are activated only in response to acute administration of L-DOPA are implicated specifically in the emergence of LID , whereas genes that are activated only in response to chronic L-DOPA are involved in the consolidation and manifestation of this condition . Interestingly , one of the receptors for the neuropeptide galanin , Galr2 was only induced in response to acute L-DOPA ( Table S1 and Figure 6F ) , while the gene encoding the ligand , Gal , was only induced by chronic L-DOPA stimulation as mentioned above . Further studies are necessary to examine which permissive determinants are required at genomic sites for H3K27me3S28p mediated expression of PcG target genes . Nevertheless , the fact that many genes regulated by PcG proteins are TFs and among them many are affected by H3K27me3S28 phosphorylation and upregulated by acute and chronic L-DOPA , suggest that LID is , at least in part , a consequence of secondary transcriptional events . These TFs might only need to be induced transiently in order to trigger transcription of other genes that lead to sustained changes in neuronal plasticity associated to long-term maladaptive responses , such as dyskinesia . Overall our novel findings reveal a previously unrecognized plasticity of PcG-repressed genes in terminally differentiated neurons . The identification of specific genes whose expression is increased upon prolonged treatment with L-DOPA and dopamine D1 receptor stimulation offer a possibility to design novel therapeutic strategies to treat Parkinson's disease and potentially other disorders caused by dysfunctional dopaminergic transmission in the brain , such as drug addiction and schizophrenia .
Female C57BL/6J mice ( 30 g ) were purchased from Taconic ( Tornbjerg , Denmark ) . Bacterial artificial chromosomes transgenic mice expressing EGFP under the control of the promoter for the D2R ( Drd2-EGFP ) or the dopamine D1R ( Drd1a-EGFP ) were generated by the Gene Expression Nervous System Atlas program at the Rockefeller University [27] and were crossed on a C57BL/6 background for three generations . DARPP-32 T34A mutant mice [51] and Msk1−/− mice [52] have already been described . The animals were housed in groups of five under standardized conditions with 12 hours light/dark cycle , stable temperature ( 20°C ) , and humidity ( 40–50% ) . All protocols utilized to generate the model of PD and dyskinesia ( including chronic administration of L-DOPA ) , were approved by the Research Ethics Committee of Karolinska Institutet and by the Swedish Animal Welfare Agency ( permit N515/12 ) . Mice were anesthetized with a mixture of fentanyl citrate ( 0 . 315 mg/ml ) , fluanisone ( 10 mg/ml ) ( VetaPharma , Leeds , UK ) , midazolam ( 5 mg/ml ) ( Hameln Pharmaceuticals , Gloucester , UK ) , and water ( 1∶1∶2 in a volume of 10 ml/kg ) and mounted in a stereotaxic frame ( David Kopf Instruments , Tujunga , CA ) equipped with a mouse adaptor . 6-OHDA-HCl ( Sigma-Aldrich Sweden AB ) was dissolved in 0 . 02% ascorbic acid in saline at a concentration of 3 µg of free-base 6-OHDA per microliter . Each mouse received one unilateral ( right hemisphere ) injection of 6-OHDA of 1 µl ( 0 . 5 µl/min ) into the medial forebrain bundle according to the following coordinates ( mm ) : anteroposterior ( AP ) , −1 . 2; mediolateral ( ML ) , −1 . 2; dorsoventral ( DV ) , −4 . 8 ( all millimeters relative to bregma ) [53] . Noradrenergic neurons were protected by injection of 25 mg/kg desipramine ( Sigma ) thirty minutes prior to 6-OHDA injection . This procedure leads to a decrease in striatal tyrosine hydroxylase immunoreactivity ≥80% and to a marked akinesia affecting the side of the body contralateral to the lesioned striatum . Animals were allowed to recover for 2 weeks before experimentation . L-DOPA ( 10 mg/kg in combination with 7 . 5 mg/kg benserazide; purchased from Sigma ) was dissolved in saline ( 0 . 9% NaCl ) , and injected intraperitoneally in a volume of 10 ml per kilogram of body weight for 1 , 3 or 9 days . SKF81297 ( 3 mg/kg ) was purchased from Tocris . Mouse embryonic stem ( mES ) cells , wildtype ( wt ) E14 ( provided by Dr . Zhou-Feng Chen and Dr . Helle Færk Jørgensen ) and Eed−/− ( provided by Dr . Anton Wutz ) were cultured on 0 . 1% ( w/v ) gelatin-coated plates in ES medium ( Glasgow Minimum Essential Medium ( Sigma ) supplemented with Glutamax-1 ( Gibco ) , non-essential amino acids ( Gibco ) , 50 mM 2-mercaptoethanol , 15% ( v/v ) ES-cell-qualified FBS ( Gibco ) , and 1% ( v/v ) penicillin/streptomycin ) in the presence of 1 , 000 U/ml of LIF ( Millipore ) . To induce histone phosphorylation , the mES cells were stimulated with 1 µg/mL anisomycin in DMSO or DMSO only as control . For ChIP , cells were cross-linked for 10 min at room temperature in culture media containing 1% formaldehyde , 10 mM Hepes ( pH 8 . 0 ) , 0 . 1 mM EGTA , and 20 mM NaCl . Cross-linking was stopped by addition of glycine to a final concentration of 0 . 125 M , followed by an additional incubation for 5 min . Fixed cells were washed 3 times with PBS and harvested in SDS lysis buffer ( 50 mM Tris at pH 8 . 1 , 0 . 5% SDS , 100 mM NaCl , 5 mM EDTA , 1 mM PMSF , 10 µg/ml leupeptin and 10 µg/ml aprotinin ) . The cells were then pelleted for 10 min at 2 , 400 g followed by the same ChIP protocol as for striatal tissue . The included primer sequences are listed in Table S3 . Mice were killed by decapitation , the heads of the animals were cooled in liquid nitrogen for 6 s and the brains were removed . Coronal slices of 1 mm thickness were obtained from a mouse brain dissection matrix ( Activational Systems Inc . , RBM-2000C ) , and three striatal punches of 2 mm diameter from sequential slices were dissected out on an ice-cold surface , sonicated in 1% SDS , and boiled for 10 min . Proteins were separated by SDS–polyacrylamide gel electrophoresis and transferred overnight to PVDF membranes ( Amersham Pharmacia Biotech , Uppsala , Sweden ) . The following antibodies were used: H3S28p ( Millipore , 07-145 ) , H3K27me3S28p ( Hansen lab ) , histone H3 ( Abcam , ab1791 ) , Atf3 ( Santa Cruz , sc-188 ) . The Npas4 antibody was provided by Prof . Greenberg , Harvard Medical School , Boston , USA . Mice were rapidly anaesthetized with pentobarbital ( 300 mg/kg ip , Sanofi-Aventis , France ) and perfused transcardially with 4% ( w/v ) paraformaldehyde in 0 . 1 M sodium phosphate buffer ( pH 7 . 5 ) . Brains were post-fixed overnight in the same solution and stored at 4°C . Forty-micrometer-thick sections were cut with a vibratome ( Leica , Nussloch , Germany ) . Free-floating sections were rinsed in tris-buffered saline , permeabilized in 0 . 2% Triton X-100 in TBS for 20 min and blocked to prevent non-specific binding by incubation in 0 . 5% Triton X-100 , 5% normal goat serum , 1% bovine serum albumin in TBS for 1 hr at RT . Sections were incubated overnight at 4°C with primary antibodies . The following antibodies were used: EGFP ( Aves Lab , GFP-1020 ) , NeuN ( Abcam , ab138452 ) . Antibodies for histone marks were the same as for Western blotting . Images from the dorsolateral striatum were obtained by sequential laser scanning confocal microscopy ( Zeiss LSM 510 Meta ) . Data was analyzed by one-way or two-way ANOVA when appropriate followed by Tukey's HSD post-hoc test . Unpaired t-test was used when comparing two means . p<0 . 05 was considered significant . Naive C57BL/6 mice were killed by decapitation , and the brains were rapidly removed . Coronal slices ( 250 µm ) were prepared with the use of a vibratome ( Leica , Nussloch , Germany ) . Dorsal striata were dissected out from each slice under a microscope . Two slices were placed in individual 5-ml polypropylene tubes containing 2 ml of Krebs-Ringer bicarbonate buffer . The samples were equilibrated at 30°C for 30 minutes , followed by incubation of the slices with either vehicle ( DMSO ) , 100 nM or 1 µM okadaic acid ( Sigma ) in 2 mL fresh buffer for 50 min . After incubation , the solutions were rapidly removed , the slices were sonicated in 1% SDS , and the samples were analyzed by Western blotting as described . Ten µg of N-terminal H3 peptides representing the first 40 amino acids of histone H3 . 1 were used per reaction , either unmodified or modified as follows: S28 phosphorylated ( H3S28p ) or K27 tri-methylated and S28 phosphorylated ( H3K27me3S28p ) . PP1 ( 1 unit ) was added to the reaction containing 10 µg of specified H3 peptide in a 15 µl de-phosphorylation buffer: 50 mM HEPES , pH 7 . 5 , 100 mM NaCl , 2 mM DTT , 0 . 01% Brij 35 and 1 mM MnCl2 . The de-phosphorylation reaction was allowed to proceed at 30°C , for 30 min and stopped by adding EDTA ( pH 8 . 0 ) to a final concentration of 5 mM . A fraction of each reaction was spotted on a nitrocellulose membrane corresponding to: 1 . 0 µg , 0 . 1 µg and 0 . 01 µg H3 . 1 peptide . The membrane was blocked as for standard Western blotting and developed using antibodies for H3S28p ( Millipore 07-145 , 1∶3 , 000 ) and secondary anti-rabbit HRP ( Vector Laboratories ) . Dot-blots were added enhanced chemiluminiscence ( ECL ) after the last wash and exposures were made using a ImageQuant LAS 4000 camera system . The quantifications were made based on the dots containing 1 µg peptide . Tissue punches for chromatin preparation was obtained as described for Western blotting . The punches were fixed for 12 min in cold 1% formaldehyde/PBS followed by glycine incubation to stop further cross-linking . The fixed punches were then washed 3× with cold PBS containing phosphatase inhibitors ( 20 nM okadaic acid , 10 µM NaF ) and subsequently snap-frozen for later use . Chromatin immunoprecipitation experiments were performed as described [54] with some modifications: Fixed striatal punches were homogenized in a nuclear extraction buffer ( 10 mM Tris ( pH 8 . 0 ) , 100 mM NaCl , 2 mM MgCl2 , 0 . 3 M Sucrose , 0 . 25% IGEPAL CA-630 ) containing protease inhibitors ( 1 mM PMSF , 0 . 1 mM aprotinin , 0 . 1 mM leupeptin ) and phosphatase inhibitors ( 20 nM okadaic acid , 10 µM NaF ) , by douncing 15 times using a 2 mL loose grind pestle followed by a 30 min incubation on ice . The homogenate was dounced another 50 times using a 2 mL loose grind pestle for nuclear release , followed by 10 min centrifugation at 2 , 400 g to pellet nuclei . The extracted nuclei were then lysed in a lysis buffer containing 50 mM Tris-HCl ( pH 8 . 0 ) , 10 mM EDTA , 1% ( wt/vol ) SDS and protease/phosphatase inhibitors , diluted in RIPA buffer ( 10 mM Tris-HCl ( pH 7 . 5 ) , 140 mM NaCl , 1 mM EDTA , 0 . 5 mM EGTA , 0 . 1% ( vol/vol ) Triton-X-100 , 0 . 1% ( wt/vol ) SDS , 0 . 1% ( wt/vol ) Na-deoxycholate ) and the DNA was sonicated to an average size of 300–500 bp using a Bioruptor standard device ( Diagenode ) ( 10 cycles 30 sec ON , 30 sec OFF , highest setting ) . 10 µg anti-rabbit IgG ( DAKO ) , 3 µg of anti-H3 ( “GERA” , Hansen lab ) , 2 . 5 µg of anti-Rnf2 ( “NAST” , Hansen lab ) anti-H3K27me3 ( 9756 , Cell Signaling ) and 2 µg anti-H3K4me3 ( Lys4 ) ( 9751 , Cell Signaling ) and anti-H3K27me3S28p ( the specificity of the batch #5 of peptide antigen purified H3K27me3S28p antibody used in this study was tested as shown in Figure S2 ) was incubated at 4°C with 25 µL washed Dynabeads protein A ( Invitrogen ) and RIPA in a total volume of 100 µL . The bead-antibody complexes were then incubated at 4°C for 2 h with 20 µL chromatin in a total volume of 250 µL . Beads were washed in 3× RIPA , 1× high salt wash buffer ( 20 mM Tris-HCl ( pH 7 . 5 ) , 500 mM NaCl , 2 mM EDTA , 0 . 1% Triton-X-100 , 0 . 1% SDS ) , 1× LiCl buffer ( 10 mM Tris-HCl ( pH 7 . 5 ) , 250 mM LiCl , 1 mM EDTA , 1% Na-deoxycholate , 1% IGEPAL CA-630 ) and 1× TE buffer . After washes , DNA was eluted from beads and de-crosslinked in 20 mM Tris-HCl , pH 7 . 5 , 5 mM EDTA , 50 mM NaCl . 1% ( wt/vol ) SDS and 50 µg/mL protease K at 68°C overnight . For input , 20 µL chromatin was de-crosslinked in 20 mM Tris-HCl , pH 7 . 5 , 5 mM EDTA , 50 mM NaCl 68°C overnight . ChIP and input DNA was then purified and eluted using Minelute PCR purification kit ( Qiagen ) . Enrichments on selected loci were measured by qPCR , 3 technical replicates , ( 7500 Fast , Applied Biosystems ) relative to a 5-point dilution series of input chromatin . The included primer sequences are listed in Table S3 . Student's t test were used to compare means of the different conditions . ChIPs for each experimental condition was performed in at least triplicates . The resulting immunoprecipitated DNA were pooled and prepared for ChIP sequencing using an Illumina kit according to the manufacturer's guidelines . 2 nanogram of starting material , as determined by PicoGreen concentrations , was used in each case . Sequencing was performed on a Genome Analyzer II ( Illumina ) at the National High-throughput Sequencing Centre in Copenhagen . Libraries were de-multiplexed and high quality reads ( Chastity score > = 0 . 6 ) were aligned to the mouse genome ( mm9 ) using Bowtie [55] allowing up to two mismatches . Reads not aligning uniquely to the mouse genome were removed and only unique reads were used for subsequent analysis . Tracks from single genomic loci were presented using the UCSC Genome Browser ( http://genome . ucsc . edu/ ) [56] . Reads were normalized to a library size of 10M reads and converted to wig-files using the program EaSeq ( Lerdrup et al , manuscript in preparation ) . All quantitation , scoring , gating , and visualization was done in the program EaSeq ( Lerdrup et al , manuscript in preparation ) . A list of all transcripts including genomic coordinates was derived from Genomatix ( see RNA-sequencing for details ) , and the amounts of ChIP-seq signal at the genomic regions corresponding to −1 kbp to +1 bkp was quantified and normalized to dataset size and a region-size of 1 kbp . Transcripts were scored positive or negative by fitting the abundance to the quantified ChIP-seq signal at all transcripts genome-wide to a normal distribution . A threshold was automatically applied at the level where the only 5% of the regions were expected to score positive by chance ( thresholds were 6 . 8 for left hemisphere H3K27me3 and 16 . 9 for right hemisphere H3K27me3S28P ) . Transcripts were gated into subpopulations depending on the level of ChIP-seq signal relative to these thresholds and/or fold change in gene expression as well as significance in Benjamini-Hochberg corrected p-values [57] . One striatal punch per hemisphere was dissected out and subsequently put in RNAlater ( Qiagen ) at +4°C over night to inhibit RNA degradation . Total RNA was extracted using a RNeasy kit ( Qiagen ) and quantified on a NanoDrop 1000 device . 200 ng of RNA was used for generation of cDNA using a TaqMan Reverse Transcription Reagents kit ( Invitrogen ) . Expression levels for individual transcripts were measured by qPCR and calculated by the ddCt-method using TATA-binding protein ( Tbp ) mRNA as housekeeping gene . The expression levels were based on 3 biological replicates . The included primer sequences are listed in Table S3 . A TruSeq RNA Sample preparation kit ( Illumina ) was used for library generation out of 0 . 5 µg of total RNA per condition . The generated libraries were sequenced on a Genome Analyzer II ( Illumina ) at the National High-throughput Sequencing Centre in Copenhagen . Reads were mapped to the mouse genome ( mm9 ) using the Genomatix Mining Station software ( Genomatix ) . Differential expression analysis was done on triplicates using the region miner task “Expression Analysis for RNASeq Data” on a Genomatix Genome Analyzer ( Genomatix ) using the DeSeq algorithm [58] . Gene ontology ( GO ) enrichment analyses were done from sets of genes with a significant ( Benjanimi-Hochberg corrected for multiple testing ) increase or decrease in expression of at least 1 . 5 fold using the DAVID functional annotation tool at http://david . abcc . ncifcrf . gov/ [59] , [60] . The Geo accession number for the ChIP-seq and RNA-seq data reported in this paper is GSE60703 . | In Parkinson's disease ( PD ) the motor impairment produced by the progressive death of midbrain dopaminergic neurons is commonly treated with the dopamine precursor , L-DOPA . Utilizing a mouse model of PD , we show that L-DOPA , via activation of dopamine D1 receptors , promotes the expression of genes normally repressed by Polycomb group ( PcG ) proteins . We propose that this effect is exerted by promoting the phosphorylation of histone H3 on serine 28 at genomic regions marked by tri-methylation of the adjacent lysine 27 , generating a H3K27me3S28p double-mark . This event leads to displacement of PcG proteins and aberrant gene expression . These findings reveal a previously unrecognized plasticity of PcG-repressed genes in terminally differentiated neurons . Furthermore , the identification of specific genes whose expression is increased upon prolonged treatment with L-DOPA and the consequential activation of dopamine D1 receptors offer a possibility to design novel therapeutic strategies to treat Parkinson's disease and potentially other disorders caused by dysfunctional dopaminergic transmission in the brain , such as drug addiction and schizophrenia . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biology",
"and",
"life",
"sciences",
"medicine",
"and",
"health",
"sciences"
] | 2014 | Dopamine Signaling Leads to Loss of Polycomb Repression and Aberrant Gene Activation in Experimental Parkinsonism |
We performed a genome-level computational study of sequence and structure similarity , the latter using crystal structures and models , of the proteases of Homo sapiens and the human parasite Trypanosoma brucei . Using sequence and structure similarity networks to summarize the results , we constructed global views that show visually the relative abundance and variety of proteases in the degradome landscapes of these two species , and provide insights into evolutionary relationships between proteases . The results also indicate how broadly these sequence sets are covered by three-dimensional structures . These views facilitate cross-species comparisons and offer clues for drug design from knowledge about the sequences and structures of potential drug targets and their homologs . Two protease groups ( “M32” and “C51” ) that are very different in sequence from human proteases are examined in structural detail , illustrating the application of this global approach in mining new pathogen genomes for potential drug targets . Based on our analyses , a human ACE2 inhibitor was selected for experimental testing on one of these parasite proteases , TbM32 , and was shown to inhibit it . These sequence and structure data , along with interactive versions of the protein similarity networks generated in this study , are available at http://babbittlab . ucsf . edu/resources . html .
The recent explosion of sequence and structure information in the public databases has made possible surveys of proteins on a genomic scale . Structural genomics initiatives have rapidly increased our knowledge of protein structures , and this has provided a foundation for better elucidating reaction mechanisms and ligand binding , for understanding substrate specificity , and for creating many new three-dimensional ( 3D ) structural models [1] . One use of such information is to inform drug discovery . For example , Vidovic et al examined human protein-tyrosine phosphatases as targets for rational drug design [2] . Rational drug design has been shown to produce selective drugs: for example , a number of effective cancer drugs have been produced that have fewer side effects than traditional , cytotoxic drugs [3] . Evaluating potential off-target effects is an important consideration in the process , and surveying homologs of target proteins can reveal unanticipated interactions [4] , [5] . Conversely , some drugs show efficacy with unanticipated targets making them useful in treating diseases other than those for which they were designed . For example , Kinnings et al . used a computational approach to compare proteins with similar binding sites to those of the targets of commercially available drugs and found that a drug approved for Parkinson's disease may be effective for treating tuberculosis [6] . Thus , the larger context provided by examining the differences between structurally related proteins may aid in the design of more selective drugs , while study of their similarities can give clues for new starting points in drug design . While the public databases are rich with sequence and structure data , retrieving specific data and synthesizing the information into views that are intuitively interpretable is not a trivial task , even for an experienced user of bioinformatics tools . The central idea behind our study is to take the approach illustrated by Vidovic et al one step further and construct genome-wide views for more than one organism; specifically , for a host and parasite , allowing cross-species comparisons . We constructed these views using sequences and structures of the proteases of the pathogen Trypanosoma brucei and its human host Homo sapiens . The full diversity of a set of sequences or structures is often termed “sequence space” or “structure space . ” To visualize the information , we used similarity networks , whereby sequences or structures are clustered graphically by similarity . Such networks represent a powerful way to visualize relationships across large sets of sequences and structures [7] . To construct the structure similarity networks , existing crystal structures and homology models , as well as newly created models , were utilized . Proteases were chosen for this computational study because a number of these proteins have been validated as druggable targets and many have available structures . Protease inhibitors are currently under investigation to treat various parasite infections , cancer , HIV , hypertension , and diabetes [8]–[18] . Here , we employ the nomenclature of the protease database MEROPS [19] in labeling proteases by the evolutionary units of family and clan ( see Methods ) . Proteases catalyze the hydrolytic breakdown or processing of proteins and account for about 2% of all expressed genes [20] . The set of an organism's proteases expressed at a particular time or circumstance has been called its “degradome” [21] . Here , we use the term to refer more generally to all the active proteases coded by an organism's genome . The parasite degradome targeted in this study is that of the protist T . brucei , which causes human African trypanosomiasis ( “HAT” ) or sleeping sickness , a disease that affects an estimated 50 , 000 to 70 , 000 people , mostly in sub-Saharan Africa [22] . HAT is one of a number of ‘neglected’ tropical diseases that primarily afflict the poor [15] , [22] . The few existing treatments for such diseases often have severe side effects . Without drug treatment , HAT is often fatal; yet the standard drug used to treat infection of the central nervous system is itself often lethal [23] . T . brucei is related to two other human pathogens , Trypanosoma cruzi and Leishmania major , that share many physical characteristics [24] . These three species are referred to as the “Tritryps” [24] . As will be illustrated here for T . brucei , knowledge about well-characterized proteins in the other Tritryp species is valuable for inferring characteristics about related but less well-characterized proteins in the target species , for example , by enabling the creation of homology models . The objectives of this study were first to compare the array of proteases in the human host with that of the T . brucei pathogen and to determine the breadth of sequence space in each organism that was covered by 3D structure . Secondly , we aimed to use the similarities and differences within and between protease sequence and structure similarity groups to obtain insights into possible new targets for drug design . The global views produced here will also be useful for guiding phylogenetic and other more detailed studies comparing proteases of the parasite and its human host . We found that structure coverage of sequence space in human and parasite is broad , making global structural comparisons both feasible and informative . To illustrate how these results may be used to better understand structurally related human and parasite proteases , we include a detailed structural evaluation of two groups of parasite proteases that may have potential as new drug targets . For one of these protease targets , TbM32 , we predicted and experimentally confirmed its inhibition by a known human drug .
Predicted T . brucei proteins were from the T . brucei genome project ( 9 , 192 sequences , file “Tb927_Proteins_May08_v4 . fas” downloaded from ftp . sanger . ac . uk ) [26] , and will be referred to hereafter as “Tb_proteins . ” To identify protease-like sequences , we used a protocol similar to that used by Berriman et al . [27] . The Tb_proteins were BLAST searched using blastp [28] against pep82 , and the results limited to those with E-value cutoff ≤1e-4 , which yielded 477 “provisional proteases . ” Note: scientific “e notation” is used to express E-values , e . g . , where 1e-4 represents 1×10−4 . Because this E-value is rather non-stringent , these hits were compared with similar sequences in Swiss-Prot ( downloaded November 2 , 2008 ) , a manually curated set of protein sequences known to have reliable annotations [29] , [30] . Additionally , the provisional proteases were searched against 219 profile hidden Markov models ( profile HMMs ) from Pfam [31] version 22 . 0 that corresponded to MEROPS peptidase families ( personal communication with Neil Rawlings ) using the program HMMER ( v2 . 3 . 2; trusted cutoff ) [32] . Because profile HMMs define the likelihood of finding particular amino acids in a column in a multiple sequence alignment of relevant sequences , they are helpful in identifying distantly related proteins by scoring more highly the presence of specific regions and residues important to a known family . [32] . T . brucei sequences were removed as false hits if they were similar to SwissProt sequences annotated with non-protease functions or if they matched a Pfam model to non-protease families . After removing false hits , 251 “putative proteases” remained . Predicted active proteases were identified using tools at the MEROPS website where metal-binding and active site residue ( MASR ) sites for a sequence can be predicted based on BLAST alignments to pre-computed alignments of families . Of 251 putative proteases , 127 were predicted to be active proteases and trimmed to remove non-peptidase regions by examining the results of our original BLAST searches of T . brucei sequences against pep82 and removing regions that were not included in the alignments to the MEROPS peptidase domains . This final set of predicted active protease sequences is called “Tbpep . ” Unlike T . brucei , human sequences were already well-curated in MEROPS with MASR data readily available for each sequence . Of the 958 human peptidase domain sequences from pep82 , 574 were predicted to be active according to the MASR data; this set is hereafter called “Hspep82 . ” Twenty additional human peptidase sequences were identified from the Mammalian Degradome Database ( “MDD” ) , a manually curated dataset of proteases from the Lopez-Otin group ( http://degradome . uniovi . es ) [33] that were found to be dissimilar to Hspep82 sequences but predicted to be active peptidases ( by BLAST searches against pep82 and MASR prediction similar to the procedure for T . brucei ) . These 20 sequences were trimmed to remove non-peptidase regions and added to the 574 sequences above . This final set of 594 predicted active human peptidase sequences is called the “HsLOpep82” dataset . PDB entries representing Tbpep and HsLOpep82 sequences were found by BLAST searches of the PDB protein sequences and those with good resolution ( ≤3 . 5 Å ) were kept for analysis . The 150 human and one T . brucei pdb files were trimmed to remove non-peptidase regions . This was done because a trial test showed that structure similarity detection between peptidases can be obscured if non-peptidase regions are included in structures being compared ( data not shown ) . For sequences without crystal structures , models were taken from ModBase ( http://modbase . compbio . ucsf . edu ) [34] , a large database of comparative structural models ( homology models ) created using the Modeller program [35] . Only good quality models , as determined by using the recommended cutoff of ModPipe Protein Quality Score ( MPQS ) >1 . 1 , were used in this analysis . There were 174 human models and 48 T . brucei models initially identified . Models were checked to make sure that they spanned the peptidase regions of the target sequences , and those with unacceptably short or incorrect regions were discarded . 141 human and 47 T . brucei models passed this check . These were then trimmed to exclude non-protease regions . Because modeling is computationally- and time-intensive , we created new models only for representatives of clusters of similar sequences for which no structure ( crystal structure or ModBase model ) existed . Representative sequences of clusters with ≤40% sequence identity ( “sequence ID , ” clustered with CD-HIT [36] ) that had no structures were submitted to ModWeb , a homology modeling web server that utilizes Modeller ( http://modbase . compbio . ucsf . edu ) [34] . This yielded 51 and 23 new human and T . brucei models with MPQS >1 . 1 , respectively . It has been shown that better quality models ( about 1 . 5 Å or better root mean square ( RMS ) error between template and model ) generally result from using templates with ≥30% sequence ID to the target sequence [37] . However , sequence identity alone can be misleading . Additional factors can be important indications of model quality such as how much of the template sequence aligns well to the target and whether inter-atomic distances in the model are similar to those seen in real proteins . The MPQS reported in Modeller is a composite score that includes a number of such factors . Of the models included here , 80% have ≥30% sequence ID to their templates . An updated model for the T . brucei M32 sequence ( TbM32 ) was created using the program Prime ( Prime 2 . 0208 , Maestro 8 . 5207 , Schrodinger LLC , Portland , OR ) . The structure for a T . cruzi M32 protease ( PDB code 3DWC ) [38] was used as the template for modeling because it possesses a higher sequence ID to TbM32 than the 1KA2 structure previously used for the ModBase model ( 72% vs . 33% ) . All-by-all blastp scores were computed on the 594 HsLOpep82 and 127 Tbpep sequences ( total of 721 nodes ) , and the data viewed with the network visualization program Cytoscape [39] using the “organic” layout setting , which clusters nodes more tightly if they are more highly connected; a BLAST E-value threshold of ≤1e-5 was required for drawing edges between any two nodes , resulting in 10 , 188 edges ( Figure 1 ) . This corresponds to ≥40% sequence ID for alignments ≥50 residues . Nodes are color- and shape-coded by species , MEROPS family ( according to the family of the best BLAST hit to pep82 sequences as described above ) , and structure representation . An all-by-all pairwise structure similarity comparison was performed using the program FAST [40] on the trimmed pdb files ( crystal structures , ModBase models , and ModWeb models ) for the 71 T . brucei and 342 human predicted active peptidases ( total 413 ) with structure representation . The data were visualized with Cytoscape ( “organic” layout; a threshold of normalized FAST score ( SN ) ≥4 . 5 was used to draw edges . This score is well above the minimum cutoff ( 1 . 5 ) stated by the authors of FAST to be statistically significant . Nodes were color- and shape-coded by species , MEROPS family , and structure representation . Protein structures were aligned , visualized , and root-mean-square deviation ( RMSD ) values calculated using the MatchMaker tool from the Chimera software package [41] . Default settings were used , with aligned pairs of atoms ≤2 Å included in the RMSD calculation . Alignments with similar RMSDs but with more aligned pairs within this threshold indicate higher overall structural similarity . Multiple sequence alignments ( MSAs ) were created using MUSCLE [42] and visualized with the program GCG SeqLab ( Wisconsin Package Version 10 . 3 , Accelyris Inc . , San Diego , CA ) . For the T . brucei protease of family M32 ( TbM32 ) and human angiotensin-converting enzyme 2 ( ACE2 , PDB code 1R4L ) , the structure-based sequence alignment was visualized with ESPript [43] ( http://espript . ibcp . fr/ESPript/cgi-bin/ESPript . cgi ) . To find structural relatives for the T . brucei C51 family in the PDB , the structure alignment server PDBeFold [44] at ebi ( www . ebi . ac . uk/msd-srv/ssm ) was utilized by submitting the ModWeb model for the amidase domain of trypanothione synthetase-amidase ( “TbTSAAm” ) and specifying 40% as the lowest acceptable match of secondary structure elements for both query and target . The TbM32 gene was amplified by PCR from genomic DNA with Phusion polymerase using sense ( 5′ - GCGCGCCATATGATGAAGGCATACAAAGAGCT - 3′ ) and antisense ( 5′ - ATGCATGTCGACTCAGTTGGCATCGTCACGGT -3′ ) primers . The PCR product was cloned into the pET28a expression vector ( Invitrogen , Carlsbad , CA ) and the N-terminal polyhistidine-M32 expressed in E . coli strain BL21-DE3 . The recombinant enzyme was purified in two steps: first using a Ni-NTA slurry , then further purified on a Ni column ( 5 mL HisTRAP FF column ) using an ÄKTA FPLC system ( GE Healthcare Life Sciences , Piscataway , NJ ) at 4°C from which the protein was eluted using a 0–250 mM imidazole linear gradient in 2 column volumes . Active fractions were analyzed by gel electrophoresis , and pure samples were combined , flash frozen and stored at −80°C for future use . Recombinant TbM32 ( 2 µM ) was assayed using the synthetic carboxypeptidase substrate FA {N- ( 3-[2-furyl]acryloyl ) }-Phe-Phe ( “FAFF” , BACHEM , Torrance , CA ) as substrate ( 100 µM ) in 50 mM Tris/HCl , pH 7 . 4 . Initial steady-state velocity ( “Vi” ) was determined by continuous assay for a range of substrate concentrations at 340 nm with a SpectraMax Plus platereader ( Molecular Devices , Sunnyvale , CA ) . Vi was calculated as milliunits/min using SoftMaxPro software ( Molecular Devices ) . For inhibition studies , protease and inhibitor were pre-incubated for 30 min at room temperature prior to adding substrate . Concentrations of inhibitors and controls were: 10 µM of the ACE2 inhibitor 28FII ( 3-{[1- ( 2-acetylamino-3-methyl-butyryl ) -pyrrolidin-2-yl]-hydroxy-phosphinoyl}-2-benzyl-propionic acid , active diastereoisomer , a gift from the laboratory of Vincent Dive [45] ) , 10 µM lisinopril ( ACE inhibitor; Toronto Research Chemicals , North York , Ontario , Canada ) , 100 µM 1 , 10P ( 1 , 10 phenanthroline , a divalent metal chelator , as a positive control; Sigma-Aldrich ) , and 1% DMSO ( negative control; Sigma-Aldrich , St . Louis , MO ) . Measurements were taken in triplicate and statistics were calculated using the software R v2 . 9 . 2 [46] by running ANOVA and a TukeyHSD post-hoc test .
The sequence similarity network ( Figure 1A ) reveals on a global scale the diversity of families of proteases predicted to be active in humans and T . brucei , showing the protease families that are most prevalent in both organisms as well as those that reflect the greatest differences between them . Figure 2 shows the overall distribution of peptidases by catalytic type . The network shown in Figure 1B shows that when models generated by comparative structural modeling ( homology models ) are included along with crystal structures , most of T . brucei and human sequence space is well covered by three-dimensional structures ( See Methods and the note on homology modeling below for a discussion of model quality ) . This made it feasible to create the global structure similarity network ( Figure 3 ) that clusters human and parasite proteases by similarity of 3D structures . In similarity networks , large numbers of proteins can be viewed in a visually meaningful way . The proteins are represented as nodes ( points ) and similarity scores above a statistical significance threshold cutoff are expressed as edges ( lines ) drawn between the nodes . The greater the number of interconnections among the proteins within a grouping , the closer they are drawn together . It should be noted that the placement of such clusters as they are laid out in network views such as shown in Figures 1 and 3 is done roughly by size so proximity between separate clusters does not imply relatedness . In the sections that follow , the overall patterns that emerge from the degradome landscapes in these network views are discussed , along with new hypotheses about parasite and host biology based on these comparisons . We also include a detailed structural examination of two protease groups , M32 and C51 , that are very different in sequence from human proteases and may have potential as drug targets . In the sequence similarity network in Figure 1A , nodes are color coded by assigned MEROPS protease family ( see Methods for definitions of family , clan , and catalytic type ) . There are five times as many human proteases ( 594 sequences ) as T . brucei proteases ( 127 ) , representing 71 and 37 different families , respectively . It has also been observed that in the degradome of the parasite Schistosoma mansoni , the parasite has fewer proteases representing fewer families than humans [27] , but little work has been done to address in detail why this may be a general trend , though this may involve the specialization of parasites . In general , serine , cysteine , and metallo catalytic type proteases dominate both degradome landscapes . There are no glutamic proteases predicted to be active in humans , and this is also the case in T . brucei . Half ( 35 ) of the total families ( 73 ) in the network have both parasite and human members . However , there are a large number of families in humans ( 36 ) that are missing in T . brucei , and two families ( C51 and M32 ) are specific to the parasite . Figure 2 shows the distribution of human and T . brucei peptidases by catalytic type , and Supplementary Table S1 shows the counts by family and the more remotely-related grouping of clan . Serine proteases comprise the most abundant catalytic type of proteases in humans with 202 members . Among all species , the serine protease catalytic type is known to be a large category of proteins containing a number of independently-evolved families from different clans representing a wide variety of functions [47] . In humans , the largest family in this catalytic type is the S01 ( trypsin and chymotrypsin ) family ( 115 members ) , with members that have well-known roles in digestion as well as in blood coagulation and immunity [47] . S01 is also the largest family by far of any catalytic type in humans , with the second largest family ( C19 ) having 51 members . In contrast to humans , cysteine proteases ( 48 ) predominate over the serine protease catalytic type ( 22 ) in the T . brucei degradome ( Figures 1A and 2 ) . Cysteine proteases have functions in virulence , immunoevasion , and enzyme activation in parasites and are the subject of active research [48] , [49] . There are 146 cysteine proteases in humans predicted to be active . Before the Tritryps genomes were available , previous work indicated that the majority of all proteases detected in the Tritryps were cysteine proteases [48] . However , our network using genomic data shows that metalloproteases are just as numerous as cysteine proteases both in T . brucei ( 48 ) and in humans ( 146 ) suggesting this may be a rich area for future studies . Notably , the S01 family is devoid of T . brucei sequences ( of several S01 homologs in T . brucei , all are predicted to be inactive ) . Although known to be generally quite distinct structurally , cysteine and serine proteases have mechanistic similarities due to the chemical relatedness of the active site sulfur and oxygen of cysteine and serine , respectively [49] . Both act as nucleophiles on the peptide bond , but sulfur is the better nucleophile . It may be that some ancestral C01 members were superseded by serine proteases that were somehow functionally superior for humans . Given the importance of the S01 proteases in human biology [47] , [50] , the absence of active members of this family in the parasite underscores significant differences with parasite biology , as has been noted previously [48] , [49] . The sequence similarity networks in Figure 1 use only sequence data from human and T . brucei to create the groupings shown resulting in some differences from assigned MEROPS family classifications which assigns some divergent proteases to the same family based on sequences from all known proteases from all species . For example , Figure 1 shows that some families ( e . g . , family S01 ) are composed of more than one cluster , revealing great diversity within these clusters that may be interconnected if sequences from other species are included . More divergent structure relationships are evident in the structure similarity network discussed later . Figure 1B shows the same sequence similarity network as in Figure 1A , but here the nodes are color-coded by species , and sequences with structure representations ( crystal structures or homology models ) are denoted with larger nodes . The human degradome is covered much better in terms of crystal structure than that of T . brucei: when the networks were initially constructed , there were 150 crystal structures for human proteases , but only a single crystal structure for T . brucei , the cathepsin L-like cysteine protease rhodesain ( from T . b . rhodesiense ) [51] . When good quality models are considered , coverage of the T . brucei degradome becomes comparable to that of humans . In Figure 1B , 61% of the clusters with T . brucei members have structure representation and 67% of clusters with human members also have structure representation . The inclusion of homology models in the network increases overall structure coverage by about 50% . There are 8 sequence similarity clusters with three or more members that have no structure representation ( Supplementary Table S2 ) and so may be good targets for structural characterization . The largest of these clusters are C85 ( 6 human , 2 T . brucei sequences ) and A22 ( 5 , 1 ) . Although models are not generally considered to have the same accuracy as crystal structures , they can give valuable structural information . It has been shown that a good quality homology model generally results from using a template ( crystal structure used to guide the modeling ) with ≥30% sequence ID to the target sequence ( protein to be modeled ) [37] . At this level , there is expected to be about 1 . 5 Å or better root mean square deviation ( RMSD ) between the model and the actual structure , and the fold and many of the details are likely to be accurate . Only good quality models were included in our analyses ( see Methods for details ) . We compared the model for a T . brucei cathepsin B-like protease in the network ( TbCatB ) with the crystal structure ( PDB 3HHI [52] ) that was solved for this protease after the networks were constructed , and found that the model was quite accurate . The overall structures aligned well ( RMSD = 0 . 67 Å over 214 atom pairs ) and the active site residues also aligned closely ( RMSD = 0 . 51 Å over 33 atom pairs ) . The network model for TbCatB was based on human procathepsin B ( PDB 3PBH; 48% sequence ID to the target ) . For comparison , two structures of the same protein , human ACE , solved by the same lab but bound to different inhibitors ( 1UZE and 1UZF bound to enalaprilat and captopril , respectively ) have an RMSD of 0 . 22 Å over 574 atom pairs . Figure 3 shows the structure similarity network in which human and T . brucei crystal structures and homology models are clustered by 3D structure similarity . Figure 3A is colored by MEROPS family , and 3B by structure ( crystal structure or model ) . There are fewer clusters than in the sequence similarity network , not only because some sequences lack structure representation , but also because some divergent sequences that are not connected in the sequence similarity network share similar structures ( and are connected with edges in the structure similarity network ) . This reflects the phenomenon that protein structure evolves more slowly than sequence [53]–[55] . Figure 3A shows that a number of structure similarity clusters have members from more than one MEROPs family , underscoring the sequence divergence within clusters . Structure similarity is often used as evidence , along with functional similarity , that proteins with divergent sequences are evolutionarily related ( i . e . , are homologs ) [56]–[58] . Most of the clusters in Figure 3 are composed of proteins that share the same catalytic type , providing further support that these proteins are homologs . Figure 4 shows the structure similarity network colored by clan . As described in Methods , the clan is the highest level in the MEROPS classification system where proteins are still considered to be evolutionarily related , and the first letter of the clan name represents the catalytic type of a clan's member families . However , as can be seen in Figure 3A , two of the clusters are of mixed catalytic type . The first cluster includes families C44 , T01 , T02 , and T03 and , consistent with the structure similarity grouping , these have been assigned to one MEROPS clan ( clan PB , Figure 4 ) because they share a common fold and conserved position of the nucleophile , even though the nature of the nucleophile in each family is different [20] . The second mixed cluster ( Figure 3A ) contains families M14 , M17 , M20 , M28 , and C15 . Unlike the first cluster discussed above , these families are assigned to different MEROPS clans ( Figure 4 ) : MC ( M14 ) , MF ( M17 ) , MH ( M20 and M28 ) , and CF ( C15 ) . This is based on differences in catalytic mechanism and non-conserved locations of metal-binding residues [20] . Structural similarity between members of these families has been detected by others and is annotated accordingly in the SCOP structural classification database [59] , but opinions differ whether they are evolutionarily related [20] , [60] . Strikingly , this is the only cluster in the network that has mixed clans ( Figure 4 ) . Viewed at the same level of structure similarity , all other clusters are composed of single clans . In fact , two other unmixed clans ( CA and MA ) are even more structurally divergent , each emerging as multiple , separate clusters ( Figure 4 ) . It is intriguing that the scaffold for this second group of mixed catalytic type shows such variation in catalytic residues and metal-binding positions while sharing similar function . While this question has previously been probed by others , it would be interesting to address this again using the broader context provided by new genomic data . One advantage of global views of degradome relationships among species is the ease with which potentially important species differences and similarities can be identified for further investigation . As indicated in Figure 1A , the sequences of two T . brucei protease families , M32 and C51 , with one and five members , respectively , are quite distant from those of any human proteases . Both M32 and C51 families are known to occur in prokaryotes and parasitic protists [61] , [62] . However , as shown in Figure 3A , despite its distance in sequence space from human proteases , the T . brucei M32 singleton ( TbM32 ) has several relatively close human structural neighbors . In contrast , the C51 cluster has none . In the sequence similarity network , TbM32 ( Tb_proteins ID Tb11 . 02 . 0100 , GI:71754837 ) fails to show a statistically significant BLAST match to any other protease ( the best match has an E-value = 0 . 62 ) . However , as seen in the structure similarity network ( Figure 3A ) , the homology model for TbM32 ( “TbM32m” ) has several close human structural neighbors: Angiotensin I-converting enzyme ( ACE ) , ACE2 , Neurolysin , Thimet oligopeptidase ( TOP ) , and Mitochondrial intermediate peptidase . RMSDs of human crystal structures aligned with TbM32m range from 1 . 217 Å to 1 . 283 Å with number of aligned alpha carbon pairs ranging from 73 to 119 . TbM32m was created using as a template the crystal structure of an M32 protease from T . cruzi ( designated here as “TcM32 , ” PDB code 3DWC ) . TcM32 is a metallocarboxyeptidase: it cleaves one amino acid from the C-terminus of a peptide and requires a metal ion for activity [61] . Because of its high sequence identity to TcM32 ( 72% ) and the good alignment between predicted active site residues to those of other M32 proteases , we predicted that TbM32 was also likely to be a metallocarboxypeptidase . The best characterized human structural neighbor to TbM32m is the anti-hypertensive drug target ACE; however , its lesser known homolog , ACE2 , is the only human peptidase in the cluster that is a carboxypeptidase . ACE is a dipeptidyl peptidase , cleaving two residues from the end of a peptide . The presence of an arginine ( R273 ) in the ACE2 S1 binding pocket , instead of the corresponding Gln in ACE , creates a smaller pocket in ACE2 that allows only one residue to fit into the active site C-terminal to the cleavage point . This difference also helps to rationalize the high selectivity of inhibitors for ACE relative to ACE2 [63] . ACE and ACE2 have roles in regulating blood pressure [64] and cardiac function [65] , respectively . A number of inhibitors have been designed for ACE [66] , and a few have been also developed for ACE2 [45] , [67] . Figure 5 ( inset ) shows the overall structural similarity between TbM32m and ACE2 . We created a structural alignment of TbM32m and the crystal structure for ACE2 bound with inhibitor MLN-4760 ( PDB code 1R4L ) [63] , which showed that a TbM32m arginine ( R348 ) likely corresponds to ACE2 R273 ( Figure 5 ) . However R348 is somewhat receded , resulting in more space in the binding pocket in TbM32m . In the TcM32 crystal structure , an arginine superimposes closely with TbM32m R348 , resulting in a similar binding pocket space . It is known that ACE inhibitors do not bind to TcM32 ( [38] and personal communication with JJ Cazzulo ) despite an apparently slightly larger pocket than ACE2 . Some insight may be given towards understanding this by the knowledge that ACE2 and some other metallopeptidases undergo significant “hinge closure” upon binding a ligand [63] . Model TbM32m was built using apoenzyme TcM32 as a template and it seems feasible that both TbM32 and TcM32 could also undergo hinge closure upon ligand binding , thereby leading to a smaller binding pocket and consequent inability to bind ACE inhibitors . Visual inspection of the structural alignment of TbM32m and ACE2 showed similar binding pocket shapes and no steric clashes of TbM32m residues with the superimposed ACE2 inhibitor , leading to a further prediction that ACE2 inhibitors might bind TbM32 . To test these two computational predictions , we first cloned and expressed TbM32 and showed that it cleaves the synthetic carboxypeptidase substrate FA-Phe-Phe . Further , it is inhibited by the metal chelator 1 , 10P ( 1 , 10 Phenanthroline ) . This is consistent with a metallocarboxypeptidase function ( Figure 6 ) . We then assayed the recombinant TbM32 with ACE2 inhibitor 28FII , a phosphinic peptide that mimics the transition state structure of ACE2 substrates [45] ( MLN-4760 is no longer available ) . The ACE2 inhibitor produced statistically significant inhibition of TbM32 whereas the ACE inhibitor lisinopril did not ( Figure 6 ) . The IC50 of 28FII with human ACE2 has not been published , but its Ki is low ( 0 . 13 nM ) [45] , suggesting its potential as an inhibitor . For comparison , the compound MLN-4760 has an IC50 of 0 . 44 nM with human ACE2 [63] . Our results show that significant inhibition of TbM32 by an ACE2 inhibitor occurs at 10 µM; while this level is higher than the IC50 of an inhibitor designed for ACE2 when used with ACE2 , these preliminary results suggest that ACE2 inhibitors may be worth consideration as lead compounds for further development against TbM32 . Although TbM32m and ACE2 are highly similar in overall structure , the α-carbon of the TbM32m R273 originates from a different secondary structure element and has a different topology than ACE2 ( Supplementary Figure S1 ) . The sequence identity between these two proteins is <10% so that structural data was needed to give insight into ligand specificity in these highly divergent proteins . The T . brucei C51 family has five members , with pairwise sequence identity among them ranging from 27%–96% . This cluster is remote from human proteases both in terms of sequence and structure . One of the five sequences has been previously identified as T . brucei trypanothione synthetase-amidase ( “TbTSA” , Tb_proteins ID Tb927 . 2 . 4370 , GI:84043680 ) and is the only protein in the cluster that has been experimentally characterized . TbTSA has a C-terminal synthetase domain that catalyzes the production of trypanothione ( TSH ) from two glutathione ( GSH ) and one spermidine ( Spd ) molecule . Its N-terminal amidase ( C51 ) domain catalyzes the reverse reactions [68] . The biological role of the amidase domain is not completely clear , but it likely plays a role in maintaining a concentration balance between these compounds [69] . Unlike TbTSA , the other T . brucei C51 sequences have only the amidase domain; a multiple sequence alignment of all five sequences ( not shown ) indicates that the active site residues predicted to be associated with peptidase/amidase activity are well aligned . We modeled all five amidase domains using as the template the TSA from L . major ( “LmTSA” , PDB code 2VOB ) , two of which are represented in Figure 3A ( see Methods ) . The amidase domain of TbTSA ( “TbTSAA” ) has 58% sequence ID to LmTSA . A structural alignment of each T . brucei C51 model to the model of TbTSAA ( “TbTSAAm” ) ( not shown ) shows conservation of residues near the active site ( an average of 9 of 17 selected pairs were strictly conserved ) , suggesting that one or more of the uncharacterized C51s may have amidase activity . RMSDs of overall alignments of the T . brucei C51 models to TbTSAAm ranged from 0 . 282–0 . 586 Å with number of aligned alpha carbon pairs ranging from 95–122 . GSH serves in anti-oxidant and detoxification roles in most animals and plants [70] . However , trypanothione ( TSH ) serves this purpose in trypanosomes [71] and does not occur in humans . Experiments by others have suggested TbTSA has promise as a drug target [69] , [72] , [73] . One gene knockout study suggested it is the trypanothione synthetase domain and not the amidase domain that is essential to the parasite; however , this study also showed that both domains are important for parasite virulence [69] . It may be that both domains , perhaps in tandem , are worthy of consideration as drug targets due to their physical connection as a two-domain protein and the biochemical relationship in their roles , i . e . , synthesis and hydrolysis of trypanothione . For example , a trypanothione-like inhibitor that can bind both domains may be worth consideration . Additionally , the amidase domain's distinctive structure among prokaryotes and parasitic protozoa , but not in humans [62] , makes it an intriguing subject for other cross-species comparisons and for exploring possible drug targets in non-trypanosomes . The closest human structure neighbor to TbTSAAm is cathepsin F ( “CatF” , PDB 1M6D ) ( FAST SN score = 2 . 6 , about 5% sequence ID ) . Cathepsins are well-studied cysteine proteases in the C01 “papain” family , with roles that range from general protein degradation to wound healing [74] . A number of inhibitors have been developed for this family [49] . The alignment of TbTSAAm with 1M6D , which is complexed with a vinyl sulfone inhibitor ( 4-morpholin-4-yl-piperidine-1-carboxylic acid [1- ( 3-benzenesulfonyl-1-propyl-allycarbamoyl ) -2-phenylethyl]-amide ) , shows some general , overall structure similarity ( RMSD = 1 . 1 Å over 16 alpha carbon pairs ) , but also some striking differences ( Figure 7 ) . Most importantly , we predict that a helix exists near the binding site in TbTSAA that could obstruct the binding of a cathepsin-like inhibitor , a feature that is absent in CatF ( 1M6D ) . Upon searching the PDB , the most similar structure to TbTSAAm co-crystallized with a ligand was found to be the amidase domain of E . coli glutathionylspermidine ( Gsp ) synthetase/amidase ( “EcGspSAA” , PDB 3A2Y , 36% sequence ID to TbTSAA ) . Superposition of TbTSAAm with EcGspSAA ( RMSD = 0 . 85 Å over 120 alpha carbon pairs ) and human CatF showed that EcGspSAA has a binding site helix similar to the one in TbTSAAm . The TSH-related Gsp binds in a different orientation and location than the protease inhibitor binds human CatF . These observations suggest that such differences in architecture may allow the design of TbTSAA-specific inhibitors that would not cross-react with human C01 peptidases .
The explosion of data in sequence and structure databases in recent years , along with advances in modeling technology , presents researchers with the opportunity for creating more global views of sequence and structure space from whole organisms than has been possible previously . It has been estimated that sufficient structural templates exist for modeling about 50% of all known proteins [75] . However , leveraging existing data and synthesizing the information into a form that is interpretable in an intuitive and accurate way can be challenging . We constructed networks presenting the first global views of the degradome landscapes of the parasite , T . brucei and its human host , allowing a side-by-side comparison of sequence and structure similarity of predicted active proteases between the two species . The networks show patterns of abundance and variety of proteases while highlighting sequence clusters in which structures are sparse and may be of higher priority for solving new structures . These networks also give clues as to how divergent proteins might be related . In addition , such global views can give insights about potential drug targets . Our results suggest that ACE2 inhibitors might serve as lead compounds for inhibitor development against TbM32 . Also , we predict that uncharacterized C51 members may have an amidase function and that structural differences relative to human peptidases may make it possible to design specific inhibitors for this family of parasite proteins . Studies such as these should prove more useful as databases of sequence , structure , and function continue to grow and species-specific proteomes become more complete . The networks are available for download and can be viewed and manipulated interactively using the freely available program Cytoscape ( www . cytoscape . org ) . | Human African trypanosomiasis ( HAT ) is caused by the protozoan parasite Trypanosoma brucei . HAT is fatal unless treated , yet the current treatment itself can cause death . New treatments are urgently needed . Our study focuses on proteases , which are enzymes that break down proteins . Because of their roles in many centrally important biological processes , proteases are targets for drugs to treat a variety of diseases including parasite infection . The recent explosion of protein sequence and structure information in public databases has made surveys of proteins on a genomic scale possible . However , collecting specific data of interest from diverse databases and synthesizing them in a way that is easy to interpret can be difficult . We used T . brucei and human protease sequences , crystal structures , and models to create network views that show how proteases cluster by similarity . Such views are valuable not only for understanding the evolution of the protein repertoire in each species , but also can give important clues for drug design . Two T . brucei protease groups ( “M32” and “C51” ) that are very different in sequence from human proteases were examined in structural detail . Based on our analyses , a human ACE2 inhibitor was selected for experimental testing on one of these parasite proteases , TbM32 , and was shown to inhibit it . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"and",
"Discussion",
"Conclusions"
] | [
"genomics",
"biology",
"computational",
"biology",
"structural",
"genomics",
"genetics",
"and",
"genomics"
] | 2012 | A Global Comparison of the Human and T. brucei Degradomes Gives Insights about Possible Parasite Drug Targets |
The burden of leptospirosis in humans and animals in Africa is higher than that reported from other parts of the world . However , the disease is not routinely diagnosed in the continent . One of major factors limiting diagnosis is the poor availability of live isolates of locally circulating Leptospira serovars for inclusion in the antigen panel of the gold standard microscopic agglutination test ( MAT ) for detecting antibodies against leptospirosis . To gain insight in Leptospira serovars and their natural hosts occurring in Tanzania , concomitantly enabling the improvement of the MAT by inclusion of fresh local isolates , a total of 52 Leptospira isolates were obtained from fresh urine and kidney homogenates , collected between 1996 and 2006 from small mammals , cattle and pigs . Isolates were identified by serogrouping , cross agglutination absorption test ( CAAT ) , and molecular typing . Common Leptospira serovars with their respective animal hosts were: Sokoine ( cattle and rodents ) ; Kenya ( rodents and shrews ) ; Mwogolo ( rodents ) ; Lora ( rodents ) ; Qunjian ( rodent ) ; serogroup Grippotyphosa ( cattle ) ; and an unknown serogroup from pigs . Inclusion of local serovars particularly serovar Sokoine in MAT revealed a 10-fold increase in leptospirosis prevalence in Tanzania from 1 . 9% to 16 . 9% in rodents and 0 . 26% to 10 . 75% in humans . This indicates that local serovars are useful for diagnosis of human and animal leptospirosis in Tanzania and other African countries .
Leptospirosis is an understudied zoonotic disease in Tanzania and across Africa . Limited reports show a high prevalence of leptospirosis in animals and humans with Africa presenting a major burden globally [1] . The highest median annual incidence of leptospirosis is in Africa standing at 95 . 5 per 100 , 000 people . Africa is followed by Western Pacific ( 66 . 4 ) , the Americas ( 12 . 5 ) , South-East Asia ( 4 . 8 ) and Europe ( 0 . 5 ) [2] . In Africa , leptospirosis has been reported in almost all geographic zones . Despite its high burden , leptospirosis is not routinely diagnosed in African hospitals . Awareness of this disease is also generally lacking among health providers , medical personnel and the general public including high-risk populations such as abattoir workers and animal handlers ( Mgode , personal observation ) . One of the major factors limiting diagnosis of leptospirosis in Africa is the need for leptospire isolation to discover the local circulating serovars that are needed for inclusion in the microscopic agglutination test ( MAT ) . MAT requires various live Leptospira serovars as antigens to detect infections caused by different serovars belonging to different serogroups [3 , 4] . Apart from serological studies , there are relatively few studies on isolation and identification of Leptospira pathogens in Africa . However , a number of Leptospira serovars have been isolated in some African countries , namely South Africa [5 , 6] , Zimbabwe [7 , 8] , Congo DRC [9] , Kenya [10 , 11] , Madagascar and other Indian ocean Islands [12 , 13] , Nigeria and Ghana [9 , 14 , 15] , Egypt [16] and Tanzania . In Tanzania , leptospirosis is widely reported in wild small mammals , domestic animals and humans [17–22] . Despite these reports , awareness of this disease is still lacking and there is an urgent quest for gathering sufficient data on leptospirosis for promoting awareness . The objectives of this study were , therefore , to determine Leptospira serovars occurring in Tanzania and their host animals . This knowledge will help to rationally design control and prevention measures and contribute to an improved MAT for diagnostic and prevalence study purposes in Tanzania and potentially other East African countries .
Cattle brought at Morogoro municipal abattoir and pigs slaughtered at two main informal slaughterhouse sites also in Morogoro were randomly sampled for this study . Sheep , goats , dogs and cats were also sampled from different localities in Morogoro . Trapping of rodent species and insectivore shrews was stratified whereby rodents were trapped in selected localities representing all geographic areas of Morogoro town including peri-urban areas , urban areas , inside houses , around houses ( outdoors ) , fallow land , swampy areas , inside and around markets , inside and around the abattoir . Sherman live traps baited with peanut butter mixed with maize bran were used to trap small rodents and insectivores ( shrew ) species . Larger rodents particularly the African giant pouched rats were trapped in same localities as small rodents using Havahart traps baited with fresh maize cobs . The traps were set for three consecutive nights at each site . All captured animals were identified to genus level and geographic coordinates of the collection site were recorded using GPS . Blood samples for serological determination of leptospirosis by microscopic agglutination test were obtained from patients providing blood in various hospitals in Morogoro for other tests such as diagnosis of typhoid fever . Participants were orally informed that their samples would be anonymously tested for leptospirosis . 3–4 ml of blood was obtained and a drop was inoculated into fresh medium for isolation of leptospires . The remaining blood was centrifuged to obtain serum for serological test . Urine from abattoir workers was collected into sterile universal bottles and transported to the laboratory for inoculation of drops into sterile EMJH medium . The isolation of leptospires from animal hosts in Tanzania started in 1996 and is ongoing . Isolation and identification of leptospires have been carried out in humans , domestic animals including cattle and pigs , and feral and ( semi ) domestic small mammals collected in natural landscapes , agricultural fields , in rural settlements and in urban areas . Wild and domestic animal collection and handling followed the guidelines of the American Society of Mammalogists [23] . Urine specimens were aseptically collected from animals anesthetised using di-ethyl ether . The urine was collected using sterile syringes and needles . Urine sampling was also done from cattle at slaughterhouse whereby the urinary bladder were taken out slaughtered animals and the neck of the bladder was tightly closed with fishing line ( thread ) to prevent spillage of urine during transportation to the laboratory . A drop of fresh urine was aseptically taken from the bladder using sterile syringe and needle and inoculated into a tube containing sterile Leptospira Ellinghausen and McCullough , modified by Johnson and Harris ( EMJH ) culture medium containing 5-Fluorouracil selective inhibitor . Kidney specimens were obtained after swabbing the sacrificed animal abdomen with 70% ethanol and dissecting using pair of sterile scissors and forceps . Smaller kidneys were put into sterile glass tube containing sterile phosphate buffered saline ( pH 7 . 0 ) and homogenised using sterile glass rods and or sterile glass Pasteur pipettes . Larger kidneys were macerated to obtain cross-sectional pieces which were homogenized . A drop of kidney homogenate was aseptically inoculated into EMJH medium as previously described [24] . Cultures were incubated at 30°C for up to 8 weeks and examined weekly for Leptospira growth . Isolation of leptospires from fish species was not conducted in the present study whereas serological testing using putative prevalent serovars indicated potential to provide first insight on leptospirosis in fishes from this area . Initially , five Leptospira isolates which were the first isolates from Tanzania of which three were from cattle and two from rodents were identified and enabled identification of many other isolates among the 52 reported in this study . Leptospira isolates SH9 and SH25 from the African giant pouched rats ( Cricetomys sp . ) and RM1 from cattle were subjected to standard taxonomical analyses recommended by the International Committee on Systematics of Prokaryotes: Subcommittee on the Taxonomy of Leptospiraceae . These were identified as serovar Kenya [18] and serovar Sokoine [19] . This enabled assigning other serologically and genetically identical isolates among the 52 isolates to serovar Kenya and Sokoine [25] . Isolates coded RM4 and RM7 were subjected to serological typing using serogrouping using reference rabbit serum and monoclonal antibodies . Multilocus sequence typing was also employed to identify sequent isolates from wild rodents and insectivores ( Crocidura spp . ) from Tanzania through determination of their genetic relatedness with known serovars . This enhanced assigning subsequent isolates to serovars Mwogolo , Lora , Kenya , Qunjian ( Canicola ) and Sokoine [25] . The typical identification procedures were as follows: Selected Leptospira isolates coded RM1 , RM4 and RM7 from cattle and SH9 and SH25 from the African giant pouched rats ( Cricetomys sp . ) were grown in EMJH medium for 5–7 days at 30°C . Fully-grown cultures with density of 3x108 leptospires per ml were checked for purity under-darkfield microscopy before injecting into pair of healthy and leptospirosis-free animals to produce antiserum for serological identification of the isolates [26] . Animals were handled in compliance with the “Animal Research: Reporting In Vivo Experiments” ( ARRIVE guidelines ) and the Helsinki Declaration [27 , 28] . The isolates were thereafter reacted with reference rabbit antiserum for common Leptospira serogroups . MAT was performed as previously described [3] . Findings of this prelinanry reaction were useful in selecting reacting serogroups for further identification of new isolates by monoclonal antibodies and cross-agglutination absorption test ( CAAT ) which uses antiserum produced using unknown ( new isolates ) and reference serovars as described below . MAT sets of specific monoclonal antibodies belonging to candidate Leptospira serogroups were used to determine isolate affiliation . Two isolates coded RM4 and RM7 were subjected to this approach employing the following monoclonal antibodies for serogroup Grippotyphosa: F71C2-4 , F71C3-3 , F71C9-4 , F71C13-4 , F71C16-6 , F71C17-5 , F164C1-1 , F165C1-4 , F165C2-1 , F165C3-4 , F165C7-5 , F165C8-3 and F165C12-4 . Other monoclonal antibodies used were of serovar Butembo serogroup Autumnalis: F43C9-5 , F46C1-1 , F46C2-4 , F46C4-1 , F46C5-1 , F46C9-1 , F46C10-1 , F48C1-3 , F48C3-3 , F48C6-4 , F58C1-2 , F58C2-3 and F61C7-1 . The monoclonal antibodies were provided by the WHO/FAO/OIE Collaborating Centre for Reference and Research on Leptospirosis , at the Royal Tropical Institute , Amsterdam , The Netherlands , which offers varieties of reference leptospirosis research materials worldwide [29] . Leptospira isolates coded SH9 and SH25 were also subjected to DNA fingerprinting described by Zuerner and Bolin [30 , 31] and had identical DNA pattern of serovar Kenya [18] . Isolates coded TE 0826 and TE 0845 have been previously identified using multilocus sequence typing as identical to serovar Mwogolo serogroup Icterohaemorrhagiae; isolates TE 1992 , TE 2324 , TE 2364 and TE 2366 as serovar Lora serogroup Australis; and isolate coded TE 2980 as serovar Qunjian [25] . Additionally , 3 isolates were previously identified using multilocus sequence typing as serovar Sokoine , and 11 isolates as serovar Kenya [25] . Further identification of Leptospira isolates was achieved using the gold standard test for identification and taxonomy of Leptospira serovars known as cross agglutination absorption test ( CAAT ) . Briefly , isolates are identified as different serovars if more than 10% of homologous titre remains in one of the test isolates after cross-absorption with sufficient amount of heterologous antigen . This means that 0–10% difference in antibodies remaining following absorption represents strains belonging to same serovar [26] . Leptospira isolate is grown into EMJH medium to density of 3x108 leptospires per ml for inoculation into laboratory rabbits to produce antibodies ( antisera ) against the isolate . Antiserum is harvested from rabbit and a titre of 1:5120 is determined by MAT using formalin killed homologous isolate . Antiserum with higher titre than 1:5120 is diluted with phosphate buffered saline . The test isolate cultured for 5–7 days in EMJH medium reaching a density of 3x108 leptospires per ml is killed by mixing with formalin ( final concentration of 0 . 5% ) and incubating at room temperature for 1 hour . Formalized culture is divided into 5 , 10 and 20 ml and centrifuged in refrigerated centrifuge at 10 , 000 rpm x g for 30 minutes to obtained sediment . The culture sediment is air dried and resuspended into PBS-formalin diluted antiserum . The suspension is incubated at 30°C overnight and thereafter centrifuged at 10 , 000 rpm x g for 30 minutes to obtain the supernatant which is the now the absorbed serum for MAT checking for absorption levels ( under or over absorption ) using live and killed antigen similar to the absorbing antigen . The supernatant with titre of 1:40–1:80 that is lower than 1% of homologous titre for the same antiserum will be chosen for control MAT with unabsorbed diluted reference antiserum with live and killed homologous antigen; and MAT with absorbed serum with live and killed homologous reference antigen . Subsequently , the produced antiserum against unknown isolate is absorbed with all positive Leptospira serovars to determine titres remaining after absorption . Detailed procedures of cross-agglutination absorption test have been described by Hartskeerl and co-workers [32] . Leptospira isolates coded SH9 and SH25 from the African giant pouched rats and isolate coded RM1 from cattle in Tanzania were identified as serovar Kenya and serovar Sokoine by this method [18 , 19] . The reference Leptospira serovar Sokoine strain RM1 which is a new serovar described in 2006 has been deposited in Leptospira culture collection at the WHO/FAO/OIE Collaborating Centre for Reference and Research on Leptospirosis of the Royal Tropical Institute , Amsterdam , The Netherlands , and the WHO Collaboration Centre for Diagnosis , Reference , Research and Training in Leptospirosis , Port Blair , Andaman and Nicobar Islands , India [19] . Serovar Kenya strain SH9 and SH25 and as well as serovars Grippotyphosa strain RM4 and RM7 , Mwogolo , Lora and Canicola reported in this study have also been deposited in the Leptospira culture collection of the Royal Tropical Institute , Amsterdam , The Netherlands . The isolates are also maintained in the culture collection of the Pest Management Centre , Sokoine University of Agriculture , Morogoro , Tanzania . Infection with local Leptospira serogroups and putative serovars in animals and humans was deduced serologically using MAT on blood samples from cattle , rodents , bats , fish and humans . Local Leptospira serovars included in the MAT were Sokoine , Kenya , Lora , Canicola and Grippotyphosa , and reference serovars were Hebdomadis , Pomona and Hardjo . The Ethical Review Board of Sokoine University of Agriculture approved use of animals . The Tanzania Commission for Science and Technology ( COSTECH ) granted the research permit for use of wild animals ( permit no . 2013-260-NA-2014-110 ) . Infected animals were also handled in compliance with the “Animal Research: Reporting In Vivo Experiments” ( ARRIVE guidelines ) and the Helsinki Declaration [27 , 28] . Wild and domestic animal collecting and handling followed the guidelines of the American Society of Mammalogists [23] . Anonymous human participants gave oral consent allowing anonymous screening for leptospirosis and determination of the prevalence . These individuals were those seeking diagnosis of other diseases in the blood at hospitals and abattoir workers .
A total of 52 Leptospira isolates were obtained from urine and kidneys of different animal species in Tanzania . The isolation rate of leptospires from different animal hosts ranged from 0 . 6% in the rodent Mastomys natalensis to 8 . 4% in the African giant pouched rat ( Cricetomys sp . ) ( Table 1 ) . Four Leptospira isolates were obtained from 589 human urine samples that is 0 . 67% isolation success rate . Eighty-three of the 589 human urine samples were from abattoir workers handling and slaughtering cattle at Morogoro municipal abattoir . Three of the 83 ( 3 . 6% ) abattoir workers were culture positive . Abattoir workers thus yielded more positive cultures than participants sampled in hospitals ( n = 506 ) from which only one isolate was obtained from an individual living at Kidodi village in the Kilombero river valley that is 150 km away from Morogoro municipality where three isolates were obtained in abattoir workers . The overall isolation rate from both abattoir workers and individuals sampled in hospitals was low ( 0 . 19% ) with four isolates from 589 individuals . Culturing from fresh 358 blood samples collected in hospitals was negative . The positive cultures from human urine were contaminated and isolates died before carrying out identification . Common Leptospira serovars with their respective animal hosts were: Sokoine ( cattle and rodents ) ; Kenya ( rodents and shrews ) ; Mwogolo ( rodents ) ; Lora ( rodents ) ; Qunjian ( Canicola ) ( rodent ) ; serogroup Grippotyphosa ( cattle ) ; and an unknown serogroup or serovar from pigs ( Table 2 ) , which is not related to common serogroups or serovars currently reported from Tanzania . The majority of the 52 Leptospira isolated obtained in this study were from the African giant pouched rat ( Cricetomys sp . ) which yielded 24 isolates out of 285 specimens ( 8 . 42% ) and shrews with 11 isolates from 298 specimens ( 3 . 7% ) . Eight isolates were obtained from 1382 Mastomys natalensis rats ( 0 . 6% ) , whereas isolation from 1021 cattle and 236 pigs sampled in cattle slaughterhouse and informal pig slaughterhouses yielded 8 isolates ( 0 . 68% ) and 2 isolates ( 0 . 6% ) , respectively . Three isolates were obtained from Rattus rattus but cultures were contaminated before performing and preliminary identification hence they are not presented in this work . No isolates were obtained from Lemniscomys spp . , Tatera spp . , Mus spp . collected in Morogoro municipality . The pattern of Leptospira isolation rates for different animal species is shown in Fig 1 . The preliminary serogrouping of the isolates showed that the RM1 isolate reacts with serum for serogroups Canicola , Icterohaemorrhagiae and Sarmin . The titres for serogroup Icterohaemorrhagiae were higher than those of other serogroups ( Table 3 ) . Isolates RM4 and RM7 reacted with rabbit serum for serovar Butembo serogroup Autumnalis as well as with serovar Grippotyphosa and Huanuco of serogroup Grippotyphosa , and to some extent , with serovar Djasiman serogroup Djasiman . Isolates coded SH9 and SH25 reacted with rabbit serum for serovars Kenya and Ballum of serogroup Ballum , and serovar Poi serogroup Javanica ( Table 3 ) . Leptospira isolates assigned to tentative serogroups in preliminary serogrouping with reference rabbit serum were further analysed using specific monoclonal antibodies for those serogroups . Isolates coded RM4 and RM7 did not react with monoclonal antibodies for serogroup Autumnalis , virtually excluding that they belonged to this serogroup . However , they reacted with 11 of the 13 monoclonal antibodies for Grippotyphosa ( Fig 2 ) , supporting that they belong to this serogroup ( Table 4 ) . Reaction profiles of RM4 and RM7 were most comparable with that of reference Leptospira serovar Grippotyphosa serogroup Grippotyphosa . Serological tests involving eight commonly occurring Leptospira serovars ( Table 5 ) in the Morogoro region revealed variations in occurrence and prevalence of Leptospira serovars in wild and domestic mammals , fish and humans . Some Leptospira serovars were found often in more than one animal species whereas other serovars were detected in relatively few animal species . Serovars found in different animal species were serovar Kenya and serovar Sokoine detected in domestic mammals , rodents , bats , fish and humans ( Table 5 ) . The seroprevalence of different Leptospira serovars in different animal species shows varying seropositivity which indicate that these serovars can be used in subsequent MAT for diagnosis of leptospirosis in animal species reported hereunder ( Fig 3 ) . An over 10-fold increase in seroprevalence of leptospirosis in humans , wild and domestic mammals in Tanzania was observed following the use of local serovars such as serovar Sokoine in MAT compared to the reference serovar Icterohaemorrhagiae antigen used in the first study of leptospirosis in animals and humans in Tanzania [17] . These observed increases in rodents went from 1 . 9% to 16 . 9% , in humans from 0 . 26% to 10 . 75% and in dogs from 37% to 39% . There was also an increase in seroprevalence in MAT with the local Grippotyphosa antigen compared to the reference serovar Grippotyphosa as follows: in rodents from 0% to 2% , in dogs from 0% to 10% , and slight increase was observed in humans from 0 . 26% to 0 . 5% . Use of the local serovar Kenya in the MAT of goats , sheep , pigs and dogs also yielded higher prevalence ( Table 4 ) , which unfortunately lack comparison since the reference serovar Kenya was not tested in Tanzanian animals . Fifty-three ( 75 . 7% ) of the 70 seropositive humans and 31 ( 62% ) of the 50 seropositive animals had lower antibody levels ( 1:20–1:80 ) against eight tested Leptospira serovars . These titres are below the cut-off point of 1:160 adopted from European setting which has been in use in Tanzania for past two decades due to absence of intensive leptospirosis studies determining cut-off point for Africa . Serovar Sokoine was the most reacting serovar in both humans and animals and contributed to 62 . 3% ( 33 out of 53 ) and 38 . 7% ( 12 out of 31 ) of the seropositivity with lower titres . Majority of the higher titres in 10 out of 17 humans ( 58 . 8% ) and 13 out of 19 ( 68 . 4% ) rodents was due to serovar Sokoine . This was followed by serovar Hardjo that contributed to 15% of lower titres and 23 . 5% ( 4 out of 17 ) of the higher titres in humans . Serovar Qunjian ( Canicola ) contributed to 29% ( 9 out of 31 ) lower titres and 21% ( 4 out of 19 ) of the higher titres in rodents . Serovar Grippotyphosa also contributed to 29% ( 9 out 31 ) lower titres in rodents and 5% ( 1 out of 19 ) of the higher titres in these animals . The following Leptospira serovars ( Table 6 ) belonging to different species and serogroups yields high positivity when used in MAT of humans and broad range of animal species . Although not exhaustive the listed representative serovars cross reacts with at least 10 other serovars previously reported in central and eastern Africa , Indian ocean islands located near eastern Africa coast , and elsewhere . For example , serovar Sokoine that is widespread in various domestic and wild animal species in Tanzania ( Table 5 ) reacts with serovars Mwogolo , Ndahambukuje and Ndambari found in Congo DRC in central Africa [9] , and reacts also with serovars Copenhageni and Icterohaemorrhagiae all belonging to serogroup Icterohaemorrhagiae . Serovar Sokoine again reacts with serovars Weaveri and Rio of serogroup Sarmin as demonstrated in MAT for preliminary serogrouping ( Table 3 ) . This cross reactivity broadens detection of leptospirosis caused by these serovars in MAT using serovar Sokoine . Serovar Kenya widely found in rodents in Tanzania was first described in neighbouring Kenya [10] . Its wide distribution in eastern Africa region makes it a good candidate for inclusion in MAT in this region . Serovar Grippotyphosa isolated in cattle in Tanzania cross-reacts with serovar Butembo found in Congo DRC in central Africa which widens the geographic region that this serovar can be applied in MAT . Isolation of additional serovars from Africa is much needed to establish a comprehensive region-specific Leptospira serovars for successful leptospirosis diagnosis in African region . The proportion of prevalence of the Leptospira serovars recommended for diagnosis of leptospirosis in humans and different animal species is indicated in Fig 3 . Leptospira serovar Pomona was absent in African giant rats , shrews ( Crocidura spp . ) and bats . Serovar Hebdomadis was not detected in bats and fish whereas Canicola was negative in shrews ( Crocidura spp . ) and bats . Serovar Hardjo was negative in rodent species and Grippotyphosa was not detected in shrew species . Leptospira serovars , which did no react with tested animal species , are not presented in Fig 3 below .
This study shows that diverse Leptospira serovars occur in a wide range of wild and domestic mammal species , fish and humans in Tanzania . The local Leptospira serovar Sokoine and serovar Kenya were the predominant serovars found in many vertebrate species , including fish and humans . Serovar Sokoine for example has been isolated from cattle , rodents ( Mastomys natalensis and Cricetomys sp . ) and shrews ( Crocidura sp . ) . This serovar has also been detected serologically in humans , cattle , rodents , shrews , bats and fish [20 , 21] . The distribution pattern of Leptospira in different animals species determined by their isolation rate and seroprevalence indicates that certain rodent species harbour leptospires more often than other species . For example , Mastomys natalensis , which are the most abundant rodent species in sub-Saharan Africa [36] yielded few Leptospira isolates and had low seroprevalence compared to the African giant pouched rat ( Cricetomys sp . ) and shrews ( Crocidura sp . ) , which are also widely distributed across Africa . This lower infection rate in Mastomys natalensis could be due to differences in habitats where these species are found , differences in foraging behaviours and/or physiology . For example , Cricetomys sp . lives in burrows , which in urban areas may extend to pit latrines and garbage collection areas where they forage . These habitats potentially increase the chances of contracting leptospires and may explain the high isolation rates observed in Cricetomys sp . and Crocidura sp . ( Fig 1 ) . Similarly , Crocidura species often appear to live in moist areas . Shrews feed on insects found on soils and wet or moist environments which is reported to support the survival of leptospires for relatively longer periods of time [37 , 38] . On the other hand , Mastomys natalensis habitats are relatively different from those of Cricetomys sp . and Crocidura sp . and is considered a peridomestic rodent found in both wild habitats such as shrubs and grasslands as well as crop fields and houses [39] . This study indicates that Cricetomys sp . and Crocidura sp . serve as major reservoirs of leptospirosis among small mammal species . Future studies aiming at isolation of leptospires to determine serovar prevalence in neglected regions should target the giant African pouched rats ( Cricetomys sp . ) and shrew species ( Crocidura sp . ) or other animal species with similar life styles , e . g . Rattus norvegicus . This study shows that certain Leptospira serovars are specific to certain animal species but that some serovars such as Sokoine and Kenya are found in a broad range of animal species including domestic animals , rodents , bats and fish . Leptospira isolates belonging to serovars Sokoine and Kenya were serologically detected by MAT in diverse animal species including cattle , sheep , goats , dogs , cats , rodents , bats , fish and humans ( Table 5 ) , ( Fig 3 ) . Serovar Sokoine belongs to serogroup Icterohaemorrhagiae that majority of its serovars are often reported in humans worldwide [9] . Furthermore , serovars Sokoine and Kenya belong to the L . kirschneri species that is considered predominant in the East and Central African region including Indian Ocean islands near the East African coast [9 , 13] . Other Leptospira serovars encountered in animals and humans include serovars Hebdomadis , Hardjo , Pomona , Lora , Australis and Canicola . The seroprevalence of these serovars in different animal species and humans shows relatively lower prevalence compared to the proportions of serovar Sokoine , Kenya and Lora . The observed occurrence and prevalence of these serovars in animals and humans indicate that they are major candidate serovars for inclusion and use as antigens in subsequent MAT for serodiagnosis of leptospirosis in animals and humans in this region . Studies focusing in isolation of new Leptospira serovars in humans and animals in Africa are urgently needed to generate a panel of region-specific Leptospira serovars for MAT in African continent . In this study inclusion of local serovars , particularly serovar Sokoine , in MAT detection of leptospiral antibodies in rodents and humans in Tanzania revealed an over 10-fold increase in leptospirosis prevalence which was higher than the prevalence reported in the first broad study of leptospirosis in Tanzania [17] . An increase in seroprevalence was also observed following use of the local serovar Grippotyphosa in determination of leptospirosis in humans and animals . Similarly , the local serovar Kenya also demonstrated higher leptospirosis prevalence in goats , sheep , pigs and dogs exceeding those of reference serovars Hebdomadis , Pomona , Hardjo and Canicola ( Table 5 ) . This further indicates the necessity for isolation of local Leptospira for use in MAT . Majority of seropositive humans and animals particularly rodents had lower titres against the tested serovars whereas serovar Sokoine and Hardjo had the highest number of individuals with lower and higher titres in humans . Serovar Sokoine had also the highest number of animals with lower and higher titres followed by serovar Canicola . Serovar Grippotyphosa had also higher number of animals with lower titres . Serovar Hardjo was not detected in rodents . These findings indicate the need for establishing the cut-off point for serological diagnosis ( MAT ) of leptospirosis in humans and animals in Africa where the disease in endemic due to abundance and diversity of reservoir hosts in Africa . Emphasize is needed to isolate and determine antibody levels using animals’ seropositivity and titres , which are common in animals yielding positive isolates . The findings suggest potential risk of human infection from animal leptospirosis indicated by isolation of three Leptospira isolates from urine of three abattoir workers out of 83 individuals ( 3 . 6% ) sampled among people involved with slaughtering of cattle at Morogoro abattoir where 12 Leptospira isolates belonging to serovar Sokoine , Grippotyphosa and Qunjian ( Canicola ) were isolated from cattle and rodents captured around the abattoir . Leptospira isolation success of 3 . 6% in abattoir workers indicates higher infection risk in this population than in the general population in which only one sample was culture positive ( 0 . 19% ) out of 506 urine samples collected from patients taking blood tests in hospitals . Indeed the overall rate of Leptospira isolation from abattoir workers and other participants from hospitals was lower ( 0 . 67% ) compared to 3 . 6% positivity observed in abattoir workers . Abattoir workers rarely ware adequate personal protective gears that can prevent direct contact with animals’ blood and urines when working at this abattoir and other two informal pig slaughterhouses where leptospires were isolated from pigs also in Morogoro . Successful isolation of leptospires from rodents captured inside and around houses which increases the risk of human infections . The observed lower antibody levels in humans may also indicate consistent exposure to leptospires . Isolation of leptospires from a patient at Kidodi village located 150 km away from Morogoro town further indicates high risk of leptospirosis although the identification of the four human isolates was not successful as they died from contamination before preliminary identification . Kidodi village is located along the Kilombero valley where sugarcane farming is the main agricultural activity that is associated with Leptospira infection . Two Leptospira isolates obtained from pigs appear to differ from serovars belonging to serogroups Grippotyphosa , Icterohaemorrhagiae and Ballum . The two isolates from pigs did not react with any of the reference rabbit serum for serovars belonging to these serogroups including serum against the most common Leptospira serovars Sokoine and Kenya . Further identification of these isolates is required to understand their taxonomic status for serodiagnostic and epidemiological purposes . Furthermore , there is an urgent need for conducting leptospirosis studies focusing on isolation and identification of leptospires from a wide range of terrestrial and aquatic animal species in understudied areas of the African continent . Such studies may enhance mapping of the distribution pattern of Leptospira serovars and the actual burden of this disease across Africa . This could be achieved by establishing Leptospira isolation facilities in microbiological laboratories , especially in universities and research institutions existing in different African countries . Successful isolation should be followed by preliminary identification of the isolates to at least genus level using darkfield microscopy . Subsequently , the isolates could be identified to serogroup level by MAT as shown in this report ( Table 3 ) . The new isolates should also be inoculated into rabbits to produce anti-leptospira antibodies for use in serological typing and MAT [26] . These experimental steps are feasible in most African countries . Further identification of isolates could also be achieved using molecular DNA fingerprinting methods [30 , 31] which can enhance comparisons of DNA fingerprint patterns of known serovars with unknown isolates from similar localities prior to the gold standard cross agglutination absorption test ( CAAT ) recommended for taxonomy of the leptospires [40] . In conclusion , diverse Leptospira serovars occur in animals and humans in Tanzania . These Leptospira serovars have implications in serodiagnosis of this zoonotic disease in both animals and humans . Serovars Sokoine , Kenya , Grippotyphosa , Lora , Pomona , Hardjo , Hebdomadis and Canicola should be included in leptospirosis diagnosis in Tanzania and neighbouring countries . Further studies focusing on isolation and identification of leptospires from terrestrial and aquatic animal species as well as humans are needed to understand potential circulating serovars in neglected regions where the disease is not recognised and not routinely diagnosed in hospitals . A recent study of causes of fevers among infants and children in northern Tanzania shows that 7 . 7% were due to leptospirosis [41] . This prevalence could increase when local Leptospira serovar Sokoine is used as antigen as has been demonstrated in humans from Morogoro region where initial surveillance conducted in 1996 using reference serovars showed lower prevalence ( 0 . 26% ) [17] , and subsequence surveillance conducted in 2006 using local serovar Sokoine showed higher prevalence ( 10 . 7% ) [35] . Similar trend was recently observed in rodents whereby previous studies using reference serovar Icterohaemorrhagiae belonging to serogroup Icterohaemorrhagie of which serovar Sokoine also belongs was 5% [17] and a higher prevalence ( 16 . 9% ) was obtained when local Leptospira serovar Sokoine was used in rodent population located 30 km away from where serovar Sokoine was isolated in Morogoro , Tanzania [42] . Generally , this study shows that use of local serovar Sokoine and Kenya in MAT reveals high leptospirosis prevalence in wide range of animal species and calls for reconsideration of previous knowledge that certain Leptospira serovars infects only certain animal species . Serovar Sokoine in this case was found in humans , domestic animals ( including pet animals ) , rodents , bats and fishes , hence making it a broad antigen for leptospirosis diagnosis in Africa . Public awareness of leptospirosis should be promoted to high-risk populations including farmers , livestock keepers , fishermen and abattoir workers . Leptospirosis has been reported also in residents of rural and urban areas interacting with rodents but not engaged with risk occupational activities [43] . Hospital staff , clinicians , veterinarians and policy makers’ needs awareness of existence and high prevalence of leptospirosis in Africa and should consider its diagnosis and treatment . For example in Tanzania and neighbouring countries , diagnosis of leptospirosis in hospitals can be achieved in collaboration with the leptospirosis laboratory at Sokoine University of Agriculture . Capacity building training on leptospirosis diagnosis , Leptospira isolation and maintenance of Leptospira cultures for use as live antigens for leptospirosis diagnosis should be emphasized in Africa in order to successful control this debilitating and fatal neglected disease . | Leptospirosis disease is widespread in humans and broad range of animal species in Africa . However , leptospirosis is highly neglected and not extensively taught in both medical and veterinary schools almost across Africa . Availability of live leptospires isolated from Africa for use in its diagnosis by the gold standard microscopic agglutination test ( MAT ) is also a major problem . This study reports on local Leptospira serovars and their natural hosts consisting of different animal species , for inclusion in diagnosis of leptospirosis in Africa . A total of 52 Leptospira isolates were obtained from fresh urine and kidneys . African giant pouched rats ( Cricetomys sp . ) and insectivore shrew species ( Crocidura sp . ) had the highest leptospires isolation success . These were identified by serogrouping , cross-agglutination absorption test and molecular typing . Common Leptospira serovars with their respective animals were: serovar Sokoine ( cattle and rodents ) ; Kenya ( rodents and shrews ) ; Mwogolo ( rodents ) ; Lora ( rodents ) ; Canicola ( rodents ) ; Grippotyphosa ( cattle ) ; and an unknown serogroup from pigs . Inclusion of local serovar Sokoine in serodiagnosis revealed a 10-fold increase in leptospirosis prevalence from 1 . 9% to 16 . 9% in rodents and 0 . 26% to 10 . 75% in humans . Future serodiagnosis of leptospirosis in Africa should include these serovars and serovar Hardjo , Hebdomadis , Pomona and other local isolates . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Leptospira Serovars for Diagnosis of Leptospirosis in Humans and Animals in Africa: Common Leptospira Isolates and Reservoir Hosts |
Prominent immune alterations associated with aging include the loss of naïve T-cell numbers , diversity and function . While genetic contributors and mechanistic details in the aging process have been addressed in multiple studies , the role of environmental agents in immune aging remains incompletely understood . From the standpoint of environmental infectious agents , latent cytomegalovirus ( CMV ) infection has been associated with an immune risk profile in the elderly humans , yet the cause-effect relationship of this association remains unclear . Here we present direct experimental evidence that mouse CMV ( MCMV ) infection results in select T-cell subset changes associated with immune aging , namely the increase of relative and absolute counts of CD8 T-cells in the blood , with a decreased representation of the naïve and the increased representation of the effector memory blood CD8 T-cells . Moreover , MCMV infection resulted in significantly weaker CD8 responses to superinfection with Influenza , Human Herpes Virus I or West-Nile-Virus , even 16 months following MCMV infection . These irreversible losses in T-cell function could not be observed in uninfected or in vaccinia virus-infected controls and were not due to the immune-evasive action of MCMV genes . Rather , the CD8 activation in draining lymph nodes upon viral challenge was decreased in MCMV infected mice and the immune response correlated directly to the frequency of the naïve and inversely to that of the effector cells in the blood CD8 pool . Therefore , latent MCMV infection resulted in pronounced changes of the T-cell compartment consistent with impaired naïve T-cell function .
Cytomegalovirus ( CMV ) is a ubiquitous herpesvirus , latently persisting in the majority of the adult human population worldwide [1] . While CMV may induce severe disease in immunocompromised patients , it is generally considered apathogenic for the adult immunocompetent host . Broadly targeted CMV-specific T-cells dominate the memory compartment of exposed subjects [2] , and CD8 T-cells specific for an immunodominant epitope are more frequent in people belonging to older age groups [3] , peaking in the very old [4]–[5] . Epidemiologic studies in healthy elderly volunteers have identified a correlation of CMV seropositivity and the immune risk profile ( IRP ) , a condition characterized by the accumulation of CD28− CD8 T-cells , poor proliferative responses to polyclonal stimulation , inverted CD4/CD8 T-cell ratios [6]–[7] , and , in some cohorts , decreased life expectancy of the host [6] , [8]–[10] . Others have shown that CMV seropositivity correlates to significant telomere shortening in the T-cell pool [11] , poor response to influenza vaccination [12] or poor immunity to a co-resident EBV infection [13] . Therefore , it has been proposed that the massive commitment of the memory T-cell pool to the control of CMV infection may contribute to an accelerated onset of immune senescence [14] , where the filling of the immunological space with CMV-specific T-cells may constrict the T-cell receptor ( TCR ) repertoire . More recent studies raised some doubt on these findings ( for a review see the recent Report from the Second Cytomegalovirus and Immunosenescence Workshop [15] ) , arguing that not all aging human populations exhibit IRP [16] , yet others showed a correlation between CMV seropositivity and frailty in the elderly [9] , [17] or increased mortality [9] . However , while the clinical studies could show the association of CMV infection and parameters associated with immune senescence , they cannot elucidate the cause-effect relationship between these phenomena . Therefore , it was not clear if CMV infection causally contributes to immune senescence or if immune senescence or predisposition for other immune alterations results in an increase of susceptibility to CMV infection . The cause-effect relationship between CMV and immune senescence may be defined by an experimental approach . Such a study would require the comparison of experimentally infected hosts to uninfected ones over the course of their lifetime , which is only feasible in animal models of infection . Due to its strict species specificity , in vivo experimental models of CMV infection and immunity rely on the infection of animals with orthologous CMV viruses . We decided to use the mouse model of infection because lifelong experiments are feasible , CMV-uninfected control hosts are readily available , and the mouse CMV ( MCMV ) model is well characterized [18] , recapitulating the essential features of CMV infection and immunity . CD62Llo ( effector memory - EM ) CD8 T-cells with MCMV specificity accumulate once the primary infection has been cleared and viral latency has been established [19] . This accretion of memory T-cells has been termed memory inflation ( MI ) , and shown to proceed continuously over the host's lifetime [20] , resulting in an accumulation of CMV-specific CD8 cells exhibiting activated CD28−CD27−CD122− phenotypes [21] . Therefore , the mouse model reflects the accumulation of differentiated CMV-specific cells observed in human CMV infection [3] . MI likely reflects an ongoing recruitment of MCMV-specific naïve and central memory cells [22]–[23] , because the MCMV specific EM cells are unable to proliferate in response to Ag upon adoptive transfer [22]–[23] , which is in line with the proliferative hyporesponsiveness of the accumulated CMV-specific CD8 cells in elderly humans [11] , [24] , or in rhesus CMV ( RhCMV ) specific CD8 T-cells in old or adult rhesus monkeys [25] . It remained unclear whether the changes of the T-cell compartment upon CMV infection are irreversible and progressive and whether CMV affects the overall functionality of the T-cell pool . We show here evidence that MCMV infection results in profound changes that affect the entire CD8 pool . The changes included a relative and absolute increase of CD8 T cell counts , a persistent increase of its EM fraction and a reduced representation of the naïve CD8 cells , leading to a distorted TCR repertoire diversity . Moreover , we show that CD8 responses to superinfection with West-Nile virus ( WNV ) , influenza or Herpes Simplex virus are diminished . These changes occurred also upon infection with an MCMV recombinant lacking all known immune evasive genes , arguing that they were not caused by direct CMV interference with the immune system . On the other hand , the strength of the response to WNV correlated inversely to the representation of the naïve pool and directly to that of the EM CD8 pool , thus linking the MCMV-induced changes in phenotype and homeostasis to the functional response to superinfection . An analysis of lymphoid compartments showed that MCMV infection resulted in an overall increase in the size of all CD8 compartment in lymph nodes , yet simultaneously in poor mobilization of CD8 cells to the draining LN upon a viral challenge .
IRP has been described as the inversion of the CD4/CD8 ratio in the peripheral blood due to an accumulation of CD8 T-cells . [10] . We therefore compared the peripheral blood of 15 month old mice experimentally infected for 9 months with MCMV , to uninfected littermate mice ( MOCK ) . To define if any change would be specific for MCMV infection , or a phenomenon that may be observed in any other infection , we included in our experiment a control group of mice infected with a recombinant vaccinia virus ( VACV ) expressing the immunodominant MCMV gene IE-1 ( VACV-ie1 ) . Flow cytometric analysis ( Fig . 1A ) revealed that the CD4 cell pool was not significantly altered ( Fig . 1B ) , whereas the percentage of CD8 cells increased significantly in the MCMV-infected group ( Fig . 1C ) , resulting in a decrease of the CD4/CD8 ratio ( Fig . 1D ) . While the ratio was not inverted , as described in some cohorts of elderly humans with IRP [10] , the observed changes tilted the CD4/CD8 balance towards a relative increase of CD8 cells in the T-cell pool . To understand whether the ratio was altered due to an increase in the count of CD8 , a decrease of CD4 T-cells , or combined changes that may have affected the non-T lymphocyte compartment , we defined their absolute counts per ml of blood . We did not observe changes in the CD4 compartment ( Fig . 1E ) , but found a significant increase in the CD8 count ( Fig . 1F ) . Cells that did not belong to either the CD4 or the CD8 pool were defined as non-T cells , and their number was also not altered ( Fig . 1G ) . Therefore , latent CMV infection resulted in an absolute and a relative increase of the peripheral CD8 pool , which decreased the CD4/CD8 ratio . Since the most prominent changes were observed in the CD8 pool , we focused on this subset . Memory inflation ( MI ) is an ongoing accumulation of MCMV-specific CD8 T-cells with EM ( CD62Llo ) phenotypes in the peripheral pool [19]–[20] . To understand the effects of MI on the EM CD8 T-cell pool , we compared the kinetics of antigen specific responses ( Fig . 2A ) to that of the frequency of EM cells ( Fig . 2B ) . We followed mice infected with MCMV , VACV , or MOCK infected littermates . As expected , in MOCK controls we observed no increase in the antigen specific cells and a modest accumulation of EM cells starting at 300 days and consistent with the age-related changes in the T-cell pool [26] . In VACV-infected control mice , we observed an initial strong response to infection , both in terms of IE-1 specific responses and the increase in the total EM pool , followed by contraction of responses upon virus clearance . Low levels of IE1-specific cells could be detected throughout 14 months post infection , while the EM fraction decreased to levels observed in uninfected controls by 9 months post infection , again consistent with the age-related changes in the immune system . In MCMV infection the response showed a distinctly different pattern: there was an expansion , brief contraction and then an ongoing accumulation ( memory inflation , MI ) of IE-1-specific cells in the peripheral pool ( Fig . 2A ) , consistent with previous observations , [19]–[20] . The total EM pool showed a significant expansion immediately upon infection , and then after a brief dip , returned to a plateau and remained consistently expanded throughout the period of observation , taking up nearly half of the total CD8 pool ( Fig . 2B ) . The IE-1 epitope is only one of the epitopes present in MCMV , and the vast response to primary infection , seen as the acute increase in the total EM cell pool , likely represented the sum of antiviral response to different epitopes . The subsequent persistence of large numbers of EM-cells may have likewise reflected persistent Ag stimulation , in which case one would expect that such EM cells may exhibit signs of recent Ag contact . We examined the activation status of the CD8 pool at a late time point upon infection ( 9 months ) , and have observed a decrease of TCR expression on CD8 T-cells as evidenced by the reduction of CD3 ( Fig . 2C ) or TCR β chain expression ( data not shown ) . Similarly , MCMV infection resulted in the increase of the activation marker Ly6C in the CD62Llo pool ( Fig . 2D , see also Fig . S1 ) , arguing for repeated activation of this subset . Overall , the MCMV infection resulted in a permanent increase of EM cells bearing an activated phenotype . To differentiate between effects upon the EM compartment caused by natural aging processes and those driven by persistent MCMV infection , we compared littermate mice infected at 6 , 12 or 16 months of age once they reached 20 months of age ( 14 , 8 and 4 months post infection , respectively ) . The relative size of the EM pool exhibited strong variability with age in mock-infected controls ( Fig . 3A ) , consistent with age-related changes ( see Fig . 2B ) . In VACV-infected control mice , we observed that the fraction of EM cells in the CD8 pool was elevated at 4 , lower at 8 and the lowest at 14 months post infection , consistent with the decrease of this fraction over time ( see also Fig . 2B ) . In MCMV infection , this decrease was notably slower and the EM pool remained significantly larger than in mock-infected controls at all time points ( Fig . 3A ) . The increase in the fraction of KLRG1+ cells in virus infected mice has been described in several models of infection [27]–[28] , and in human CMV specific CD8 T-cells [29] and was shown to remain elevated only in cells responding to persistent viruses , but not to CD8 cells reacting against acute viruses [29] . The persistence of activated T-cells upon MCMV infection was confirmed by the pattern of persistent surface expression of KLRG1 , a marker expressed on the surface of activated cells . While VACV infection resulted in a moderate and transient increase of the KLRG1+ subset of CD8 cells , their fraction was persistently elevated in MCMV infected mice , ( Fig . 3B ) . It was conceivable that the persistence of the EM subset affected the size of the naïve T-cell pool . Thus , we determined the fraction of the CD44−CD11a−CD62L+CD127+ pool ( representative gating in Fig . S2 ) at 20 months of age in mice infected at 6 , 12 or 16 months of age with MCMV or VACV and compared it to the frequency of naïve CD8 cells in mock-infected mice . We observed that both infections resulted in lower frequencies of naïve cells at 4 months post infection , yet this decrease was persistent only in MCMV infection ( Fig . 3C ) . Moreover , at 14 months post infection VACV infected mice showed significantly higher fractions of naive cells than MCMV infected ones . Therefore , only the MCMV infection resulted in an irreversible decrease of the fraction of naïve cells , and that occurred regardless of the age at which infection occurred , indicating that CMV infection may readjust the homeostatic balance of the CD8 T-cell pool . It has been shown that human CMV infection results in significant changes in the pool sizes of TCR families , defined by the variable ® chain ( Vβ ) expression [30] . Moreover , while young mice exhibit remarkably consistent frequencies of T-cells with defined Vβ , aging results in increased variability of TCR Vβ pool sizes [31] . We compared the variability of the various Vβ fractions of CD8 cells in MCMV and in VACV-infected mice at 14 months post-infection and 20 months of age ( representative blot in Figure 4A , representative comparison of Vβ14 frequencies Fig . 4B ) . The coefficients of variability of Vβ8 , Vβ9 , Vβ10 , Vβ13 , and Vβ14 subsets were compared by the F-test of variances and three out of five tested Vβ families ( Vβ9 , Vβ10 and Vβ14 ) showed significantly higher variability in MCMV infected mice than in VACV infection ( Fig . 4C ) , while the other two showed no significance . These results argued for a significant increase in the variability of the size of Vβ TCR families and hence a significant change of the TCR repertoire in the CD8 pool , concomitant to the increase of EM and the reduction of naïve cell representation in the CD8 pool of MCMV infected mice . Changes in the blood compartment of MCMV infected mice matched the population changes observed in HCMV-infected human subjects , arguing that the mouse model may offer translational insights about the effect of CMV infection . Recently , Remmerswaal et al . showed that lymph nodes in humans do not contain effector memory T cells specific for CMV [32] . To examine the effect of the controlled mouse CMV infection on the lymphocyte populations in lymphoid organs , including the spleen or lymph nodes ( LN ) , we compared the percentages of naïve , CM and EM cells in the spleen and inguinal LN of MCMV , VACV and MOCK infected mice at 6 months post infection ( Fig . 5A ) . The spleen population distribution partially matched those observed in the blood , where MCMV resulted in decreased naïve and increased EM percentages , although the difference was significant only between MCMV and VACV infected mice ( Fig . 5A ) . The LN showed a distinctly different pattern , with an elevated proportion of CM in MCMV and VACV infected mice and , surprisingly , the highest proportion of EM cells in MOCK-infected controls , while the naïve pool contained a similar proportion of naïve cells in the LN of all mice ( Fig . 5A ) . Therefore , we concluded that the decrease of naïve and increase of EM cell representation in MCMV infection was specific for the blood , and to a lesser extent for the spleen compartment , but not for the LN . To evaluate whether the change in the fractions of T-cells were a result of the loss of naïve cells , increase in EM , or a combination of both , we quantified the absolute counts of cell subsets in all compartments tested in the previous section . We observed an obvious absolute reduction of naïve in the blood of MCMV and VACV infected mice ( Fig . 5B ) , but also an increase in the size of the EM pool in the blood of MCMV infected mice , partially reflecting the changes in proportions ( Fig . 5A ) . On the other hand , in spleen and LN we observed an absolute increase of all CD8 subsets in MCMV and VACV infected mice over the MOCK controls , with the increase more pronounced in the MCMV that in the VACV group ( Fig . 5B ) . This was in line with our observation that spleen and LN were enlarged in infected mice as compared to the mock controls and that the mean total CD8 number in the LN was more than threefold increased in MCMV infected mice compared to the mock-infected controls ( not shown ) . Hence , while we observed a loss of naïve cells in the blood compartment of MCMV infected mice , which was not the case in the secondary lymphoid organs of the same mice . The CMV-induced perturbations in the representation and pool size of naïve and memory cells in multiple compartments . The altered CD8 T-cell repertoire could have affected the functional response of the naïve pool to heterologous microbial challenge . To test this assumption , we challenged 129Sv6 mice with viruses unrelated to MCMV or VACV and monitored T-cell responses by means of peptide in vitro restimulation . Five months following MCMV , VACV or mock infection , we challenged adult ( 8-month old ) and aging ( 17-month old ) mice with influenza virus . CD8 responses to the influenza NP366 peptide were significantly weaker in aging mice , than in adult ones ( Fig . 6A ) , but only in the mock-infected and in the VACV infected controls . Mice carrying latent MCMV infection showed weak responses to influenza both in the adult and in the aging group . More importantly , CD8 responses were significantly weaker in MCMV-infected than in control-infected mice , arguing that latent MCMV infection results in poor T-cell responses upon superinfection with influenza . Weak CD8 responses to another flu-derived immunodominant peptide ( PA224 ) showed that this loss of response was not limited to the NP366 peptide in old ( Fig . 6B ) nor in adult mice ( data not shown ) . To exclude the possibility that the suppression of naive-cell responses to novel antigen depends on the mouse strain and challenge virus , we challenged BALB/cxDBA/2 F1 mice with WNV . We observed significantly weaker CD8 responses to the WNV peptides only in mice latently infected with MCMV , ( Fig . 6C ) . In a third unrelated experiment we used DBA/2xC57BL/6 F1 mice and Herpes simplex virus type 1 ( HSV-1 ) as the challenge virus . Mice were infected with MCMV or VACV at 3 months of age , challenged 5 months later with HSV-1 , and monitored for the frequency of CD8 cells specific for the immunodominant , Kb-restricted , HSV-1 peptide SSIEFARL . In MCMV infected mice we observed a trend towards weaker responses at early time post infection ( not shown ) and significantly reduced long-term CD8 memory at 7 months post challenge ( Fig . 6D ) . To understand if the function of naïve cells could be restored over very long periods of time , we compared mice infected at 6 , 12 , 16 or 20 months of age with MCMV or VACV to mock-infected controls , by challenging them with WNV at 22 months of age ( 18 , 10 , 6 and 2 months post MCMV infection , respectively ) . MCMV infection resulted in a reduction of CD8 responses to WNV peptides in all MCMV groups ( Two-way ANOVA relative to mock controls p<0 . 0001 ) , and WNV-specific responses were higher in all VACV-infected groups ( Fig . 6E ) , with the possible exception of mice infected with VACV at 20 months of age . Interestingly , this phenomenon seemed specific for MCMV infection but not another chronic Herpesvirus infection , as old mice persistently infected with HSV-1 were indistinguishable from uninfected littermates in terms of CD8 responses to WNV challenge ( Fig . S4 ) . Blood cells from same mice shown in Fig . 6E were stimulated with anti-CD3 antibodies for 6 h to control for the possibility that CMV may have induced general hyporesponsiveness of CD8 T cells , yet the responses were essentially identical in all mice groups ( Fig . S5 ) . More importantly , the fraction of WNV-peptide responding cells was normalized to the percentage of anti-CD3 responsive cells , to obtain the percentage of maximum response . Again , we could observe the weakest response in MCMV infected mice , arguing that the poor response was specific to the CD8 ability to recognize WNV peptides , and not a result of overall immune suppression in MCMV infected mice ( Fig . 6F ) . In conclusion , our results strongly argued that MCMV infection suppresses immediate and memory CD8 responses to superinfection with unrelated virus . The poor CD8 response to superinfection in MCMV infected mice could be explained by two plausible causes . One was that the changes in the CD8 pool upon MCMV infection ( Fig . 1–4 ) resulted in their poor immune response to neoantigens . The other was that MCMV immune-evasive genes [33] suppressed the immune response to WNV or influenza superinfection . While immune evasive genes are expressed with early kinetics , and not during MCMV latency [34] , superinfection may have resulted in MCMV reactivation in professional antigen-presenting cells as they matured [35] . In that case , the expression of MCMV-encoded immune evasive genes would have inhibited DC function and T-cell priming , but MCMV recombinants lacking immune evasive genes would not compromise the immune response to superinfections . Since the known MCMV genes interfering with MHC-I peptide presentation belong to the m2 or to the m145 gene families , we deleted by targeted mutagenesis all the MCMV genes belonging to these two families [36] and used the recombinant virus , termed here ΔMCMV , to define the effects of MCMV immune evasins on the CD8 responses . The CD8 responses to a WNV challenge in mice infected with ΔMCMV was suppressed just as in wild type ( WT ) MCMV infection ( Fig . 7A ) . Thus , poor CD8 response to WNV challenge was unlikely due to immune evasion genes encoded by MCMV . To address the alternative scenario , in which the changes in the peripheral CD8 pool result in poor CD8 responses to superinfections , we cross correlated the fraction of naïve cells to the fraction of CD8 cells responding to the WNV challenge ( Fig . 7B ) . We observed a significant ( p = 0 . 0007 ) and direct linear correlation ( r = 0 . 36 ) of naïve cell pool sizes and response intensities which argued that the size of the naïve pool has the potential to influence the response to WNV . Moreover , the size of the EM pool correlated more strictly ( r = −0 . 49 ) and significantly ( p<0 . 0001 ) to the WNV responses , but the correlation was inversed ( Fig . 7C ) , arguing that the presence of large pre-existing EM pools may impair the immune response to neoantigen . We considered next that the naïve response must initiate in the draining LN of mice upon which the naïve cells are activated to become effector cells which migrate out of the LN into the bloodstream and towards virus infected tissues . While our data showed a correlation of blood CD8 subsets to the size of the CD8 response to WNV , the changes in the blood CD8 subsets of MCMV infected mice did not reflect the situation in the LN ( see fig . 5 ) . Therefore , we reasoned that only by measuring the changes of the CD8 populations in the draining and non-draining LN upon a virus challenge we may explore if MCMV blocked the activation of CD8 cells , or their migration out of the draining LN . Intranasal influenza challenge results in lung infection upon which mediastinal LN act as draining LN , whereas inguinal or mesenteric LN are not involved in the primary responses and may be used as control LN . Comparison of the CD8 population in the draining and control LN of mice acutely infected with influenza at 6 months post MCMV or VACV infection showed that the CD8 compartment was significantly larger in the draining LN of the VACV-infected mice , but not in those from MCMV infected mice ( Fig . 8A ) . This argued that the recruitment and/or activation of CD8 cells in the draining LN of MCMV infected mice may be impaired . This difference was specific for the mediastinal LN of Flu-challenged mice , because a comparison of CD8 counts in MCMV and VACV infected mice in the absence of a flu challenge revealed no difference in the count of CD8 in mediastinal LN ( Figure S6 ) . To understand which subset of CD8 cells was specifically contributing to the stronger CD8 response in the draining LN of VACV infected mice , we compared the naïve , CM and EM subset of CD8 cells in draining and control LN . We observed that draining LN presented similar numbers of naïve cells and an overall higher number of CM cells in both MCMV and VACV infected groups ( data not shown ) , but a specific increase of the EM fraction in the draining LN that was limited to the VACV infected group ( Fig . 8B ) . Therefore , our results argued that an infection with flu resulted in a less vigorous activation of EM cells in the draining LN of MCMV infected mice , than in those of the VACV infected controls .
In this study , we demonstrate that experimental CMV infection diminishes the CD8 response to heterologous virus infections . It has been shown that herpesvirus infections have the potential to improve the immune protection against bacterial infections [37] due to elevated innate immune responses . In our study we observed no positive effects of latent MCMV infection on the immune protection of mice upon WNV challenge , if anything the survival was slightly shorter and lower ( L . C-S . & J . N-Z . , unpublished data ) . During the preparation of this manuscript , we became aware of the results by the Karrer group ( Mekker et al , Submitted as companion manuscript ) , whose results further support this interpretation – these authors found that mice latently infected with MCMV show diminished clearance of lymphoid choriomeningitis virus ( LCMV ) upon challenge ( Mekker et al . submitted ) . The differences between our data and those of Barton et al [37] likely reflect differences in challenge models and perhaps more importantly , the age of experimental animals and/or the length of CMV infection and the time of testing after primary infection . While that study focused on effects occurring relatively soon after infection ( 45 days post infection ) in young adult mice , we focused on effects occurring 5 months or later upon infection , in middle aged or old mice . We considered the possibility that the absolute number of CD8 cells responding to WNV or flu remained unchanged , and that the lower frequencies of responding cells merely mirrored an absolute increase of memory cells due to the CMV-induced T-cell expansions . However , while CMV infection increased the size of the peripheral blood CD8 pool by a factor of 1 . 7 ( Fig . 1C ) , this could not have explained the observed 10 or 12-fold differences in the frequency of flu-specific cells ( see Fig . 6A ) , which was consistent with the absolute loss of CD8 response to LCMV challenge observed in old MCMV-infected mice ( Mekker et al . submitted ) . The phenotype analysis of the blood compartment argued for two main possibilities: ( i ) that a loss of naïve T cell numbers , and/or their TCR repertoire diversity , caused poor responses to superinfection with emerging viruses ( this scenario could explain the difference between MCMV and mock-infected mice based on numerical differences , but not between MCMV and VACV infected mice , since both infections reduced the naïve CD8 count in the blood; potential TCR repertoire changes remain to be analyzed in sufficient detail ) ; or ( ii ) that accumulation of large EM cell populations inhibited the response of the remaining naïve cells , which was in line with findings of others ( Mekker et al . submitted ) . The analysis of compartments other than the blood showed an overall accumulation of CD8 cells with no loss of the naïve subset , excluding the reduced naïve cell numbers as the mechanism underlying poor responses in MCMV infected mice . Moreover , in LN of MCMV infected mice we observed a specific enrichment of the CM and not the EM fraction . This is in line with data by Snyder et al . showing that the MCMV-specific EM fraction of memory-inflated cells cannot renew on its own , but rather mirrors the proliferation of a CMV-specific subset of CM cells in the LN of infected mice [22] . Finally , and most importantly , upon an influenza challenge we observed an overall expansion of the CD8 pool in the draining LN , which was impaired in MCMV-infected mice . Detailed analysis showed that this impairment was mainly due to poor expansion and/or accumulation of effector CD8 cells . Therefore , while our results do not formally exclude the possibility that the poor immune responses were caused by other mechanisms , they suggest that latent MCMV infection specifically contributes to the poor activation of CD8 cells in the LN of mice carrying latent MCMV . Mekker et al . also observed that young animals infected for two months with MCMV prior to challenge showed less reduction in CD8 responses to LCMV challenge than old animals , while we observed an opposite trend in adult and old mice challenged with influenza at 5 months post CMV infection . This difference likely reflects the differences in age at challenge and length of MCMV infection prior to infection , or in the choice of viruses used for challenge . Interestingly , when we normalized the age of mice at challenge , but infected mice with MCMV at different ages , we observed that the CD8 response to superinfection with WNV was diminished in MCMV infected animals , but was not worsening with the passage of time upon MCMV infection ( Fig . 5E and F ) , Similarly , VACV infection resulted in a transient increase of EM and of KLRG1+ fractions , and the only groups of MCMV and VACV infected animals that showed no significant difference to one another in responses to WNV challenge were the groups of animals infected for two months prior to challenge ( Fig . 6E and F ) . Nothwithstanding these minor discrepancies , our results described above and the data from Mekker et al . are highly consistent and demonstrate that MCMV infection results in several phenotypical and functional changes of the immune system that are usually associated with aging . Surprisingly , we observed that MI does not reflect a progressive expansion of the entire EM pool . In fact , after its initial rapid expansion , the fraction of EM cells in the CD8 pool did not increase further than the levels seen 14 days after infection ( Fig . 1C ) . On the other hand , the CD8 cells recognizing the immunodominant epitope YPHFMPTNL increased in the same mice from approximately 2 to 16% of the CD8 compartment . This implies that after the initial expansion of EM cells , the main change within the CD8 compartment did not involve their further expansion , but rather the replacement of the relatively polyclonal antiviral response by the oligoclonal response against defined epitopes , while the overall magnitude of the response remained unaltered . Therefore , our results raise the question whether the increase of CMV pp65 Ag specific cells observed with advancing age [3] , indeed reflects an increase in the population of CMV specific cells , or merely reflects focusing of CMV-specific CD8 responses upon a narrower range of defined immunodominant targets . Numerous clinical studies showed that CMV seropositivity coincides with poor responses to other viruses [13] , poor responses to vaccines [12] , or poor life expectancy in the very old [6] , [10] , [38] . We showed recently that the phenotype and function of CMV-specific T-cells from old and immunosenescent rhesus monkeys are indistinguishable from those in adult and immunocompetent ones [25] . However , the rhesus model did not allow the comparison of CMV infected and uninfected hosts , because essentially all captive rhesus monkeys become naturally infected with CMV early in their life . Therefore , the effects of CMV infection on the aging immune system could only be defined in mouse models of infection , where SPF colonies allow the maintenance of MCMV negative controls throughout their lifetime . To our knowledge , along with the study by Mekker et al . , this was the first attempt to establish an experimental model to study this question and the first experimental proof that a persistent herpesvirus infection may lead to an irreversible change in the CD8 pool and impair their ability to respond to emerging infections . While Mekker et al showed clearly that MCMV infection results in a loss of relative and absolute responses to superinfection with unrelated viruses , we showed that the poor responses were exclusive to infection with MCMV and could not be observed in infections with other viruses , such as vaccinia or Herpes simplex , and showed that poor responses are observed in several murine strains and F1 hybrids , arguing that the phenomenon is independent of the mouse gentoype . Mekker et al . showed by adoptive transfer experiments that MCMV infection does not compromise the ability of donor cells to respond to an antigen that they recognize , arguing that MCMV did not impair CD8 function in general , but only the endogenous responses to viruses , whereas we showed that none among the known MCMV immune evasive genes contributes to the suppression of CD8 responses to challenge . Instead we could show a decrease in the recruitment and/or activation of CD8 cells in draining LN upon a viral challenge . Therefore , our results and those from Mekker et al . may have profound implications for our understanding of the immune aging process in a microbiological environment that reflects the real-life exposure of the average individual . While we focused in this study on the CD8 subset , our data do not exclude the possibility that latent MCMV infection may affect the CD4 T-cell and the B-cell subsets of lymphocytes . Future studies will allow us to elucidate the effect of persisting CMV infections on T-helper subsets and on humoral immune responses , a question particularly important in light of the relevance of these subsets for the efficiency of vaccination strategies . Based on our data , we propose a model in which the persistent CMV infection causes an ongoing recruitment of EM cells in in lymph nodes , which impairs the ability of the naïve cells to mount responses to unrelated virus . Future studies should elucidate whether and to what extent these effects are due to competition of T-cell populations within the lymph node , or alterations in the migration and/or antigen presenting function of professional APC , or whether there also may be a component of TCR repertoire diversity reduction , for which there is some support ( Smithey , M . J . et al . , submitted for publication ) . The newly established balance may offer benefits in terms of stronger innate immune responses [37] , yet at the same time diminishes the efficiency of the adaptive branch to target neoantigens . All of these changes , combined with age-related cell-autonomous changes in T-cells [39] or other parts of the immune system ( rev in [40] ) could contribute to the clinical evidence of poor prognosis and poor CD8 function in very old CMV-positive hosts observed in some studies [6] , [9] , [41] . Our model should provide a useful tool to elucidate the effects of persistent CMV infection on the immune system and the mechanisms of this interaction .
All animal experiments at OHSU were performed according to federal ( U . S . Animal Welfare Act ) and institutional guidelines , following the Institutional animal care and usage committee ( IACUC ) requirements , under the protocol #0724 . All animal experiments performed at HZI were performed in compliance with the German animal protection law ( TierSchG BGBI S . 1105; 25 . 05 . 1998 ) and were approved by the responsible state office ( Lower Saxony State Office of Consumer Protection and Food Safety ) under permit number 33 . 9-42502-04-11/0426 . The mice were housed and handled in accordance with good animal practice as defined by FELASA and the national animal welfare body GV-SOLAS . MCMV , molecular clones MW97 . 01 [42] and Δm1-17+144-158 [36] were grown on MEF cells and purified by sucrose cushion ultracentrifugation . WNV , strain 385-99 , provided by Dr R . T . Tesh ( U . Texas Med . Branch , Galveston , TX ) , Herpes simplex type 1 virus - strain 17 ( HSV-1 ) , syn+ , provided by Dr . J . McGeoch ( University of Glasgow , Scottland , UK ) , VACV strain Western reserve ( VACV-WR ) and MCMV-ieI-Vaccinia Virus ( VACV-ie1 ) [43] were grown and titrated on Vero cells . Influenza PR/8/34 strain ( Flu ) , grown on hen eggs , was purchased from Charles River . Mice were purchased from NCI intramural breeding colony ( Frederick , MD ) and housed in barrier BSL2 or BSL3 housing at OHSU West Campus , or purchased from Janvier ( France ) and housed in barrier BSL2 housing at HZI . The SPF status of experimental mice was confirmed by monitoring sentinels housed on same blue-line racks for the duration of the experiments . Experiments were performed on BALB/c , 129Sv6 , ( BALB/c×DBA/2 ) F1 , or ( DBA/2×C57BL/6N ) F1 mice . Unless otherwise indicated , data from ( BALB/c×DBA/2 ) F1 mice are shown , and this cohort was additionally controlled on a cage basis for murine Norovirus by serology during the course of the experiment ( all CMV and mock-infected mice were negative ) . Mice were intraperitoneally ( i . p . ) infected with 105 PFU of MCMV or 106 PFU of VACV . Challenge assays were performed with 100 PFU of WNV or 106PFU of HSV-1 applied i . p . , or intranasally with 300 EID50 of the Flu virus . We stimulated blood leukocytes corresponding to 100 µl blood with 10−6 M concentration of peptides in RPMI ( 5% FBS , 5 µg/ml Brefeldin A ) for 6 h at 37°C . We used influenza peptides NP366–374 ( peptide sequence ASNENMETM ) or PA224–233 ( SSLENFRAYV ) , the HSV-1 gB peptide SSIEFARL , the CMV peptide YPHFMPTNL ( IE-1 derived ) or the WNV peptides identified by screening ( see below ) . Negative controls consisted of cells that were incubated in the same conditions in the presence of an ovablumin peptide ( OVA , SIINFEKL ) . Upon stimulation , the cells were washed and subjected to surface staining and intracellular cytokine staining ( ICCS ) . To identify H-2d restricted immunodominant WNV peptides , we screened a library of 684 peptides ( 15-mers overlapping by 10 peptides ) spanning the entire WNV genome and identified two immunodominant 15mers that induced cytokine responses upon in vitro restimulation of splenocytes from WNV infected mice . Two optimal H2-Kd-restricted restricted nonamers ( KYVDYMSSL and GYISTKVEL ) were predicted using a public web-service for epitope prediction [44] ( available at www . syfpeithi . de ) and their ability to induce in vitro cytokine responses was experimentally validated by in vitro restimulation of splenocytes and ICCS . Erythrocytes were lysed by brief ( 5 s ) osmotic shock in dH2O , and blood leukocytes were stained at +4C for 30′ with fluorophore-conjugated monoclonal antibodies . We used as our standard an 8-color panel anti ( α ) CD3-APC-Alexa750 ( eBioscience , San Diego , CA ) , αCD4-Pacific Blue ( BD , San Jose , CA ) , αCD8-Percp-Cy5 . 5 ( BD , San Jose , CA ) , αCD11a-PE-Cy7 ( eBioscience , San Diego , CA ) , αCD44-Alexa700 , αCD62L-PE-Texas Red ( Invitrogen , Carlsbad , CA ) , αCD127-PE , and αKLRG1-biotin followed by Streptavidin-Qdot525 . In some experiments , we also used αLy6C-FITC ( eBioscience , San Diego , CA ) , αCD27-APC ( Biolegend ) or YPHFMPTNL-Ld tetramers coupled to APC ( NIH tetramer core facility ) . For ICCS , cells were washed after the surface staining with the 8-color phenotype panel described above and then fixed for 5′ in 100 µl of IC-fixation buffer ( eBioscience , San Diego , CA ) , incubated for additional 5′ in permeabilization buffer ( eBioscience , San Diego , CA ) , washed twice and stained with anti-IFNγ-APC ( BD , San Jose , CA ) for 1 h . Upon two washes in permeabilization buffer , cells were acquired in an LSR2 flow-cytometer ( BD , San Jose , CA ) and analysis was performed by FlowJo software ( TreeStar , Ashland , OR ) . In select bleeds , an aliquot ( 100 µl ) of whole blood was directly analyzed in an AC•T 5diff hematologic analyzer ( Beckman-Coulter ) or CD8+CD4− leukocytes from a defined fraction of spleen or lymph-node homogenates were counted in an Accuri C6 flow-cytometer ( BD ) to define total lymphocyte counts per sample . PCR for each of the 22 variant ( V ) regions of the TCR® chain ( Vβ ) was performed essentially as described [45] , with minor modifications . We isolated CD8+ , CD3+ , CD4− splenocytes from 129Sv6 mice infected for 14 months with MCMV or their uninfected littermates using a FACS-ARIA II sorter ( BD Biosciences , San Jose , CA ) , isolated their mRNA and subjected it to a panel of 22 reverse transcriptase ( RT ) PCR reactions specific for each Vβdomain . PCR products were subjected to electrophoresis on a polyacrylamide gel ( PAGE ) , and analyzed by densitometry in an ABI 3100 sequencer apparatus . V regions of different TCR clones differ in their sequence , but also in size , due to random addition and subtraction of nucleotides during somatic recombination . Thus , PCR of V TCR regions , results in products of varying lengths , where products with the median length of about 10 aa ( 30 nt ) are the most abundant , followed in a descending order by clones that are progressively shorter or longer than this median length . PCR followed by PAGE and densitometry , allows the differentiation of polyclonal cell populations , with bell-shaped distribution of PCR products in PAGE densitometry , from the oligoclonal populations , where reduced diversity does not allow the formation of a Gaussian curve . Statistical analysis was performed with GraphPad Prism software . | The cytomegalovirus ( CMV ) is a widespread virus of the herpesvirus family , which latently infects the majority of the adult human population worldwide . While CMV causes severe disease in AIDS patients , in recipients of organ transplants , or when infection occurs during pregnancy , this virus is considered apathogenic for the general population . Several reports indicated that CMV infection may be associated with poor immune function , and even survival , in older adults , yet it remained unclear whether CMV infection may have the potential to impair the immune function . Here we test this possibility in a mouse model of CMV ( MCMV ) infection . We found that mice carrying latent MCMV exhibit significant changes in CD8+ lymphocytes , which are crucial in the recognition of , and protection against viruses . These changes were similar to changes observed in aging and resulted in poor cytotoxic T-lymphocyte response to infection with unrelated viruses , such as the West Nile virus or the influenza virus . Our study provides the first experimental demonstration that latent CMV infection may impair the immune defense of the host . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"infectious",
"diseases",
"immune",
"cells",
"west",
"nile",
"fever",
"influenza",
"immunology",
"biology",
"viral",
"diseases",
"cytomegalovirus",
"infection",
"immunomodulation"
] | 2012 | Cytomegalovirus Infection Impairs Immune Responses and Accentuates T-cell Pool Changes Observed in Mice with Aging |
Cyclin-dependent kinase 9 ( Cdk9 ) promotes elongation by RNA polymerase II ( RNAPII ) , mRNA processing , and co-transcriptional histone modification . Cdk9 phosphorylates multiple targets , including the conserved RNAPII elongation factor Spt5 and RNAPII itself , but how these different modifications mediate Cdk9 functions is not known . Here we describe two Cdk9-dependent pathways in the fission yeast Schizosaccharomyces pombe that involve distinct targets and elicit distinct biological outcomes . Phosphorylation of Spt5 by Cdk9 creates a direct binding site for Prf1/Rtf1 , a transcription regulator with functional and physical links to the Polymerase Associated Factor ( PAF ) complex . PAF association with chromatin is also dependent on Cdk9 but involves alternate phosphoacceptor targets . Prf1 and PAF are biochemically separate in cell extracts , and genetic analyses show that Prf1 and PAF are functionally distinct and exert opposing effects on the RNAPII elongation complex . We propose that this opposition constitutes a Cdk9 auto-regulatory mechanism , such that a positive effect on elongation , driven by the PAF pathway , is kept in check by a negative effect of Prf1/Rtf1 and downstream mono-ubiquitylation of histone H2B . Thus , optimal RNAPII elongation may require balanced action of functionally distinct Cdk9 pathways .
Regulation of RNA polymerase II ( RNAPII ) transcription elongation affects the expression of many eukaryotic genes [1] . A central regulator of metazoan elongation is positive transcription elongation factor b ( P-TEFb ) , a complex comprising cyclin-dependent kinase 9 ( CDK9 ) and a cyclin partner . P-TEFb was identified on the basis of its biochemical activity in overcoming promoter-proximal RNAPII pausing imposed by DRB-sensitivity inducing factor ( DSIF; a complex of Spt4 and Spt5 ) and a negative elongation factor ( NELF ) [2]–[5] . Cdk9 orthologs are present and essential in Saccharomyces cerevisiae ( Bur1 ) and Schizosaccharomyces pombe ( Cdk9 ) , where they have imputed positive roles in RNAPII elongation [6]–[12] . Cdk9 phosphorylates multiple proteins involved in elongation , including Spt5 and the Rpb1 subunit of RNAPII [13] . In addition to regulating elongation , these phosphorylations are implicated in the recruitment of chromatin-modifying enzymes and mRNA 3′-end processing factors [11] , [14]–[16] . Cdk9 can also phosphorylate and activate the histone H2B ubiquitin-conjugating enzyme RAD6 , indicating multiple links between Cdk9 and chromatin modification [17] , [18] . Cdk9 phosphorylates both Rpb1 and Spt5 within repeated amino acid motifs at their C-termini . The Rpb1 C-terminal domain ( CTD ) repeat has consensus sequence Y1SPTSPS7 and is involved in coordinating co-transcriptional events [14] . Cdk9 can phosphorylate multiple sites within the CTD heptad , but its roles in RNAPII elongation have been commonly attributed to Ser2 phosphorylation [13] . Spt5 contains a C-terminal repeat domain ( which varies in size and sequence among taxa ) that is phosphorylated by Cdk9 on a conserved threonine residue within the repeated motif [4] , [11] , [16] , [19] . The relative contributions of specific Cdk9 targets to its various functions in transcription are not known . The Polymerase-Associated Factor ( PAF ) complex is a key mediator of Cdk9 function during elongation . PAF is minimally composed of conserved subunits Paf1 , Leo1 , Cdc73 , and Ctr9 [20] , and its functions broadly overlap with those of Cdk9 . In vitro , PAF cooperates with the phosphorylated Spt5 CTD to relieve promoter-proximal RNAPII pausing , and stimulates RNAPII elongation on DNA and chromatin templates [21]–[24] . Like Cdk9 , PAF also functions in mRNA 3′ end processing and co-transcriptional chromatin modification [25]–[28] . In budding yeast , PAF recruitment to transcribed genes requires Bur1-dependent Spt5 phosphorylation , which promotes a direct interaction between the Spt5 CTD and the Rtf1 subunit of PAF [15] , [16] , [18] , [29]–[31] . Rtf1 is a stable component of PAF in budding yeast , but is biochemically separate from PAF in metazoans [22] , [26] , [32]–[37] . Both PAF and Rtf1 are present at transcribed genes and are required for co-transcriptional formation of mono-ubiquitylated histone H2B ( H2Bub1 ) and methylation of histone H3 lysine 4 ( H3K4me ) [25] , [26] . Mutation or knockdown of PAF1 and RTF1 revealed shared functions in telomeric gene silencing in S . cerevisiae , somite development and hematopoiesis in zebrafish , regulation of flowering time in A . thaliana , and mouse embryonic stem cell pluripotency [25] , [38]–[41] . In contrast , zebrafish Rtf1 and PAF were found to have distinct functions in cardiac development [34] . Characterization of PAF mutants in S . cerevisiae revealed pleiotropic deficiencies for paf1Δ and ctr9Δ mutants but few phenotypes overall in an rtf1Δ strain [42] . Indeed , an rtf1Δ mutation was found to suppress the phenotypes of paf1Δ , suggesting that the relationship between Rtf1 and PAF could , in some circumstances , be antagonistic [37] . Lethality of bur1Δ in budding yeast is also suppressed by rtf1Δ , as are phenotypes caused by mutation of the associated cyclin Bur2 [43] , [44] . Thus , Rtf1 may counteract an important shared function of Cdk9 and PAF . Although both Rtf1 and PAF are required for formation of H2Bub1 and H3K4me , genetic analyses in S . cerevisiae indicate a particularly close functional connection between Rtf1 and these histone modifications [45] , [46] . While the precise gene regulatory functions of H2Bub1 and H3K4me have not been elucidated , H2Bub1 , like Rtf1 , has been found to oppose Cdk9 . In the fission yeast S . pombe , phenotypes associated with loss of H2Bub1 are suppressed by reduction of activity of Cdk9 . These include a general 3′ shift of RNAPII density within genes , a pattern that could reflect enhanced elongation through coding regions caused by aberrant Cdk9 function [11] . H2Bub1 itself depends upon Cdk9 activity toward Spt5 , and also contributes positively to Spt5 phosphorylation ( Spt5-P ) levels . Thus , H2Bub1 and Spt5-P may be part of an auto-regulatory mechanism that fine-tunes Cdk9 functions . The Cdk9 function antagonized by H2Bub1 involves a target other than Spt5 , suggesting the involvement of a distinct Cdk9 pathway [11] . Using the fission yeast S . pombe as a model system , we report that PAF and Rtf1 ( which we designate Prf1 in S . pombe ) are involved in two distinct and opposing regulatory pathways , both of which depend on Cdk9 activity . We further show that these divergent functions involve different Cdk9 targets , and mirror the opposing functions of Cdk9 and H2Bub1 in elongation .
To further explore the relationship between Cdk9 activity and H2Bub1 in S . pombe , we assessed the biochemical and functional properties of orthologs of known PAF complex subunits . We first tagged the genes encoding homologs of Paf1 ( systematic ID # SPAC664 . 03 ) , Ctr9 ( SPAC27D7 . 14c; named tpr1+ in S . pombe ) , and Rtf1 [SPBC651 . 09c; which we henceforth refer to as Prf1 , for PAF-related factor 1 , to avoid confusion with the unrelated gene SPAC22F8 . 07c currently annotated as rtf1+ [47]] with the tandem affinity purification ( TAP ) tag and purified the factors from whole-cell extracts [48] . Purified material was analyzed by SDS-PAGE alongside that from an untagged strain . Polypeptides uniquely present in the TAP purifications were excised from the SDS gel and analyzed by tandem mass spectrometry . Material purified from paf1-TAP and tpr1-TAP strains contained orthologs of the previously defined PAF complex components Paf1 , Leo1 ( SPBC13E7 . 08c ) , Tpr1 , and Cdc73 ( SPBC17G9 . 02c ) ( Figure 1A ) . Prf1 was not visible by SDS-PAGE in the Tpr1-TAP or Paf1-TAP preparations , nor were PAF components visible in the Prf1-TAP preparation ( Figure 1A ) . Thus , Prf1 and PAF are predominantly separate in S . pombe cell extracts . To further confirm this finding we carried out in-solution mass spectrometry analysis on the Tpr1-TAP and Prf1-TAP preparations . Material derived from an untagged strain was analyzed in parallel as a control . This analysis identified the core PAF subunits Paf1 , Leo1 , Tpr1 , and Cdc73 with high confidence in the Tpr1-TAP preparation , but not in the control or Prf1-TAP preparations ( see Table S1 ) . Conversely , Prf1 was identified with high confidence only in the Prf1-TAP purification . We also investigated whether we could detect association between Prf1 and PAF in a single-step isolation procedure . We constructed a strain carrying both prf1-TAP and tpr1-13xmyc fusion genes . Whole-cell extracts from this strain , as well as from a tpr1-13xmyc strain , were incubated with IgG-sepharose beads to isolate Prf1-TAP and purified material was analyzed by western blotting . Whereas Prf1-TAP was efficiently purified , no Tpr1-13xmyc protein could be detected , further arguing that Prf1 and PAF do not stably associate in S . pombe cell extracts ( Figure 1B ) . We found that deletion of the S . pombe prf1+ gene , or of the tpr1+ and cdc73+ genes encoding PAF components , resulted in a total loss of H2Bub1 as gauged by immunoblotting , comparable to that observed in htb1-K119R ( ubiquitin acceptor site ) or brl2Δ ( E3 ubiquitin ligase ) mutants ( Figure 2A ) . The presence of normal levels of H2Bub1 in the prf1-TAP and tpr1-TAP strains affirmed the functionality of the respective TAP-tagged proteins ( Figure S1 ) . As expected from the established dependence of H3K4me on H2Bub1 in yeast [49] , [50] , the prf1Δ , tpr1Δ , and cdc73Δ mutations also strongly diminished tri- and dimethyl forms of H3K4me , albeit to varying extents ( Figure S2 ) . Thus , Prf1 and PAF have a shared function in co-transcriptional histone modification . The paf1Δ and leo1Δ mutations caused only partial reduction in H2Bub1 and H3K4me levels , indicating divergent roles of PAF complex subunits in promoting these modifications ( Figures 2A and S2 ) . Because the H3K36me modification was maintained in these mutants , we infer a specific role for S . pombe PAF and Prf1 in the H2Bub1/H3K4me chromatin modification axis ( Figure S2 ) . Loss of H2Bub1 caused by htb1-K119R or brl2Δ mutations is associated with cell morphology and septation defects ( including “twinned” septa and unseparated chains of cells ) [11] . Microscopic examination of DAPI/calcofluor-stained cells revealed that prf1Δ caused qualitatively and quantitatively similar phenotypes ( Figures 2B and 2C ) . To our surprise , the PAF complex mutations tpr1Δ and cdc73Δ , which eliminated H2Bub1 and reduced H3K4me levels , caused minimal cell morphology and septation defects . This indicates functional divergence between PAF and other factors promoting H2Bub1 . To further probe the functional relationships between Cdk9 , PAF , Prf1 , and H2Bub1 , we assessed growth of the relevant mutant strains under conditions previously shown to make requirement for Cdk9 activity more stringent , including minimal medium ( EMM ) and medium containing the nucleoside biosynthesis inhibitor mycophenolic acid ( MPA ) [51] . We found that tpr1Δ and cdc73Δ phenocopied the hypomorphic cdk9-T212A mutant ( which lacks the phosphorylation site for the CDK-activating enzyme Csk1; [7] ) under these conditions , arguing for a functional link between these factors and Cdk9 activity ( Figure 3 ) . The paf1Δ and leo1Δ mutants shared the dramatic EMM sensitivity of tpr1Δ and cdc73Δ but were only mildly sensitive to MPA . Importantly , htb1-K119R and brl2Δ mutants lacked the phenotypic signature associated with reduced Cdk9 activity and loss of PAF , as did prf1Δ ( Figure 3 ) . The brl2Δ and prf1Δ mutations did cause moderate growth defects on EMM , perhaps due to H2Bub1-independent functions of Brl2 and Prf1 proteins . Thus , whereas PAF is functionally aligned with Cdk9 , Prf1 is intimately connected to factors directly involved in H2Bub1 formation . To further define the distinct functions of PAF and Prf1 we investigated mechanisms that target these factors to transcribed genes . Prompted by findings in budding yeast linking PAF chromatin association to Bur1 activity [15] , [16] , [30] , we performed chromatin immunoprecipitation ( ChIP ) in prf1-TAP , tpr1-TAP , and paf1-TAP strains that also carried a Cdk9 variant ( cdk9as ) engineered to be sensitive to inhibition by bulky ATP analogues [12] . We found that chromatin recruitment of all three proteins to the constitutive act1+ and adh1+ genes depended on Cdk9 activity ( Figure 4A ) . Acute inhibition of Cdk9as in these strains had no effect on overall levels of the respective TAP-tagged proteins ( Figure S3A ) and did not greatly impair RNAPII occupancy at act1+ or adh1+ ( Figure S3B ) . This argues that chromatin association of both PAF and Prf1 depends on Cdk9 activity . We reasoned that different Cdk9 targets might mediate the distinct functions of PAF and Prf1 . The Spt5 CTD is a specific target of Cdk9 in S . pombe; the Spt5 CTD consists of tandem copies of a nonapeptide motif of consensus sequence T1PAWNSGSK9 . We reported previously that changing the Thr1 phosphoacceptor site to alanine reduced H2Bub1 levels [11] . We analyzed Prf1 and PAF recruitment to transcribed chromatin by ChIP in strains harboring Spt5 CTD mutations . The strains were constructed by replacing the endogenous spt5+ CTD domain ( which contains 18 nonapeptide repeats ) with a truncated ( but fully functional ) CTD domain containing eight consensus repeats or otherwise identical repeat arrays with T1A or T1E mutations [19] . The strains also carried the cdk9as allele , which allowed assessment of the contribution of specific loss of Spt5-P to the effect of general Cdk9as inhibition . The spt5-T1A mutation reduced chromatin association of Prf1 at act1+ and adh1+ to a level close to that observed upon inhibition of Cdk9as ( 5- to 10-fold; Figures 4B and S4A ) . Thus , Prf1 recruitment to transcribed chromatin strongly depends on phosphorylation of the Spt5 CTD by Cdk9 . The phosphomimetic spt5-T1E mutation also impaired ChIP of Prf1 , but rendered the residual crosslinking of Prf1 Cdk9-independent . Thus , although a negative charge at all Thr1 positions within a truncated CTD is not sufficient for normal Prf1 recruitment it does allow some Prf1 binding through an altered , Cdk9-independent mechanism . In contrast , ChIP of Tpr1-TAP revealed Cdk9-dependent chromatin association in the presence of spt5-T1A and spt5-T1E mutations , as well as in a strain in which the Spt5 CTD is deleted altogether ( spt5ΔC ) ( Figures 4B and S4A ) . This argues that PAF recruitment to chromatin involves a Cdk9 target other than the Spt5 CTD . The modest ( ∼2-fold ) decrease in ChIP of Tpr1-TAP in the spt5-T1A and spt5ΔC mutants ( that was not apparent in the phosphomimetic spt5-T1E mutant ) suggests that Spt5-P does make some contribution to PAF recruitment . These data point to the existence of two functionally distinct Cdk9 pathways ( which we will refer to as the PAF pathway and the Prf1 pathway ) that involve distinct phosphoacceptor targets . We note that global levels of Spt5-P were reduced by prf1Δ , as was found previously for htb1-K119R and brl2Δ mutations ( Figure S5A ) . This implicates Prf1 as part of the positive feedback loop linking Spt5-P and H2Bub1 . Mutations in PAF components either did not affect ( tpr1Δ and cdc73Δ ) or increased ( paf1Δ and leo1Δ ) global levels of Spt5-P , further highlighting the specific functional connection between Spt5-P and the Prf1 pathway . The Rpb1 CTD is a target of Cdk9 . In S . pombe , Cdk9 contributes to global levels of Rpb1-Ser2 phosphorylation and promotes the formation of multiply phosphorylated Rpb1 CTD repeats [12] , [51] . Budding yeast PAF components Cdc73 and Ctr9 both interact with phosphorylated Rpb1 CTD peptides in vitro and enhance global levels of Rpb1-Ser2 phosphorylation in vivo [30] , [32] , [52] . This latter function is shared by the corresponding PAF components in S . pombe ( but not by Prf1; see Figure S5B ) . To investigate the role of Rpb1 CTD phosphorylation in Prf1 and PAF recruitment , we carried out ChIP assays in strains harboring phosphoacceptor site mutations in all the Rpb1 CTD repeats . The “wild-type” control in these experiments carried a truncated but fully functional Rpb1 CTD domain [53] . ChIP assays using an rpb1-S2AS7A double mutant ( in which both serines 2 and 7 are changed to alanine in every CTD repeat ) showed very little effect on chromatin association of either Prf1-TAP or Tpr1-TAP at the active genes analyzed ( although a ∼2-fold , gene-specific reduction of Prf1-TAP association was observed at act1+ ) ( Figures 4C and S4B ) . We also tested the effect of an rpb1-S5A mutation , the lethality of which was bypassed by fusion of rpb1-S5A to the MCE1 gene encoding the mammalian mRNA capping enzyme [53] . As compared to the wild-type rpb1-MCE1 fusion , the rpb1-S5A-MCE1 fusion had little impact on crosslinking of either Prf1 or Tpr1 to the two loci tested ( Figure S6 ) . These data suggest that Rpb1 CTD phosphorylation on serines 2 , 5 , or 7 is not required for Prf1 or PAF association with transcribed chromatin in vivo . The marked dependence of Prf1 chromatin association on Spt5-P led us to test whether this could reflect a direct interaction between the Spt5 CTD and Prf1 . Biotinylated peptides corresponding to the unmodified or the Thr1-P form of the Spt5 CTD were coupled to streptavidin beads and incubated with TAP-purified , native Prf1 or PAF . We detected binding of Prf1 ( ∼10% bound ) to the unmodified Spt5 CTD peptide that was enhanced by phosphorylation ( ∼35% bound ) ( Figures 5A and 5B ) . This interaction was specific , as we detected no binding of Prf1 to Rpb1 CTD phospho-peptides modified at Ser2 , Ser5 , Ser7 , or Ser5 and Ser7 . Given that TAP-purified Prf1 is not visibly associated with other proteins ( Figure 1 ) , this argues that there is a direct interaction between Prf1 and Spt5-P . PAF ( purified from the tpr1-TAP strain ) also bound ( ∼5% of total ) to the unmodified Spt5 CTD peptide but failed to bind detectably to the Spt5-P peptide , suggesting that , in vitro , phosphorylation inhibits PAF association with the Spt5 CTD ( Figures 5A and 5B ) . Similar results were obtained with PAF purified from the paf1-TAP strain ( Figure S7 ) . Thus , we presume that the modest negative effect of the spt5-T1A mutation on Tpr1-TAP crosslinking to active genes was due to another Spt5 CTD-binding factor . Consistent with our ChIP data , we did not detect PAF association with unmodified or singly modified Rpb1 CTD peptides , and detected weak association with the peptide that was doubly phosphorylated at Ser5 and Ser7 . We previously argued that the relationship between Cdk9 and H2Bub1 is mediated by two Cdk9 pathways , one involving Spt5-P and another involving Spt5-P and an additional target [11] . Our data suggested the two pathways were functionally opposed . We wished to test whether the Prf1 and PAF pathways described here might mediate these opposing functions . We first assessed genetic interactions between prf1Δ and reduction of Cdk9 activity . We introduced the cdk9as allele into the prf1Δ background and analyzed double mutants by DAPI/calcofluor staining after inhibition of Cdk9as with 3-MB-PP1 . Cdk9as inhibition strongly suppressed the cell morphology and septation phenotypes caused by prf1Δ ( as it did for brl2Δ; Figure 6A ) . A hypomorphic cdk9-T212A mutation similarly rescued these phenotypes . These results indicate that the cell morphology and septation defects in prf1Δ , brl2Δ , and htb1-K119R strains depend upon Cdk9 activity , and suggest that Prf1 , Brl2 , and H2Bub1 itself act to limit some aspect of Cdk9 function . We next determined whether PAF mutations , which phenocopy cdk9-T212A , could also suppress the cell morphology and septation phenotypes of htb1-K119R , brl2Δ , and prf1Δ mutants . We found that cdc73Δ , when combined with htb1-K119R , brl2Δ , or prf1Δ , suppressed aberrant septation . Similarly , brl2Δ septation phenotypes were effectively suppressed by tpr1Δ ( Figure 6A ) . The dependence of these phenotypes on both Cdk9 activity and PAF is consistent with the Prf1 pathway acting in opposition to the PAF pathway . We also tested whether htb1-K119R , brl2Δ , and prf1Δ could reciprocally suppress cdk9 and PAF mutant phenotypes . None of these mutations affected the growth of cdk9 or PAF mutants on minimal media . Whereas htb1-K119R suppressed the MPA sensitivity of cdk9-T212A , prf1Δ did not , and brl2Δ enhanced this phenotype ( Figure S8 ) . These results indicate that the Prf1 pathway does not generally limit PAF pathway function . In order to ascertain how these two pathways regulate RNAPII elongation , we carried out RNAPII ChIP experiments in strains harboring mutations in either the Prf1 or PAF pathways . We quantified the ChIP by qPCR with primers in nup189+ and SPBC354 . 10+ , genes that exemplify the 3′ shift in RNAPII distribution that we documented previously [11] . Consistent with our previous data , the htb1-K119R mutation led to an increase in RNAPII occupancy near the 3′ end of the nup189+ gene relative to the 5′ end ( Figures 6B and S9A ) . A similar effect was also observed for brl2Δ and prf1Δ mutations , implying that the 3′ RNAPII shift is characteristic of the Prf1 pathway . In contrast , cdk9-T212A , cdc73Δ , and tpr1Δ mutations led to a relative decrease in RNAPII density toward the 3′ end of the gene , consistent with defective elongation ( Figure 6B ) . The effects on RNAPII distribution were generally less significant at SPBC354 . 10+ , but we nonetheless observed a similar pattern in which RNAPII occupancy increased toward the 3′ end of the gene in Prf1 pathway mutants and decreased or was unaffected in PAF pathway mutants ( Figure S9A ) . We also found that the Prf1 and PAF pathway mutations differentially affected the steady-state levels of mRNA produced from these genes . PAF pathway mutants exhibited modest decreases in mRNA levels , whereas no change or modest increases were observed in the Prf1 pathway mutant strains ( Figure S9B ) . These changes are likely to reflect altered RNAPII elongation , since the absolute levels of RNAPII occupancy near the 5′ ends of these genes did not differ between Prf1 and PAF pathway mutant strains ( Figure S9C ) . Overall , our data suggest that the PAF and Prf1 pathways have differential effects on behavior of the RNAPII elongation complex ( Figure 7 ) .
Prf1/Rtf1 and PAF do not form a stable complex in S . pombe whole-cell extracts , a finding consistent with the properties of these proteins in metazoans . Biochemical studies of human Rtf1 have shown that it can be assembled into a complex with PAF in vitro , and that a C-terminal fragment of Rtf1 can be detected in association with PAF purified from nuclear extracts [22] . Although the physiological significance of these findings has not been explored , they suggest the possibility of a regulated interaction between full-length Rtf1 and PAF . Modulation of this interaction could underlie the shared and distinct functions of Rtf1 and PAF in different biological contexts . The similarities we have found between the biochemical properties of Prf1 and PAF in fission yeast and those of their metazoan orthologs underscore the utility of the S . pombe system as a model for study of these factors . S . pombe Prf1 and PAF also show functional resemblance to their human counterparts in that they promote the H2Bub1/H3K4me histone modification axis but not H3K36me , a modification that is PAF-dependent in S . cerevisiae [26] , [43] . We found that individual subunits within S . pombe PAF were functionally distinct: loss of either Tpr1 or Cdc73 was associated with more severe phenotypic consequences in our assays than loss of Paf1 or Leo1 . This contrasts with PAF functional organization in S . cerevisiae , where paf1Δ causes pleiotropic phenotypes [42] . Further investigation into physical interactions between individual PAF subunits and other factors will help to clarify these functional differences . Our data show that Prf1/Rtf1 directly and specifically associates with the phosphorylated Spt5 CTD , a finding that is consistent with recent reports on budding yeast and human orthologs of this protein [30] , [31] , [54] . In vitro , budding yeast Rtf1 was found to bind Spt5-P and the phosphorylated Rpb1 CTD with comparable affinities , and Rtf1 mutations that abrogate its interaction with Spt5-P prevented the association of multiple PAF subunits with transcribed chromatin in vivo [30] , [54] . The budding yeast Ctr9 and Cdc73 proteins could also interact with phosphorylated peptides corresponding to either the Spt5 CTD or the Rpb1 CTD [30] . The apparently overlapping mechanisms involved in recruitment of Rtf1 and other PAF components to chromatin in budding yeast are in line with the robust association of Rtf1 with PAF found in this organism . Our findings in S . pombe point to a specific and primary role for the Prf1/Spt5-P interaction in Prf1 recruitment to transcribed chromatin , and suggest an alternate recruitment mechanism for PAF . The direct interaction between Prf1/Rtf1 and Spt5-P helps explain the critical role of Spt5-P in establishment of H2Bub1 levels . This interaction also raises mechanistic questions regarding the reciprocal dependence of Spt5-P on H2Bub1 . Loss of Spt5-P in the htb1-K119R mutant is accompanied by a reduction in crosslinking of Cdk9 to transcribed chromatin [11] . In the prf1Δ strain , it is possible that loss of Spt5-P is also attributable to the absence of the Prf1/Spt5-P interaction , which could provoke dephosphorylation of Spt5-P . The spt5-T1A mutation did not result in cell morphology and septation defects observed in prf1Δ , brl2Δ , or htb1-K119R mutant strains , suggesting that some Prf1 function is maintained in the absence of Spt5 CTD phosphorylation [11] . This is consistent with our ChIP results , which demonstrated a substantial but incomplete loss of Prf1 chromatin recruitment in spt5-T1A mutant strains . Since residual Prf1 recruitment was further reduced by Cdk9as inhibition , other Cdk9 targets are likely to play some role in this process , although the bypass of Cdk9 dependence in the spt5-T1E mutant suggests that the Spt5 CTD is the primary Cdk9 target involved . Multiple domains of S . cerevisiae Rtf1 can be independently recruited to chromatin , indicating that mechanisms other than Spt5 phosphorylation are also important for Rtf1 chromatin association [31] , [46] , [55] . It is also possible that Prf1 may retain some function even when it is not stably bound to chromatin , as has been observed for the S . cerevisiae PAF complex [56] . We found that PAF recruitment to chromatin in vivo was only modestly sensitive to Spt5 CTD phosphorylation at both transcribed loci tested . Since we found no interaction between the purified PAF complex and Spt5-P peptide in vitro , we presume that this reflects an indirect effect on PAF association with chromatin mediated by another factor that binds to Spt5-P . This is unlikely to be Prf1 , as PAF chromatin association is not affected by the spt5-T1E mutation ( unlike that of Prf1 ) . Because crosslinking of Tpr1-TAP and Paf1-TAP to transcribed loci was dramatically reduced by Cdk9as inhibition even in an spt5-T1A mutant , PAF recruitment must require at least one other Cdk9 target in addition to Spt5 . Given that the Rpb1 CTD is an established target of Cdk9 activity , we were surprised to find that simultaneous elimination of Ser2 and Ser7 in the Rpb1 CTD repeat , or removal of Ser5 , had no substantial effect on PAF recruitment in vivo . This could reflect PAF association with forms of the Rpb1 CTD repeat phosphorylated on both Ser5 and another residue , as has been suggested by recent biochemical experiments with budding yeast PAF components [30] . Alternatively , the role of the Rpb1 CTD may be redundant with the Spt5 CTD [19] , [57] . Pleiotropic phenotypes associated with mutant combinations that include Rpb1-Ser5 have precluded thorough genetic tests of these possibilities . Recruitment of the PAF complex to chromatin may also involve another , as yet unidentified , Cdk9 target . Septation and cell morphology phenotypes of Prf1 pathway mutants were reduced by cdk9-T212A , cdc73Δ , and tpr1Δ mutations , indicating that these components act aberrantly in the absence of the Prf1 pathway . Prf1 pathway mutations did not generally suppress cdk9 or PAF mutant phenotypes , suggesting that the Prf1 pathway has a specialized role in limiting certain aspects of PAF pathway function . ChIP of RNAPII reveals differential effects of the two pathways on RNAPII distribution within the nup189+ and SPBC354 . 10+ gene-coding regions , with the PAF pathway promoting RNAPII occupancy at promoter-distal sequences and the Prf1 pathway limiting occupancy at these locations . This observation supports our earlier studies on the opposing effects of Cdk9 activity and H2Bub1 on intragenic RNAPII distribution , and implies that these findings also characterize the two Cdk9 pathways we have described . We propose that Cdk9-dependent recruitment of PAF , through phosphorylation of the Spt5 CTD and other targets , promotes RNAPII elongation through gene coding regions , as supported by our RNAPII ChIP and mRNA expression data . Spt5 CTD phosphorylation also recruits Prf1/Rtf1 and promotes formation of H2Bub1 , which may act as an auto-regulatory mechanism to balance the positive effects of the PAF pathway . Previous studies documenting a positive role for PAF in elongation have shown Rtf1 to be dispensable for this function , consistent with the distinctive behavior of Prf1 in S . pombe [21]–[23] . Although we did not observe significant effects of the Prf1 pathway mutations on expression of nup189+ and SPBC354 . 10+ ( with the exception of a slight increase for SPBC354 . 10+ in the htb1-K119R mutant ) , it is possible that loss of the putative Cdk9 auto-regulatory function in these mutants would be associated with defective mRNA processing or export , rather than a change in mRNA levels per se . Further elucidation of the mechanisms through which these pathways affect gene expression will be an important goal of future research . A negative role for the Prf1 pathway in elongation could be related to the functions of H2Bub1 in directing nucleosome assembly and histone deacetylation within gene coding regions [58]–[61] . Indeed , H2Bub1 and the Hos2 deacetylase complex have been linked to reduced RNAPII processivity in budding yeast [61] . It is notable that an interaction between PAF and the elongation factor TFIIS that stimulates elongation in mammalian cells is antagonized by the H2Bub1 ubiquitin ligase RNF20 [62] . Although we find no role for TFIIS in opposing H2Bub1 functions in S . pombe ( data not shown ) , a role for the Prf1 pathway and H2Bub1 in altering PAF function in elongation may be conserved . In metazoans , Cdk9 activity has an essential role in release of RNAPII from promoter-proximal pause sites . Phosphorylation of multiple targets , including the Rpb1 CTD , Spt5 , and NELF , is thought to drive RNAPII from the pause , implying coordinate action of Cdk9 phosphorylation events in pause release [4] , [63] . Interestingly , the experiments defining the role of PAF in P-TEFb-dependent pause release in vitro used PAF complex devoid of Rtf1 [21] . We suggest that the P-TEFb pause release function in metazoans , which involves multiple Cdk9 targets and PAF , is analogous to the PAF pathway we have documented in S . pombe . The metazoan analog of the Prf1 pathway might involve Rtf1 recruitment via Spt5-P subsequent to pause release , as part of a mechanism to keep the stimulatory effects of PAF on elongation in check . Thus , the roles of Cdk9 targets in regulating the RNAPII elongation complex may diverge during the course of elongation .
S . pombe strains used in this study are listed in Table S2 . Liquid cultures used standard YES media ( yeast extract 5 g/L , D-glucose 30 g/L , supplemented with 250 mg/L each of histidine , leucine , adenine and uracil ) . Minimal media ( EMM ) was purchased from MP Biomedicals . EMM complete was supplemented with 250 mg/L each of histidine , leucine , adenine , and uracil . Mycophenolic acid ( MPA; Bioshop ) was dissolved in dimethyl sulfoxide ( DMSO ) and used at a final concentration of 25 µg/mL . The TAP and 13xmyc tags were introduced at the relevant chromosomal loci as described previously [64] . All other genetic manipulations were carried out using standard methods [65] . Paf1-TAP , Prf1-TAP and Tpr1-TAP were purified from whole cell extracts as described [48] . For mass spectrometry analysis of polypeptides excised from polyacrylamide gels , purified material was concentrated by addition of 0 . 02% sodium deoxycholate , incubation on ice for 30 min , and precipitation with trichloroacetic acid ( 15% final concentration ) overnight . The precipitate was collected by centrifugation , washed twice with ice-cold ethanol and then resuspended with 10 µl of SDS-PAGE sample buffer . Proteins were resolved by SDS-PAGE on a 4–20% gradient acrylamide gel and visualized by Coomassie ( Bio-rad ) or silver staining before mass spectrometry analysis . To analyze the purified material in solution , the TAP purification procedure was modified as follows: Glycerol and detergent were omitted from the final wash and elution steps , and EGTA was replaced with EDTA in the elution buffer . The eluted material was subjected to mass spectrometry analysis , details of which are provided in the Supplemental Material ( see Text S1 ) . For single-step TAP purification , whole cell extracts were incubated with IgG-sepharose beads ( GE Healthcare ) and the beads were collected and washed as described [48] . Beads were suspended in SDS sample buffer and bound material was analyzed by SDS-PAGE and western blotting . Immunoblotting was performed as previously described [11] . The following commercial antibodies were used: H2Bub1 ( Active Motif #39623 ) , H3K4me1 ( Abcam #ab8895 ) , H3K4me2 ( Abcam #ab32356 ) , H3K4me3 ( Abcam #ab8580 ) , H3K36me3 ( Abcam #ab9050 ) , histone H3 ( Abcam #ab1791 ) , histone H2B ( Millipore #07-371 ) , TAP tag ( Fisher Scientific #PICAB1001 ) , Rpb1 ( 8WG16; Covance #MMS-126R-200 ) , Rpb1-Ser2-P ( H5; Covance #MMS-129R ) , myc ( 9E10; Covance #MMS-150P ) . Antibodies against S . pombe Spt5 and Spt5-P were described previously [11] . DAPI/calcofluor staining was carried out as described previously [11] . For spot tests , 4×105 cells were harvested from exponentially growing cultures , and resuspended in YES media . Ten-fold serial dilutions of the culture were spotted onto the appropriate plates and incubated at 30°C for 3–6 days . ChIP was carried out as previously described [48] with minor modifications . Briefly , 1 . 5×107 cells grown in YES media were fixed with 1% formaldehyde for 30 min . Glycine ( pH 2 . 5 ) was added at a final concentration of 125 mM to stop the crosslinking . Cells were washed twice with cold 1XTBS ( 20 mM Tris-HCl , pH 7 . 5 , 150 mM NaCl , 1 mM Tween-20 ) and snap frozen in liquid nitrogen . Cell pellets were resuspended in 0 . 5 mL of lysis buffer [50 mM HEPES , pH 7 . 6 , 150 mM NaCl , 1 mM EDTA , 1% Triton X-100 , 0 . 1% sodium deoxycholate , 1 mM PMSF , protease inhibitor cocktail ( Roche ) ] , 200 µl of sterile glass beads were added , and cells were lysed at 4°C using a mini-beat beater ( 4×30 s with 1 min rests on ice ) . The lysates were transferred to 1 . 5 mL tubes and centrifuged at top speed for 15 min to collect the chromatin . Chromatin pellets were resuspended in 1 mL lysis buffer , sonicated for 20 minutes using a Bioruptor waterbath sonicator ( Diagenode ) ( 30 s ON/OFF with 2 min rest on ice after 10 min ) and centrifuged at top speed for 5 minutes . The supernatant ( 700–800 µl ) was incubated for 2–3 h with either 50 µl of IgG sepharose beads ( GE Healthcare ) ( for TAP immunoprecipitation ) or with 15 µl of protein G Dynabeads ( Invitrogen ) coupled to 8WG16 antibody ( for RNAPII immunoprecipitation ) at 4°C with rocking . The beads were washed successively with 0 . 5 mL of each of the following buffers: lysis buffer +0 . 1% SDS , lysis buffer +0 . 1% SDS+500 mM NaCl , LiCl buffer ( 10 mM Tris-HCl , pH 7 . 5 , 1 mM EDTA , 250 mM LiCl , 0 . 5% sodium deoxycholate , 0 . 5% NP-40 ) and TE ( 10 mM Tris-HCl , pH 7 . 5 , 1 mM EDTA ) . Chromatin was eluted with 100 µl of elution buffer ( 50 mM Tris-HCl , pH 7 . 5 , 10 mM EDTA , 1% SDS ) at 65°C for 30 min . Reversal of crosslinks , DNA purification , and quantitative PCR were as described [48] . Primers specific for the act1+ , adh1+ , nup189+ , and SPBC354 . 10+ genes have been described previously [11] , [66] . 15 µl of streptavidin Dynabeads ( Invitrogen ) were washed three times with 400 µl of 1xPBS/0 . 1% Triton-X100 . Beads were coupled to 50 µg biotinylated peptide in 380 µl of 1xPBS . Incubation was at room temperature for 3 h with rocking . Spt5 CTD peptides were synthesized at the Keck Foundation Research Laboratory ( Yale University ) and had the sequence [biotin]-GSKTPAWNSGSKTPAWNS . In the phospho-peptide both Thr residues were phosphorylated . Rpb1 CTD peptides were a generous gift from Francois Robert ( IRCM , Montreal ) and had the following sequences: [biotin]-YSPTSPSYSPTSPSYSPTSPS , [biotin]-YSPTSPSY ( pSer ) PTSPSYSPTSPS , [biotin]-SPTSPSYSPT ( pSer ) PSYSPTSPSY , [biotin]-TSPSYSPTSP ( pSer ) YSPTSPSYSP , [biotin]-PTSPSYSPT ( pSer ) P ( pSer ) YSPTSPSYS . Beads were washed twice with washing buffer ( 20 mM HEPES pH 7 . 6 , 20% v/v glycerol , 1 mM EDTA , 0 . 1% Triton X-100 , 250 mM KOAc , 10 mM β-glycero-3-phosphate , 1 mM PMSF ) and resuspended with 150 µl of the binding buffer ( 20 mM HEPES pH 7 . 6 , 0 . 1% Triton X-100 , 1 mM PMSF , 10 mM β-glycero-3-phosphate ) . Binding reactions contained 20 mM Hepes pH 7 . 6 , 0 . 075% Triton X-100 , 0 . 75 mM PMSF , 7 . 5 mM β-glycero-3-phosphate , 5% glycerol , 37 . 5 mM KOAc , 2 . 5 mM MgOAc , 0 . 25 mM EDTA , 0 . 25 mM DTT and 50–200 ng of purified factors ( ∼50 ng for Prf1-TAP or 200 ng for Paf1-TAP and Tpr1-TAP ) in a volume of 200 µl . The mixtures were incubated at 4°C for 1 h with shaking . Beads were washed 4 times with 1 mL of washing buffer , resuspended with 20 µl of SDS-PAGE sample buffer , boiled for 5 minutes at 95°C , and subjected to SDS-PAGE and immunoblotting with a TAP antibody . Signal intensities were quantified using ImageJ software and used to calculate percent bound . | The expression of many eukaryotic genes is regulated during the elongation phase of transcription . Cyclin-dependent kinase 9 ( CDK9 ) enzymes are key positive regulators of RNA polymerase II ( RNAPII ) elongation , and are required for co-transcriptional events such as mRNA processing and chromatin modification . Cdk9 phosphorylates RNAPII and the elongation factor Spt5 , but the molecular pathways that underlie Cdk9 functions in transcription remain poorly defined . Using fission yeast Schizosaccharomyces pombe as a model system , we have identified two Cdk9-dependent pathways that modulate RNAPII elongation . One pathway involves the direct interaction of phosphorylated Spt5 with Prf1/Rtf1 , a multi-functional transcriptional regulator . A second pathway , involving a different Cdk9 target , directs the Polymerase Associated Factor ( PAF ) complex to transcribed chromatin . The Prf1 and PAF pathways are functionally distinct and have opposing effects on the RNAPII elongation complex . Our data provide novel insights into Cdk9 functions and suggest that Cdk9 auto-regulatory mechanisms are an important feature of RNAPII elongation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | The PAF Complex and Prf1/Rtf1 Delineate Distinct Cdk9-Dependent Pathways Regulating Transcription Elongation in Fission Yeast |
The developmental time of vector insects is important in population dynamics , evolutionary biology , epidemiology and in their responses to global climatic change . In the triatomines ( Triatominae , Reduviidae ) , vectors of Chagas disease , evolutionary ecology concepts , which may allow for a better understanding of their biology , have not been applied . Despite delay in the molting in some individuals observed in triatomines , no effort was made to explain this variability . We applied four methods: ( 1 ) an e-mail survey sent to 30 researchers with experience in triatomines , ( 2 ) a statistical description of the developmental time of eleven triatomine species , ( 3 ) a relationship between development time pattern and climatic inter-annual variability , ( 4 ) a mathematical optimization model of evolution of developmental delay ( diapause ) . 85 . 6% of responses informed on prolonged developmental times in 5th instar nymphs , with 20 species identified with remarkable developmental delays . The developmental time analysis showed some degree of bi-modal pattern of the development time of the 5th instars in nine out of eleven species but no trend between development time pattern and climatic inter-annual variability was observed . Our optimization model predicts that the developmental delays could be due to an adaptive risk-spreading diapause strategy , only if survival throughout the diapause period and the probability of random occurrence of “bad” environmental conditions are sufficiently high . Developmental delay may not be a simple non-adaptive phenotypic plasticity in development time , and could be a form of adaptive diapause associated to a physiological mechanism related to the postponement of the initiation of reproduction , as an adaptation to environmental stochasticity through a spreading of risk ( bet-hedging ) strategy . We identify a series of parameters that can be measured in the field and laboratory to test this hypothesis . The importance of these findings is discussed in terms of global climatic change and epidemiological consequences .
The developmental time of an organism plays an important role in population dynamics and evolutionary biology because of its direct influence on the population growth rate , synchronization of reproduction , or with resources availability and sensitivity to climatic conditions . Variability of the time of development can be selected when variable conditions for survival or reproduction occur , e . g . due to climatic events or unstable population dynamics ( e . g . , [1]–[10] ) . Individual variation in developmental time can have important ecological and epidemiological consequences so its study is particularly relevant in insect disease vectors such as triatomines . The triatomines ( Triatominae , Reduviidae ) , vectors of Chagas disease , have been studied to a great extent for at least 70–80 years , with an important scientific knowledge accumulation about their general biology ( since 1930 [11] ) , physiology ( since 1933 [12] ) and genetics ( since 1948 [13] ) . Life history traits such as fecundity , juvenile and adult survival , fasting capacity , developmental time , mortality patterns , and life span , have also been statistically estimated under controlled conditions in the laboratory for a variety of triatomine species ( about 400 scientific articles have been written on these aspects since 1910 ) . Recently , much work has been done on some key evolutionary aspects of triatomines , such as phylogeny , speciation , dispersal and domiciliation , which are now better understood in terms of population genetics ( e . g . , [14]–[21] ) . However , no effort has been done to apply evolutionary ecology concepts for a better understanding of other relevant aspects of their biology , such as development , reproduction and survival strategies . In this work we adopted this approach to shed some light on the evolution of the variability in developmental time in triatomines , and its ecological and epidemiological consequences . Most species of triatomines live in environments ( climate , food sources , predation ) likely to result in conditions for survival and fecundity unpredictably variable in time , to a more or less wide extent depending on their geographical distribution and habitat . Life-history stochastic theory ( e . g . , [6] , [22]–[27] ) predicts that in unpredictable environments such as those probably faced by many triatomine species , the time to reach the adult ( reproductive ) stage is of major importance because a strong reduction in fitness would be expected if they reach that stage during inadequate conditions for parent reproduction and survival of offspring . When the quality of these conditions varies randomly , variability in life-cycle duration within a population resulting from individual variation in the developmental time is usually interpreted as an adaptive spreading of risk strategy called diversified bet-hedging ( [25] , [28] for synthesis ) . Diapause in a given proportion of individuals and/or individual variability in the juvenile development duration are usual strategies to produce such life cycle variability in response to environmental stochasticity [5] , [6] . However , the study of diapause , which leads to a developmental delay , though it is considered as a major adaptive trait in many insects ( e . g . , [3] , [4] ) , has been neglected in the triatomines . A few authors [29]–[32] have suggested the possibility of a diapause in some triatomine species but this hypothesis has not been tested nor its possible adaptive value been considered . We have observed in triatomine insectaries a delay in the molt of a small proportion of individuals . Descriptions available in the literature do not allow to conclude if this delay could represent natural individual variation in developmental times or laboratory artifacts . In order to investigate the existence of such variability and the underlying adaptive value , we have: ( 1 ) carried out a survey on the observation and possible reasons of its occurrence among researchers of the Chagas disease with experience in triatomine rearing , ( 2 ) statistically described the available data on life-cycles for eleven species of triatomines , ( 3 ) analyzed relationship between development time pattern and climatic inter-annual variability , and ( 4 ) developed a mathematical optimization model , including diapause , to get possible adaptive explanations and predictions . We illustrate the importance of investigating the developmental delay in insect vectors in the light of evolutionary ecology by focusing our paper on the bet-hedging diapause hypothesis as well as of its links to the epidemiology of Chagas disease and climatic global change . However , alternatives to bet-hedging diapause strategies are also discussed .
The model predicts the optimal frequency of diapause ( i . e . , the frequency of a delayed development such that it maximizes the mean geometric fitness ) under the assumption , among others , that such a frequency can be coded genetically . The output of the model is a given probability of diapause ( x ) for a given genotype , comprised between x = 0 ( no diapause ) and x = 1 ( obligatory diapause ) , and an intermediate ( 0<x<1 ) probability of facultative diapause in between . Our simulations , that were carried out covering a wide range of values in demographic parameters , show that a non-zero optimal frequency of diapause is expected ( i . e . , diapause should evolve ) if survival throughout the diapause period and the probability of random occurrence of “bad” environmental conditions ( for fecundity and/or survival in active juveniles ) are sufficiently high . Fig . 4 shows an example when the environmental stochasticity is applied only on fecundity ( fecundity is reduced 1000 times in the “bad” periods ) . Fig . 4A ( in which the probability of “bad” periods was fixed to 0 . 7 ) shows that the expected frequency of diapause is higher than 0 . 2 ( i . e . , 20% of individuals should show delayed development ) when survival throughout diapause is equal or larger than 0 . 8 . Fig . 4B , in which the survival throughout diapause is fixed to 0 . 9 , shows that the frequency of diapause is higher than 0 . 1 when the probability of random occurrence of “bad” conditions is higher than 0 . 6 . Similar trends are obtained if environmental stochasticity is applied only on the survival of non-diapausing juveniles ( results not shown ) . We can compare the model predictions when stochasticity affects separately only the fecundity or only the survival of non-diapausing juveniles with the predictions when it affects both parameters simultaneously and independently . In both cases the same expected frequency of diapause is obtained from the model , but only if the environmental stochasticity affecting survival and fecundity simultaneously and independently is assigned a lower probability of “bad” periods . E . g . , if the survival of diapausing juveniles is fixed to 0 . 9 , and that the level of stochasticity applied separately on survival of non-diapausing juveniles and on adult fecundity with a probability of “bad” periods of 0 . 7 , then the model predicts that fitness is maximized when there is an expected frequency of diapause of 0 . 4; however , to obtain the same expected frequency of diapause of 0 . 4 when the environmental stochasticity affects the survival of non-diapausing juveniles and the fecundity simultaneously and independently , the level of stochasticity applied to those two parameters has to be fixed at only 0 . 45 ( 64% smaller than 0 . 7 ) .
The results from the survey ( Tables 1 and 2 ) and the empirical information we have collected from unpublished and published data ( Fig . 2 ) suggest that a life cycle development delay exists in some triatomines ( albeit affecting a relatively small number of individuals ) , and that such delay may even take place under the constant environmental conditions used in the laboratory . This delay could be viewed either as a simple non-adaptive phenotypic plasticity in development time due to a lower feeding capacity in some individuals ( or other non-adaptive process ) , or as a form of adaptive diapause , possibly associated with a physiological mechanism in relation with the postponement of the initiation of reproduction in some individuals , minimizing the risk to reproduce when a unfavorable environmental condition occurs . Indeed , as we show with our model , the development delay observed in triatomines could be an adaptation to environmental stochasticity ( e . g . , unpredictability of host availability , predation risk or conditions above or below the non-diapausing triatomine tolerance for climatic conditions ) through a spreading of risk ( bet-hedging ) strategy . According to bet-hedging theory ( [6] , [24] , [25] , [40]–[42] ) , variability in the life cycle duration with the bi-modal pattern observed in some triatomine species may be the result of an adaptive variability in the phenotypic expression ( diapause versus non-diapause and/or variability in diapause duration ) by a given genotype . More generally , the differences in developmental time in triatomines could be explained by a concave fitness set on a coarsely grained environment as expected by the concept of fitness sets [43] . Indeed , in such case , the optimum is a mixed strategy in which the two specialized phenotypes occur in proportions depending on the probability of different ( “good” or “bad” ) environment states . In our model the level of environmental stochasticity needed for diapause to be selected is relatively high , and probably unrealistic in most cases . However , it is worth noting that diapause could also evolve in an environment without stochasticity in response to density-dependence , or because of a complex interaction between environmental stochasticity and density dependence [7]–[10] , [44] . More detailed models should assume such density-dependence , and may predict a smaller level of environmental stochasticity to select for diapause . Thus one of the main challenges will be to estimate in the field the level of environmental stochasticity and the importance of the density-dependence mechanisms . Development delays in the 5th instar nymphs have been qualitatively observed in 86% of the cases collected in the survey . This stage is one of the most “resistant” stages in triatomines; e . g . , the 5th instar nymph is one of the stages most resistant to starvation in many triatomine species: this characteristic has been confirmed in T . brasiliensis [45] , [46] , in T . lecticularia [47] , in T . tibiamaculata , P . megistus , and in R . neglectus [48] , in Dipetalogaster maximus [49] , in Cavernicola lenti [50] , in T . vitticeps [51] , and in T . melanosoma [52] . Moreover , the 5th instar nymphs are the most resistant to insecticides in T . infestans and R . prolixus [53]–[55] . Thus , one should expect the evolution of a diapause of variable length , if favored ( e . g . , because of environmental stochasticity ) , to take place most probably in this stage than in others . It is also important to remark that although the survey responses derived mostly from studies conducted in the laboratory under controlled conditions of temperature and food , at least two experimental studies conducted in semi-natural conditions ( chicken coops in the field ) with T . infestans [56] and T . dimidiata [57] have also shown developmental delays in the 5th instar nymphs . The qualitative data of the survey presented here were taken as a first step to address the problem of a potential developmental delay in triatomines but it was not intended to investigate development delay per se ( e . g . , the difference in the frequency of individual delaying their development among species ) . The observed bi-modal pattern in developmental time of various triatomine species could be expected both from maladaptive mechanism ( e . g . , those associated to infection by symbionts or viruses ) and from adaptive mechanisms as diapause based on variability among individuals with a given genotype . Viral infections may affect developmental times has been confirmed at least in T . infestans [58] . If infection or parasitism affect feeding capacity and/or other physiological vital functions , the infected individuals with developmental delay are expected to show a lower energetic level than those without delay ( not infected and/or not parasitized ) . Conversely , if developmental delay results from adaptive diapause , one may expect that individuals with delay have equal or more energetic resources than those without delay , as shown in the European chestnut weevil [59] , [60] and a bee species living in arid areas [61] , because energy is needed to survive throughout diapause and to survive and reproduce after the diapause [3] , [4] , [59] . In consequence , we postulate that individuals in a diapause state should have equal or better energetic condition than non-diapausing individuals . Under these conditions individuals in a diapause state are expected to be more resistant to environmental stresses ( biotic or abiotic ) than active individuals [3] , [4] . One may also expect that diapausing triatomines should be able to take refuge until molting into the adult stage takes place , which would increase their resistance . Kissing bugs egg parasitoidism by microhymenoptera such as Ooencyrtus trinidadensis , Telenomus fariai and T . costa-limai [62]–[67] is a good example of a non-climatic ( i . e . , biological ) mortality risk that might provide a high level of stochasticity , leading to the selection of a bet-hedging developmental diapause . Telenomus fariai is more adapted to different climatic conditions ( e . g . , tropical versus temperate ) than to different triatomine host species [68] , and effective parasitism by this microhymenopteran species depends on an adequate time synchronization between the adult female parasitoid and the egg stage of the triatomines; a triatomine strategy of a few adult females delaying their appearance ( as bet-hedging developmental diapause ) would be adaptive in allowing some females to produce eggs during period with no or few parasitoids . The temporal ( among years ) climatic variability could also be one of determinants of selection for diapause bet-hedging . However , the triatomine species position in our four categories of development time patterns does not seem to fit straightforwardly in relation to the combination of variability ( coefficient of variation ) of the precipitation and temperature . However , this climatic analysis must be viewed only as a preliminary study using existing information to relate life history traits and the climatic environment . To rigorously test the hypothesis of a bet-hedging diapause strategy and its alternative hypotheses ( see below ) , and because laboratory strains of triatomines have been reared for many generations under stable conditions very different from field conditions , we need ( 1 ) to sample natural populations of different triatomine species to estimate their developmental time distribution , ( 2 ) a knowledge of the local climatic temporal variability of the corresponding populations , and ( 3 ) the relationship between the developmental time pattern ( and its corresponding fitness ) and the local climatic variables . Indeed , it is the local level of environment stochasticity ( in climatic and biological factors ) as well as the microhabitat conditions which select for a given frequency of diapause as recently showed by local adaptation theory ( e . g . , [10] ) . If we consider , as it is usually proposed , that the domestic environment is more stable than the sylvatic one , sylvatic species should show more frequent diapause than domestic species . Because T . infestans and R . prolixus are two of the most domiciliated triatomines ( a habitat that “cushions” most of the climatic variability ) one could expect these two species not to show bi-modality in the development time pattern if climatic variability was the key factor to select bet-hedging diapause . However , these two species show a clear bi-modal development time pattern , but we cannot conclude that the frequencies in delayed development observed in these species are maladapted without confronting them to the predictions of a model that considers realistically their demography and the environmental conditions they encounter , as discussed above . For instance , it is likely that the stable conditions encountered in the domestic environment lead to a more prominent role of density-dependence ( able to select variation in development time ) , because of higher and more regular intakes of food . For simplicity , we have chosen in this paper not to deal with the possibility that the frequency of diapause could vary seasonally if the level of environmental predictability can change with the season . Such plasticity in the frequency of diapause with respect to environmental cues may be viewed as a mixing of bet-hedging and predictive plasticity [69] . More generally , if the triatomines can use environmental cues to anticipate if a future period will be good or bad for development or reproduction , a predictive diapause mechanism would likely be selected as predicted by plasticity theory [24] . The observed bi-modal pattern in developmental duration could also be explained , even in a stable environment , by a genetic polymorphism in developmental time without diapause . For instance , this polymorphism could be maintained if individuals with a long developmental time have a higher fecundity than those with short developmental time . The stability of such a polymorphism should depend on the form of the trade-off between fecundity and age of first reproduction [70] , [71] . Environmental variability could reinforce ( or allow ) such a polymorphism , because individuals with a longer developmental time would be favored when , for instance , no food is available for the adults resulting from competition between short developmental time individuals . This hypothesis is supported by the theoretical work from Taylor [72] showing that selection may favor high variation in the length of the juvenile period , e . g . , when the duration of suitable habitat is variable , individuals that develop quickly can escape at times when habitat duration is short; conversely , when the duration of suitable habitat is longer , slower developing individuals may be favored because they are larger and , therefore , have higher fecundity , as is common in insects . Therefore , future studies should estimate the duration of suitable habitat for triatomines in searching for the selective pressures that shape their development strategies to handle climatic variability . We hope that this paper will stimulate future experimental research aiming to test adaptive developmental delay hypotheses in triatomines . For instance , to test the bet-hedging diapause hypothesis , the question if variability in life cycle duration is expressed by a given genotype , and if such a mixed strategy is the fittest , should be addressed . Ideally pure strains should be used to quantify the frequency of diapausing individuals , and the independence of this frequency to predictive environmental signal should be tested . Since pure strains are difficult to obtain , several groups of individuals randomly sampled could be used and exposed to variable environmental factors . To test the fitness superiority of a mixed strategy , one must show: ( 1 ) that the resistance to environmental stress in nymphs with development delay is higher than in individuals without development delay , and that they can molt to the adult stage and reproduce , and ( 2 ) that the variability in development time matches the expectation of a realistic model including a mechanism of density-dependence . Both laboratory and field experiments will be needed because the frequency of developmental delay under laboratory conditions is possibly underestimated , as shown in studies on insects with prolonged diapause [73] , [74] . Evolutionary ecology concepts may allow the understanding of the evolution of developmental delay in the triatomines and other vector species; this understanding is not only important for basic and academic considerations , but also because of the epidemiological and applied ecological consequences . For instance , development delay , resulting from bet-hedging diapause strategies , may contribute to stabilize triatomine population dynamics by minimizing extinction risk [6] , [27] , thus increasing the overall risk of disease transmission . Furthermore , diapausing individuals with low metabolic rate and protected in refuges may be more resistant to insecticide as seen in some insect species [3] , [4] . Therefore , diapausing individuals could represent competent reservoirs of pathogens . After insecticide spraying effects have subsided , diapausing individuals could permit the reproduction after they exit their diapause condition , reducing the effectiveness of the control measures and maintaining the transmission of the disease . This problem should be encountered in areas where environmental stochasticity is high but also in areas where environmental stochasticity may increase in the future because of global climatic change . Indeed , some global climatic models predict not only an increase or decrease in mean temperature ( which affects the developmental time and the resistance to insecticides in triatomines; see [75] , [76] ) but also changes in the variability of temperature and precipitation , with an increase of exceptionally hot and dry seasons in some years . Therefore , global climatic changes may result in an increase of environmental stochasticity . Since bet-hedging diapause is an adaptive response to this kind of stochasticity , the possible adaptive responses of triatomines in relation to global climatic changes may consist of: ( 1 ) an increase in the frequency of diapause ( expressed as a developmental delay ) in areas changing from moderate to highly stochastic environment , and ( 2 ) a selection for diapause in areas changing from low to moderate or high stochasticity . Such ecological changes will result in epidemiological consequences since diapausing individuals will represent competent pathogen reservoirs as discussed above . The likelihood of defecation and of T . cruzi prevalence in triatomines increases with stage [77] and so also does the transmission of T . cruzi . Under the hypothesis of the trade-off between fecundity and age of first reproduction ( alternative to bet-hedging diapause , see above ) , as the individuals that take a longer time to develop will have a higher number of feeding instances than those individuals without developmental delay , an increase of the overall vectorial capacity of triatomines is expected . Additionally , a fluctuating environment compels insect populations away from a stable age distribution , and Taylor [78] has shown that the return time to a stable age distribution increases with a delay in the age of first reproduction; in consequence , a diapause resulting in a development delay , due to a seasonal environment could lead to an unstable age structure of triatomine populations , possibly increasing the risk of transmission due to a proportional increase of adults and older juveniles in the population . Recent work predicts and confirms bet-hedging strategies in pathogen microorganisms [79]–[81]; thus research is also needed in Trypanosoma cruzi , the etiological agent of Chagas disease , as well as in other pathogens responsible of other diseases , in order to investigate the possibility of bet-hedging increasing the fitness of the pathogen in an uncertain world . Research focused on the capacity of the pathogen to manipulate developmental time of insect vectors is also needed . Epidemiological consequences of these two possible pathogen adaptive strategies may be extremely important both for academic and applied considerations . | The developmental time of vector insects is important to their population dynamics , evolutionary biology , epidemiology of the diseases they transmit , and to their responses to global climatic change . In various triatomine species vectors of Chagas disease ( Triatominae , Reduviidae ) , a delay in the molt of a small proportion of individuals has been observed , and from an evolutionary ecology approach , we propose the hypothesis that the developmental delay is an adaptation to environmental stochasticity through a spreading of risk ( bet-hedging ) diapause strategy . We confirmed , by means of a survey among specialists , the existence of the developmental delay in triatomines . Statistical descriptions of the developmental time of 11 species of triatomines showed some degree of bi-modality in nine of them . We predicted by means of an optimization model which genotype , coding for a given frequency of developmental diapause , is expected to evolve . We identified a series of parameters that can be measured in the field and in the laboratory to test the hypothesis of an optimal diapause frequency . We also discuss the importance of these findings for triatomines in terms of global climatic change and epidemiological consequences such as their resistance to insecticides . | [
"Abstract",
"Introduction",
"Results",
"Discussion"
] | [
"ecology/evolutionary",
"ecology",
"public",
"health",
"and",
"epidemiology",
"evolutionary",
"biology/evolutionary",
"ecology"
] | 2010 | Adaptive Developmental Delay in Chagas Disease Vectors: An Evolutionary Ecology Approach |
The insulin/IGF-1 signaling ( IIS ) pathway is a conserved regulator of longevity , development , and metabolism . In Caenorhabditis elegans IIS involves activation of DAF-2 ( insulin/IGF-1 receptor tyrosine kinase ) , AGE-1 ( PI 3-kinase ) , and additional downstream serine/threonine kinases that ultimately phosphorylate and negatively regulate the single FOXO transcription factor homolog DAF-16 . Phosphatases help to maintain cellular signaling homeostasis by counterbalancing kinase activity . However , few phosphatases have been identified that negatively regulate the IIS pathway . Here we identify and characterize pdp-1 as a novel negative modulator of the IIS pathway . We show that PDP-1 regulates multiple outputs of IIS such as longevity , fat storage , and dauer diapause . In addition , PDP-1 promotes DAF-16 nuclear localization and transcriptional activity . Interestingly , genetic epistasis analyses place PDP-1 in the DAF-7/TGF-β signaling pathway , at the level of the R-SMAD proteins DAF-14 and DAF-8 . Further investigation into how a component of TGF-β signaling affects multiple outputs of IIS/DAF-16 , revealed extensive crosstalk between these two well-conserved signaling pathways . We find that PDP-1 modulates the expression of several insulin genes that are likely to feed into the IIS pathway to regulate DAF-16 activity . Importantly , dysregulation of IIS and TGF-β signaling has been implicated in diseases such as Type 2 Diabetes , obesity , and cancer . Our results may provide a new perspective in understanding of the regulation of these pathways under normal conditions and in the context of disease .
Insulin/IGF-1 signaling ( IIS ) is a conserved neuroendocrine pathway that regulates longevity , development and energy metabolism across phylogeny [1] , [2] . In the roundworm Caenorhabditis elegans ( C . elegans ) , activation of the DAF-2 insulin/IGF-1 receptor tyrosine kinase intiates an AAP-1/AGE-1 PI 3-kinase signaling cascade involving the downstream serine/threonine kinases PDK-1 , AKT-1 , and AKT-2 [3]–[7] . Activated AKT-1 and AKT-2 phosphorylate DAF-16 , the single Forkhead Box O ( FOXO ) family transcription factor homolog in C . elegans [8] . Phosphorylation of DAF-16 results in its inactivation and sequestration in the cytosol [9] , [10] . Under low signaling conditions , DAF-16 translocates into the nucleus , where it can transactivate/repress hundreds of target genes [9]–[13] . The dauer is an alternative survival stage that worms can enter upon poor environmental conditions such as crowding [14] . Mutations in the kinases upstream of DAF-16 such as daf-2 , age-1 , pdk-1 , akt-1 and akt-2 result in an increase in lifespan , dauer formation , fat storage and/or stress resistance , and loss-of-function mutations in daf-16 completely suppress these phenotypes [15]–[18] . In addition to the IIS pathway , dauer formation in C . elegans is also regulated by the DAF-7/TGF-β-like signaling pathway [19]–[21] . Activation of TGF-β signaling is achieved through binding of the DAF-7 BMP-like ligand to the DAF-1/DAF-4 , the Type I/II receptors , which phosphorylate and activate the downstream receptor-associated SMAD ( R-SMAD ) proteins DAF-8 and DAF-14 , presumably through a conserved SSXS phosphorylation motif that has been shown to be important for R-SMAD activation in mammals [22]–[24] . Upon activation , R-SMADs can associate with a Co-SMAD to regulate the transcription of hundreds of genes [23] , [25] . In C . elegans , DAF-8 and DAF-14 act to antagonize the transcriptional activity of the DAF-3 Co-SMAD and the DAF-5 SNO-SKI repressor [22] , [24] , [26]–[29] . Reduction of function mutations in daf-7 , daf-1 , daf-4 , daf-8 and daf-14 show temperature-sensitive constitutive dauer formation and mutations in daf-3 and/or daf-5 completely suppress this phenotype [19] , [21] , [30] . Genetic epistasis studies have suggested that the TGF-β pathway acts in a parallel manner with IIS to modulate dauer formation [31]–[33] . The PTEN lipid phosphatase homolog DAF-18 , which antagonizes signaling at the level of AGE-1/PI 3-kinase , is a major negative regulator of IIS . In contrast to the kinases in this pathway , loss-of-function mutations in daf-18 reduces lifespan , fat storage , dauer formation and stress resistance [32] , [34]–[39] . Besides DAF-18 , few negative modulators of the pathway have been identified . In particular , less is known about serine/threonine phosphatases that counterbalance kinase activity in the IIS pathway . We recently performed a directed RNA interference ( RNAi ) screen for serine/threonine phosphatases that regulate C . elegans IIS using dauer formation as an output [39] . We identified the PP2A regulatory subunit PPTR-1 as an important regulator of AKT-1 dephosphorylation as well as DAF-16-dependent phenotypes [39] . Here we characterize another candidate from this screen , pdp-1 , as a positive regulator of dauer formation . PDP-1 is homologous to pyruvate dehydrogenase phosphatase ( PDP ) in higher organisms , an enzyme that positively regulates the pyruvate dehydrogenase enzyme complex ( PDHc ) . RNAi of the other components of PDHc do not result in changes in dauer formation . Interestingly , we report that although PDP-1 is a robust modulator of multiple IIS-regulated processes as well as DAF-16 activity , genetic epistasis studies place pdp-1 in the DAF-7/TGF-β pathway . Through this study , we find that IIS and TGF-β signaling are more tightly connected than previously suggested , with distinct roles for the Co-SMAD DAF-3 in modulating the IIS pathway . Our data suggests that PDP-1 modulates the gene expression of several insulins , and that insulins may be a potential mediator of the crosstalk between these two pathways .
Our RNAi screen was designed to identify serine/threonine phosphatases that modulated dauer formation of daf-2 ( e1370 ) , a non-null , temperature-sensitive mutant of the C . elegans insulin/IGF-1 receptor gene , daf-2 [39] . We were particularly interested in phosphatases that would negatively regulate IIS similar to DAF-18/PTEN , and for all RNAi based assays described below , daf-18 RNAi was used as a positive control [39] . From this screen , we identified pdp-1 as a modulator of daf-2 ( e1370 ) dauer formation ( Figure 1A and Figure S2 ) . BLAST analyses using amino acid sequence revealed that PDP-1 is homologous to fly and mammalian PDP ( ∼52% positive and ∼38% identical ) . pdp-1 RNAi significantly reduces dauer formation of daf-2 ( e1370 ) worms , similar to daf-18 RNAi ( Figure 1A and Figure S2 ) . This phenotype is not allele-specific , as pdp-1 RNAi results in suppression of dauer formation in a second allele of daf-2 , daf-2 ( e1368 ) ( Figure 1B and Figure S2 ) . Similar to the results with the RNAi , a mutation in pdp-1 also affects dauer formation - pdp-1 ( tm3734 ) ; daf-2 ( e1370 ) double mutants form significantly fewer dauers when compared to the daf-2 ( e1370 ) parental strain ( Figure S2 ) . Given its homology to PDP in higher organisms , we wondered whether the effect of pdp-1 knockdown on daf-2 dauer formation was a consequence of modulating the activity of the PDHc . The PDHc is a multi-subunit enzyme complex consisting of three major enzymes: E1 pyruvate dehydrogenase , E2 dihydrolipoyl acetyltransferase and E3 dihydrolipoyl dehydrogenase that regulate energy metabolism [40] . PDHc converts pyruvate to acetyl-coA , which can either enter the TCA cycle or be used for fatty acid synthesis . In mammals , regulation of PDHc activity is primarily achieved through reversible phosphorylation/dephosphorylation of the E1α subunit by pyruvate dehydrogenase kinase ( PDHK ) and PDP , with phosphorylation inactivating the enzyme complex [40] . All of the components of the PDH complex have conserved C . elegans homologs , encoded by the genes T05H10 . 6 ( E1α ) , C04C3 . 3 ( E1β ) , F23B12 . 5 ( E2 ) , LLC1 . 3 ( E3 ) , pdhk-2 ( PDHK ) and pdp-1 ( PDP ) . To test whether modulation of PDHc activity affects daf-2 dauer formation , we grew daf-2 ( e1370 ) worms on PDHc RNAi . Quantification the RNAi efficiency of the PDHc components revealed that we achieved 60–90% knockdown ( Figure S1 ) . To our surprise , RNAi of the E1α subunit had no effect on daf-2 dauer formation , while pdp-1 RNAi resulted in dauer suppression ( Figure 1C and Figure S2 ) . In addition , RNAi of either the other E1 subunit E1β , or the E2 subunit , did not affect daf-2 dauer formation ( Figure 1C and Figure S2 ) . Knockdown of the E3 subunit resulted in lethality ( data not shown ) . Interestingly , pdhk-2 RNAi resulted in slight suppression daf-2 ( e1370 ) dauer formation but had no effect on dauer formation of daf-2 ( e1368 ) mutants ( Figure 1C and Figure S2 ) . Therefore pdhk-2 modulates the IIS pathway in an allele-specific manner and we did not perform further characterization of this gene . To further evaluate the components of the PDH complex , we examined their expression patterns . The expression pattern of PDP-1 does not completely overlap with that of the E1 and E2 subunits of PDHc ( Figure S3 ) . PDP-1 expression was enriched in the head and tail neurons , head muscle and the intestine . We did not observe any expression in the pharynx . In contrast , the expression of the E1 and E2 subunits , was observed throughout the body of the worm and was significantly enriched in the pharynx . Taken together , PDP-1 modulates daf-2 dauer formation and this function is likely to be independent of its role in regulating the PDHc . In addition to dauer formation , the IIS pathway also regulates longevity , stress resistance and fat storage [17] , [18] . Mutations in daf-2 and age-1 result in a significant extension in lifespan , enhanced resistance to various stresses and increased fat storage [7] , [35] , [41]–[44] . These phenotypes are suppressed by loss-of-function mutations in daf-18 and daf-16 [32] , [34] , [35] , [39] . We therefore investigated whether dosage modulation of pdp-1 would affect additional outputs of the pathway . We first tested the role of PDP-1 in regulating lifespan ( Figure 2 and Figure S4 ) . The lifespan of wild-type worms was not affected by pdp-1 RNAi and slightly reduced by a mutation in pdp-1 ( Figure 2A and 2D ) . In contrast , the mean and maximal lifespan of long-lived daf-2 ( e1370 ) and age-1 ( hx546 ) mutants was significantly reduced by pdp-1 RNAi ( Figure 2B and 2C ) . Similarly , pdp-1 ( tm3734 ) ; daf-2 ( e1370 ) double mutants lived significantly shorter than the parental daf-2 ( e1370 ) strain ( Figure S4 ) . To examine the effect of increased dosage of pdp-1 , we generated transgenic worms bearing a translational fusion containing pdp-1 fused to gfp and driven by its own promoter ( pdp-1::gfp ) . In addition , we also crossed the pdp-1::gfp worms to daf-2 ( e1370 ) mutants to generate the daf-2 ( e1370 ) ; pdp-1::gfp strain . Overexpression of pdp-1 results in a significant extension in lifespan compared to wild-type worms ( Figure 2D and Figure S4 ) . Interestingly , pdp-1 overexpression further extends the lifespan of daf-2 ( e1370 ) mutants ( Figure 2B and Figure S4 ) . In both of these cases , the increased lifespan was suppressed by daf-16 RNAi ( Figure S5 ) . Therefore , dosage modulation of pdp-1 regulates lifespan in a DAF-16 dependent manner . Next , we asked if PDP-1 modulated additional outputs of the IIS signaling pathway . We first tested whether PDP-1 regulates stress resistance by assaying the survival of pdp-1 mutants and transgenic animals when exposed to heat stress at 37°C ( Figure 2E and Figure S7 ) . Dosage modulation of pdp-1 affects the response to heat stress , with a pdp-1 mutation decreasing and pdp-1 overexpression slightly enhancing thermotolerance ( Figure 2E ) . Importantly a pdp-1 mutation drastically reduced the thermotolerance of daf-2 mutants ( Figure 2E ) . To examine the role of pdp-1 in regulating fat storage , we used both Oil Red O [45] and Sudan Black [7] staining ( Figure 2F and 2G and Figure S7 ) . pdp-1 mutants had similar levels of fat compared to wild-type worms , while overexpression of pdp-1 slightly enhanced fat storage ( Figure S7 ) . In contrast , a pdp-1 mutation drastically reduced the increased fat of daf-2 ( e1370 ) mutants ( Figure 2F and 2G and Figure S7 ) . This was observed in dauers , larval stage 3 ( L3 ) animals and adults , suggesting that PDP-1 is an important regulator of fat storage in daf-2 mutants . We did observe any further enhancement of the increased fat storage in the daf-2 ( e1370 ) ; pdp-1::gfp worms ( Figure S7 ) . Importantly , the increased fat storage of pdp-1::gfp and daf-2 ( e1370 ) ; pdp-1::gfp worms was suppressed by daf-16 RNAi , similar to daf-2 mutants ( Figure S7 ) . Thus , PDP-1 modulates all four well-characterized outputs of the IIS pathway . In addition to these phenotypes , pdp-1 ( tm3734 ) mutants exhibit a slow movement phenotype , which we quantified using locomotion assays ( Figure S6 ) . This slow movement was rescued by the pdp-1::gfp transgene . In addition , we performed brood size analysis of wild-type , pdp-1 ( tm3734 ) mutants , daf-2 ( e1370 ) mutants , and pdp-1 ( tm3734 ) ; daf-2 ( e1370 ) double mutants ( Figure S6 ) . pdp-1 ( tm3734 ) worms showed a slight decrease in the number of progeny compared to wild-type worms . However , when compared to daf-2 mutants , only 5% of the pdp-1 ( tm3734 ) ; daf-2 ( e1370 ) eggs yielded progeny ( Figure S6 ) . daf-2 mutants have a slightly reduced brood size [46] , [47] , and a mutation in pdp-1 severely enhances this phenotype . Taken together , PDP-1 regulates multiple outputs of IIS and acts as a negative regulator the pathway , similar to DAF-18/PTEN . The FOXO transcription factor DAF-16 is the major target of the C . elegans IIS pathway [2] , [48] . Under conditions of reduced IIS , DAF-16 is able to translocate into the nucleus , where it regulates the expression of hundreds target genes [12] , [13] , [49] , [50] . We therefore asked whether PDP-1 modulates DAF-16 subcellular localization as well as activity ( Figure 3A and Figure S8 ) . daf-2 ( e1370 ) ; daf-16::gfp worms were grown on vector , daf-18 and pdp-1 RNAi , and DAF-16 nuclear/cytosolic localization was visualized using fluorescence microscopy and quantified . Throughout the body of the worm , while DAF-16::GFP was mostly nuclear on vector RNAi , its localization was enriched in the cytosol on pdp-1 RNAi , similar to daf-18 RNAi ( Figure 3A and Figure S8 ) . The gene superoxide dismutase 3 ( sod-3 ) is a direct DAF-16 target [11] . To test whether PDP-1 modulates transcriptional activity of DAF-16 , we used a Psod-3::gfp reporter strain in a daf-2 ( e1370 ) background [51] . daf-2 ( e1370 ) ; Psod-3::gfp worms were grown on vector , pdp-1 , daf-18 and daf-16 RNAi and GFP expression was visualized using fluorescence microscopy and scored as low , medium or high ( Figure 3B and Figure S8 ) . GFP expression was markedly lower on pdp-1 RNAi compared to vector RNAi , similar to daf-18 and daf-16 RNAi , suggesting that PDP-1 positively modulates DAF-16 transcriptional activity . To further validate these results , we used quantitative real-time PCR ( Q-PCR ) to look at the expression levels of well-known DAF-16 target genes [52] in daf-2 ( e1370 ) , pdp-1 ( tm3734 ) ; daf-2 ( e1370 ) and daf-16 ( mgDf50 ) ; daf-2 ( e1370 ) worms ( Figure 3C ) . Notably , the expression of sod-3 , sod-5 and hsp-12 . 6 was significantly reduced in pdp-1 ( tm3734 ) ; daf-2 ( e1370 ) mutants relative to daf-2 ( e1370 ) . Therefore PDP-1 positively regulates a subset of DAF-16 targets . Thus far our data indicates that PDP-1 regulates multiple outputs of IIS as well as DAF-16 activity . Using dauer formation as the readout , we performed genetic epistasis experiments to identify the substrate of PDP-1 . We first tested whether pdp-1 acted directly through the IIS pathway by focusing on kinase mutants downstream of daf-2 ( Table 1 and Figure S9 ) . pdk-1 ( sa680 ) , daf-2 ( e1370 ) ; akt-1 ( ok525 ) and daf-2 ( e1370 ) ; akt-2 ( ok393 ) mutants were maintained on vector , daf-18 and pdp-1 RNAi and dauer formation of these strains was assayed at the appropriate temperatures . Interestingly , pdp-1 RNAi resulted in suppression of dauer formation of pdk-1 ( sa680 ) mutants , daf-2 ( e1370 ) ; akt-1 ( ok525 ) and daf-2 ( e1370 ) ; akt-2 ( ok393 ) worms ( Table 1 and Figure S9 ) . DAF-16 is downstream of the AKT kinases in the pathway , but we were unable to detect a physical interaction between PDP-1 and DAF-16 ( data not shown ) . We next examined a TGF-β pathway that also regulates dauer formation [19]–[21] using genetic epistasis analyses with mutants of this pathway . In these assays , TGF-β pathway mutants were maintained on vector RNAi , pdp-1 RNAi and daf-3 RNAi ( as a positive control; Table 2 and Figure S10 ) . We first tested daf-7 mutants , which contain a mutation in the gene encoding the TGF-β ligand [53] . Dauer formation of daf-7 ( e1372 ) mutants was suppressed on pdp-1 RNAi similar to daf-3 RNAi , suggesting that pdp-1 does not function at the level of daf-7 ( Table 2 and Figure S10 ) . Next , we tested dauer formation with mutants of the SMADS daf-8 and daf-14 [22] . We grew daf-14 ( m77 ) mutants on vector , pdp-1 and daf-3 RNAi . Interestingly , pdp-1 RNAi had no effect on daf-14 dauer formation , while daf-3 RNAi still resulted in suppression ( Table 2 and Figure S10 ) . We next looked at dauer formation of daf-8 ( m85 ) mutants and again observed that pdp-1 RNAi had no effect , while daf-3 RNAi suppressed dauer formation ( Table 2 and Figure S10 ) . Therefore , our genetic epistasis results indicate a genetic interaction between pdp-1 and daf-14/daf-8 . To confirm these results , we investigated whether pdp-1 RNAi could suppress dauer formation of daf-2 ( e1370 ) ; daf-3 ( mgDf90 ) double mutants ( Table 2 and Figure S10 ) . In this strain , input from the TGF-β pathway is removed due to the daf-3 null mutation , and dauer formation is presumably mediated through activated DAF-16 [39] . Therefore , if pdp-1 was indeed acting in the TGF-β pathway , we would not see any effect of pdp-1 RNAi on daf-2 ( e1370 ) ; daf-3 ( mgDf90 ) double mutants . Expectedly , pdp-1 RNAi had no effect on daf-2 ( e1370 ) ; daf-3 ( mgDf90 ) double mutants ( Table 2 and Figure S10 ) . DAF-3 itself is unlikely to be a substrate for PDP-1 , as similar to mammalian Co-SMADs , it lacks the SMAD phosphorylation motif [28] . Therefore , our genetic epistasis analysis supports a model whereby pdp-1 acts in the DAF-7 TGF-β pathway at the level of daf-8 and daf-14 . How does a phosphatase in the TGF-β signaling pathway have such robust effects on the outputs of the IIS pathway and DAF-16 ? A number of studies have previously identified roles for the TGF-β pathway in lifespan and fat storage [7] , [54] , [55] . However , genetic epistasis analysis on dauer formation placed DAF-7 TGF-β signaling and IIS as two parallel pathways where components of one pathway did not affect the other [14] , [56] , [57] . Yet in our studies , PDP-1 was able to regulate multiple outputs of IIS . Therefore , we decided to further investigate the potential crosstalk between the IIS and TGF-β signaling pathways . First , we focused on DAF-3 and DAF-5 , which are positive regulators of dauer formation similar to PDP-1 , and asked whether mutations in daf-3 or daf-5 could also affect phenotypes of the IIS pathway [14] , [28] , [29] . We tested lifespan , fat storage , dauer formation and stress resistance of TGF-β pathway mutants in a wild-type as well as daf-2 ( e1370 ) background . ( Figure 4A–4C , Figure S11 , 12 , S13 and Table S1 ) . As previously reported , the lifespan of daf-3 and daf-5 single mutants is slightly shorter than wild-type worms ( Table S1 ) [55] . In our hands , mutations in the upstream components of the TGF-β pathway such as daf-7 and daf-14 enhance dauer formation but do not significantly extend lifespan ( Table S1 and Figure S4 ) . Intriguingly , mutations in daf-3 and daf-5 have opposite effects on daf-2 ( e1370 ) phenotypes . When compared to the daf-2 ( e1370 ) parental strain , daf-2 ( e1370 ) ; daf-3 ( mgDf90 ) mutants lived significantly longer . This was also observed in daf-2 ( e1370 ) ; daf-3 ( e1376 ) worms , which is a weaker allele of daf-3 . In contrast , daf-5 ( e1386 ) ; daf-2 ( e1370 ) double mutants live much shorter than daf-2 ( e1370 ) worms ( Figure 4A , Figure S13 and Table S1 ) . A mutation in daf-5 also decreased the increased lifespan of age-1 ( hx546 ) worms , with age-1 ( hx546 ) ; daf-5 ( e1385 ) double mutants living significantly shorter than the parental strain ( Figure S13 ) . Importantly , for daf-2 worms , the effect of a daf-3 null mutation on lifespan was more pronounced at 20°C where signaling through the IIS pathway is further reduced . Therefore , under low IIS conditions , DAF-3 as well as DAF-5 can modulate longevity . We next tested the role of DAF-3 and DAF-5 on fat storage , dauer formation and stress resistance . Oil Red O staining for fat storage showed comparable levels between daf-2 ( e1370 ) and daf-2 ( e1370 ) ; daf-3 ( mgDf90 ) worms , but markedly lesser amounts of fat in daf-5 ( e1386 ) ; daf-2 ( e1370 ) worms ( Figure 4B top and bottom panel and Figure S12 ) . Similarly , age-1 ( hx546 ) ; daf-5 ( e1385 ) had less fat than age-1 ( hx546 ) worms ( Figure S12 ) . Both daf-3 and daf-5 single mutants have slightly reduced levels of fat when compared to wild-type worms ( Figure S12 ) . A similar trend was seen with our data for dauer formation . daf-2 ( e1370 ) ; daf-3 ( mgDf90 ) worms show significant enhancement of daf-2 ( e1370 ) dauer formation across several temperatures tested , whereas a daf-5 mutation or daf-5 RNAi results in reduced daf-2 ( e1370 ) dauer formation ( Figure 4Ci , Figure 4Cii and Figure S11 ) . In addition , daf-5 ( e1386 ) ; daf-2 ( e1370 ) worms fail to completely arrest at the restrictive temperature of 25°C ( data not shown ) . A mutation in daf-5 also significantly reduces thermotolerance of daf-2 ( e1370 ) worms at 37°C ( Figure S13 ) . Taken together , similar to PDP-1 , DAF-3 and DAF-5 modulate multiple outputs of the IIS pathway . Unexpectedly , we find that while DAF-3 promotes dauer formation under conditions of reduced TGF-β signaling , it negatively regulates dauer formation and longevity under conditions of reduced IIS . To further explore the crosstalk between both pathways , we next asked whether DAF-18 and DAF-16 , which are components of the IIS pathway , affect TGF-β pathway signaling . For this , we assayed dauer formation of TGF-β pathway mutants on daf-18 and daf-16 RNAi ( Table 3 and Figure S10 ) . Interestingly , dauer formation of daf-7 ( e1372 ) , daf-14 ( m77 ) and daf-8 ( m8 5 ) worms was robustly suppressed by daf-16 RNAi . We observed similar results for dauer formation daf-7 ( e1372 ) and daf-14 ( m77 ) mutants on daf-18 RNAi . However , in the case of daf-8 ( m85 ) mutants , daf-18 RNAi had no effect on dauer formation of ( Figure S10 ) , suggesting a complex crosstalk between both pathways . The enhanced dauer formation of daf-2 ( e1370 ) ; daf-3 ( mgDf90 ) is suppressed by both daf-18 and daf-16 RNAi but not pdp-1 RNAi ( Table 3 and Figure S10 ) . Therefore , we not only observe DAF-3 and DAF-5 affecting various phenotypes of the IIS pathway , but also the converse , where DAF-16 and DAF-18 robustly regulates TGF-β dauer formation . These results unravel a more complex interaction between the two pathways , where DAF-16 is likely to be the major downstream effector regulating longevity , dauer formation and other physiological outputs . How can these two pathways , once considered to be parallel to each other , be mechanistically linked ? Thus far our data suggests that PDP-1 , a component of the TGF-β pathway can modulate multiple phenotypes of IIS by positively regulating DAF-16 . In addition , we observe extensive crosstalk between the two pathways at multiple levels . A feed-forward model that has been proposed to connect TGF-β signaling to the IIS pathway suggests insulins as a possible link [55] , [58] . The C . elegans genome encodes 40 insulin genes [59] , [60] ( WormBase 215: www . wormbase . org ) . Studies using mutants and RNAi have characterized some of the insulins as agonists or antagonists of the IIS pathway [13] , [59]–[61] . Importantly , microarray studies have identified several insulin genes that are regulated by TGF-β signaling , including ins-1 , ins-4 , ins-5 , ins-6 , ins-7 , ins-17 , ins-18 , ins-30 , ins-33 , ins-35 and daf-28 [55] , [57] . We tested changes in the levels of these insulins using Q-PCR in TGF-β pathway mutants such as daf-3 ( mgDf90 ) , daf-14 ( m77 ) as well as pdp-1 ( tm3734 ) and compared them to wild-type worms ( Figure 5A–5C , Figure S14 , Tables S2 and S3 ) . Interestingly , both pdp-1 ( tm3734 ) and daf-3 ( mgDf90 ) showed elevated levels of several insulins as compared to wild-type worms ( Figure 5A and Figure S14 ) . In contrast , expression of these insulins was markedly reduced in daf-14 ( m77 ) mutants ( Figure 5B and Figure S14 ) . We next looked at the effects of overexpressing DAF-3 and PDP-1 on insulin gene expression ( Figure 5C and Figure S14 ) . The levels of several insulins are markedly reduced in daf-3::gfp and pdp-1::gfp animals when compared to wild-type worms . Therefore , dosage modulation of DAF-3 and PDP-1 modulates insulin gene expression . INS-4 , for example , has been reported as a positive regulator TGF-β pathway and a suppressor of dauer formation of daf-7 and daf-8 mutants [62] . ins-4 transcript levels were elevated in pdp-1 and daf-3 mutants but reduced in daf-14 . To investigate insulin gene expression regulated by DAF-16 , we tested daf-2 ( e1370 ) , pdp-1 ( tm3734 ) ; daf-2 ( e1370 ) and daf-16 ( mgDf50 ) ; daf-2 ( e1370 ) mutants . Several insulins were changed relative to daf-2 ( e1370 ) worms , with the trend between pdp-1 ( tm3734 ) ; daf-2 ( e1370 ) and daf-16 ( mgDf50 ) ; daf-2 ( e1370 ) being quite similar ( Figure 5D and Figure S14 ) . Interestingly , ins-7 levels were elevated both double mutants ( Figure 5E and Figure S14 ) . Previous studies have shown ins-7 to be an agonist of the IIS pathway as well as a DAF-16 target gene [13] , [63] . In contrast , ins-1 levels were drastically reduced , and INS-1 has been characterized as a potential antagonist of IIS [59] . We did not observe a significant change in ins-18 , another potential DAF-16 target [13] . We also did not detect any appreciable differences in insulin gene expression in daf-16 ( mgDf50 ) single mutants ( Figure S14 ) . In addition , we were unable to detect ins-33 and ins-35 transcripts in all the strains tested , and the trend observed with daf-28 was inconclusive ( Table S2 and S3 ) . Taken together , our results suggest the possibility that insulins downstream of TGF-β signaling mediate at least part of the cross talk between the two pathways . Therefore , PDP-1 would modulate to regulate expression of several insulins that can potentially feed into or antagonize the IIS pathway to regulate DAF-16 and its associated phenotypes .
We identified pdp-1 from a RNAi screen for serine/threonine phosphatases that modulate daf-2 dauer formation . C . elegans PDP-1 is homologous to mammalian pyruvate dehydrogenase phophatase ( PDP ) , a metabolic enzyme that is a positive regulator of the pyruvate dehydrogenase enzyme complex ( PDHc ) . Remarkably , other components of the PDHc in C . elegans do not affect daf-2 dauer formation . Microarray and SAGE studies on dauers have indicated that genes involved in anaerobic metabolism are upregulated while genes involved in the TCA cycle and mitochondrial oxidative phosphorylation are downregulated , suggesting that PDHc activity may not be critical for dauer diapause [64]–[66] . Further , annotations indicate that the C . elegans genome encodes approximately 60 serine/threonine phosphatases , in contrast to the 400 plus protein kinases , suggesting that phosphatases are likely to have a number of cellular substrates [39] , [67] . We find that PDP-1 also regulates longevity , fat storage and stress resistance in addition to dauer formation . Interestingly , these phenotypes are more severe in mutants such as daf-2 and age-1 , where IIS is reduced . Further , PDP-1 positively regulates DAF-16 activity . We reason that PDP-1 function is critical under conditions of stress or low food availability , when DAF-16 activation is required [39] . Intriguingly , genetic epistasis analyses place PDP-1 in the DAF-7/TGF-β pathway , at the level of the R-SMAD proteins DAF-14 and DAF-8 . A recent functional RNAi screen for serine/threonine phosphatases that modulate BMP signaling identified PDP as a SMAD1 phosphatase in Drosophila S2 cells and mammalian 293T cells [68] . Our study complements these findings and reveals a molecular conservation in the role of PDP-1 in regulating TGF-β signaling . Early genetic epistasis studies had suggested that TGF-β signaling and IIS pathways are parallel signaling pathways that modulate dauer diapause [31] . Importantly , in these studies , the conclusion was that both these pathways acted independently , and it was the IIS pathway that regulated longevity and stress resistance [31] , [32] . However , the effect of PDP-1 on DAF-16 activity led us to re-investigate the interaction between the IIS and TGF-β signaling . Previous studies have shown that DAF-3 and DAF-5 are negatively regulated by TGF-β signaling , and function similarly as repressors of gene expression to ultimately promote dauer formation [28] , [29] , [69] , [70] . We find that under conditions of reduced IIS , DAF-3 and DAF-5 affect various outputs of the IIS pathway in opposite ways . DAF-3 in particular regulates IIS depending upon the level of signaling through the pathway ( Figure 6 ) . In our hands , mutants of the TGF-β signaling pathway do not exhibit a pronounced increase in lifespan . However , components of this pathway are important for the long lifespan of mutants in the IIS pathway , as well as other phenotypes such as dauer formation , fat storage and stress resistance . Our epistasis studies reveal that daf-18 and daf-16 RNAi can strongly suppress dauer and fat storage of TGF-β pathway mutants . Together , these results point to a feed-forward model where signals through the TGF-β pathway are relayed to modulate activity of the IIS pathway as well as DAF-16 . Indeed , recent studies have suggested that TGF-β pathway regulates the expression of insulins , leading to a feed-forward model , where signals from the TGF-β pathway are relayed to modulate activity of the IIS pathway as well as DAF-16 [55] , [58] . In support of this model , we find TGF-β signaling regulates the expression of several insulin genes with DAF-3 and PDP-1 negatively modulating insulin gene expression . This is in agreement with previous studies that identify DAF-3 as a repressor of gene expression [69] , [70] . The expression of several insulins is also modulated by DAF-16 , with pdp-1 ( tm3734 ) ; daf-2 ( e1370 ) and daf-16 ( mgDf50 ) ; daf-2 ( e1370 ) worms showing similar trends in insulin levels . Therefore , in the absence of PDP-1 , increased levels of agonists or reduced levels of antagonists hyperactivate the DAF-2 pathway to negatively regulate DAF-16 , thereby affecting the enhanced lifespan , stress resistance , dauer formation and fat storage of daf-2 mutants . Our results suggest a model where under favorable growth conditions , signals through the TGF-β pathway activate the SMAD transcriptional complex to regulate the expression of insulins that activate the IIS pathway to phosphorylate and inhibit DAF-16 activity , thereby promoting growth , reproduction and normal lifespan ( Figure 6 , top panel ) . However , when food is limiting or under harsh survival conditions , TGF-β signaling is downregulated by PDP-1 to activate DAF-3 and DAF-5 , to regulate the repression of insulin genes that may feed into the IIS pathway ( Figure 6 , middle panel ) . DAF-3 has also been reported to negatively regulate daf-7 and daf-8 gene expression in a feedback loop [24] . We find that pdp-1 expression is elevated in daf-3 ( mgDf90 ) mutants , suggesting a similar feedback regulation ( Figure S15 ) . Repression of TGF-β and insulin gene expression by DAF-3 results in a reduction in signaling through the IIS pathway , and promotes DAF-16 nuclear localization . DAF-16 then regulates the transcription of hundreds of target genes that ultimately modulate longevity , stress resistance , dauer formation and fat storage . Under low TGF-β signaling and IIS conditions , DAF-3 and DAF-5 regulate these outputs in an opposite manner , with DAF-5 synergizing and DAF-3 antagonizing DAF-16 function ( Figure 6 lower panel ) . With our Q-PCR data , we found that PDP-1 affected only a subset of the DAF-16 target genes tested . These could represent genes that are regulated by DAF-16 and SMAD proteins . SMAD proteins have low affinity for binding DNA , and the orchestration of cellular signals into defined outputs requires their association with additional co-factors [71] . Mammalian SMAD proteins can bind several co-activators and co-repressor proteins to modulate gene transcription [23] . Specifically , a synergy between mammalian FOXO ( FOXO1 , FOXO3a and FOXO4 ) and SMAD2/3 was identified for the regulation of several genes involved in cell cycle regulation and the response to stress [72] . Importantly , these interactions required the function of the co-SMAD protein SMAD-4 , which is homologous to DAF-3 [72] . Therefore , DAF-3 and DAF-5 could also directly modulate the IIS pathway at the transcriptional level . A clear interpretation of our results is complicated by three main factors . First , the sheer number of insulins in the worm makes it difficult to assess whether they are functionally distinct . Secondly , the role of temperature in modulating the readouts of the pathway has not been closely explored . For example , we observe the effects of pdp-1 RNAi on daf-2 lifespan at 15°C but the effect decreases at a higher temperature , as the pathway gets more inactive . It is therefore likely that a certain level of signaling through the pathway is required to activate and target PDP-1 to its substrate ( s ) . At higher temperatures such as 20°C or 25°C , there may be extremely low levels of phosphorylated substrate available for PDP-1 . Similarly , the effect of a daf-3 null mutation on daf-2 phenotypes is more pronounced at higher temperatures but not at 15°C . Third , the lack of null alleles may provide an incomplete picture of the phenotypes observed . For example , previous studies using non-null alleles of daf-16 only partially suppressed dauer formation of TGF-β pathway mutants and therefore DAF-16 was thought to only affect the IIS pathway [31] . Therefore , temperature , level of signaling and the kind of mutants used ( null versus weak ) are important additional inputs that need to be considered to better understand the crosstalk between the IIS and the TGF-β pathways . In conclusion , our studies show that PDP-1 acts through the TGF-β pathway to negatively regulate IIS and promote DAF-16 activity . PDP-1 may mediate this function in part by negatively regulating TGF-β signaling to repress expression of several insulins that feed into the IIS pathway . In humans , dysregulation of TGF-β signaling and the insulin/IGF-1 signaling axis have been implicated in the onset of age-associated diseases such as Type 2 Diabetes and cancer [73]–[77] . Future studies exploring the interactions between these two pathways as well as the factors that modulate these interactions may ultimately provide a better understanding of the pathophysiology of these diseases .
All strains were maintained at 15°C using standard C . elegans techniques [78] . For all RNAi assays , worms were maintained on the RNAi bacteria for two generations except for the assays on the PDHc RNAi . Strains used in this manuscript are listed in Table S4 . RNAi plates were prepared as previously described [39] . All RNAi clones were sequenced and verified before any assays were carried out . L4 worms were picked onto fresh RNAi plates and maintained for two generations prior to the assay , with the exception PDHc RNAi plates . Worms exhibit lethality when maintained on the following RNAi clones: T05H10 . 6 ( E1α ) , C04C3 . 3 ( E1β ) , F23B12 . 5 ( E2 ) , or LLC1 . 3 ( E3 ) [79] . To circumvent this problem , strains were maintained on vector RNAi for two generations and transferred to E1α , E1β , E2 or E3 plates prior to the assay . For the pdp-1 ( tm3734 ) ;daf-2 ( e1370 ) double mutant , daf-2 ( e1370 ) males were mated to pdp-1 ( tm3734 ) hermaphrodites at 15°C . A total of 30 F1 progeny were picked onto individual plates and allowed to have progeny at 25°C . From the F2 progeny on each plate , dauers were selected and transferred to fresh plates and incubated for an additional 24 hours at 25°C . The next day , the dauers were allowed to recover at 15°C until they reached adulthood . Subsequently , adult worms were picked onto individual plates and transferred to 25°C and allowed to have progeny . Among the F3 progeny , we observed that some plates had 100% dauers at 25°C , while worms in some of the plates exhibited a developmental delay and could not form complete dauers even after 5–6 days at 25°C . Worms from both sets of plates were recovered , picked to individual plates and allowed to self at 15°C . Parents were then tested for pdp-1 ( tm3734 ) deletion by PCR . As anticipated , the pdp-1 ( tm3734 ) ;daf-2 ( e1370 ) double mutants are unable to form 100% dauers at 25°C . The daf-2 ( e1370 ) ;pdp-1::gfp strain was made by crossing daf-2 ( e1370 ) males to pdp-1::gfp hermaphrodites at 15°C . About 30 F1 animals were transferred to individual plates and allowed to have progeny at 25°C . From the progeny , F2 dauers were selected from each plate and allowed to recover at 15°C . The recovered adult worms were then checked for the presence of GFP , and GFP-positive worms were transferred to individual plates and incubated at 25°C . Plates where 100% of the progeny were dauers and GFP positive were selected and established as the strain for the assays . Strains were maintained on RNAi plates for two generations or regular OP50 plates at 15°C . Dauer assays were performed by picking approximately 100 eggs onto 2 fresh plates and incubated at the appropriate temperature . The pdk-1 ( sa680 ) , daf-7 ( e1372 ) and daf-14 ( m77 ) worms have a strong Egl phenotype . For dauer assays on these strains , gravid adult worms growing on the RNAi plates were washed off the plate with sterile PBS onto a 1 . 5 mL eppendorf tube . After 2 washes at 2000 g for 30 seconds , the adults were vortexed for 5 mins in 5 mL of 1 N sodium hydroxide and 3% sodium hypochlorite ( final concentration ) . The samples were then washed twice with sterile PBS and eggs were aspirated with a glass pipette onto fresh RNAi plates . For all dauer assays , plates were scored for the presence of dauers or non-dauers after 3 . 5–5 . 5 days , depending upon the strain . Dauer assays were performed at the temperature indicated . Significance was determined by Student's t-test . Strains were maintained at 15°C and synchronized by picking eggs onto fresh RNAi or OP50 plates . Approximately 60 young adult worms were transferred per plate to a total of three fresh RNAi or regular OP-50 plates containing 5-fluorodeoxyuridine ( FUDR ) at final concentration of 0 . 1 mg/mL [80] . All RNAi-based lifespan assays were carried out at 15°C . Lifespans on OP50 plates were performed at the temperature indicated . Survival was scored by tapping with a platinum wire every 2–3 days . Worms that died from vulval bursting were censored from the analysis . Statistical analyses for survival were conducted using the standard chi-squared-based log rank test . Strains were maintained on RNAi or regular OP50 bacteria at 15°C , as described above . From these plates , approximately 30 young adult worms were picked onto fresh RNAi or regular plates and upshifted to 20°C for 6 hrs . The plates were then transferred to 37°C and heat stress-induced mortality was determined every few hours till all the animals died . Statistical analyses for survival were conducted using the standard chi-squared-based log rank test . Strains maintained RNAi or on regular OP50 plates were synchronized by picking eggs on to fresh plates and grown synchronously at 15°C . The plates were then upshifted to 20°C for 8 hours , at the L2 stage to get L3 worms and at the L4 stage to get young adult worms . Worms were then washed off the plates into microcentrifuge tubes and incubated in 1× PBS buffer for 20 minutes on a shaker at RT . After 2 washes at 3000 rpm for 30 seconds with 1× PBS , the strains were fixed according to the type of staining performed . Oil Red O and Sudan black staining was performed as previously described [39] , [45] , [81] , [82] . After incubation overnight at RT , worms were mounted on slides and visualized using the Zeiss Axioscope 2+ microscope . For Sudan Black Staining , we used Image J software to measure the average pixel intensity for a 84-pixel radius below the pharynx of each animal in the anterior intestine area . Next , an 84-pixel radius of the background was measured , and subtracted from the values obtained for the staining . At least 10 animals were measured for each RNAi clone . Significance was determined by Student's t-test . For Oil Red O Staining , Image J was used to separate out each color image into its RGB channel components . As previously described [45] , Oil Red O absorbs light at 510 nm and therefore , the green channel was used for further analysis . We measured the average pixel intensity for a 84-pixel radius below the pharynx of each animal in the anterior pharynx area . We next measured a 84-pixel radius of the background , which was later subtracted from the values obtained from the staining . At least 10 animals was measured for each RNAi clone . Significance was determined by Student's t-test . DAF-16 localization assays were performed as previously described [39] , [52] . daf-2 ( e1370 ) ; daf-16::gfp worms were maintained on RNAi plates at 15°C similar to the dauer assays . Approximately 30 L4 worms were transferred to fresh RNAi bacteria and the plates were shifted to 20°C for 1 hr . The worms were visualized under a fluorescence microscope ( Zeiss Axioscope 2+ microscope ) . Worms were classified into four categories based on the extent of DAF-16::GFP nuclear-cytoplasmic distribution: completely cytosolic , more cytosolic than nuclear in most tissues , more nuclear than cytosolic in most tissues and completely nuclear . Quantification of Psod-3::gfp was performed as previously described [39] . daf-2 ( e1370 ) ;sod-3::gfp worms were grown at 15°C on RNAi as described above . Approximately 30 L4 animals were transferred to fresh RNAi bacteria and shifted to 25°C for 1 hr . The expression of sod-3::gfp was visualized using Zeiss Axioscope 2+ microscope . GFP expression was categorized as follows: High: GFP expression seen throughout the worm Medium: Weak expression detected in the body of the worm along with the head and the tail Low: Low GFP expression only detected in the head and tail Promoter and ORF entry clones of pdp-1 obtained from the promoterome and ORFeome were combined using multisite Gateway cloning ( Invitrogen ) into the pDEST-DD03 or the R4-R2 GFP destination vectors to create the Ppdp-1::gfp or Ppdp-1::pdp-1ORF::gfp constructs [83] , [84] . All constructs contain the unc-119 minigene . The vectors were verified by sequencing as well as restriction digestion . Transgenic worms were generated by ballistic transformation into unc-119 ( ed3 ) mutant worms as previously reported ( Biorad , USA ) [83] . Integrated lines that were obtained were used for further analyses . For the pdp-1::gfp translational fusion strain , additional lines were generated by integration of extrachromosomal array lines by UV irradiation as previously described [85] . All translational fusion lines were backcrossed 4× to wild-type prior to analysis . For all RT-PCR experiments , strains were maintained at 15°C . Eggs were obtained from gravid adult worms by hypochlorite treatment described earlier . The eggs were seeded onto large plates maintained at 15°C until the worms entered the L4 stage . The plates were then upshifted to 20°C for 8 hours until they became young adults . Worms were then collected with sterile 1×PBS and washed twice at 2000 g for 30 seconds . The supernatant was removed , and 0 . 5 mL of AE buffer ( 50 mM acetic acid , 10 mM EDTA ) , 0 . 1 mL of 10% SDS , and 0 . 5 mL of phenol was added to the worm pellet and the mixture was vortexed vigorously for 1 min , followed by incubation at 65°C for 4 min . Total RNA was purified by phenol:chloroform extraction and ethanol precipitation . The quality of the RNA isolated was determined by checking the 28 S and 18 S RNA on an agarose gel . 2 ug of total RNA was used for making cDNA using the SuperScript cDNA synthesis kit ( Invitrogen , USA ) . The expression of the DAF-16 target and insulin genes was checked by RT-PCR using the SYBR Green PCR Master Mix and 7000 Real-Time PCR System ( Applied Biosystems , USA ) . The relative expression of the genes tested was compared to actin as an internal loading control . Significance was determined by Student's t-test . Primers used for the RT-PCR experiments are listed in Table S5 . Young adult wild-type and pdp-1 ( tm3734 ) worms were picked onto 6 individual plates each . After 5 minutes , the worms were picked off the plate . The average distance covered was calculated by measuring the traces on the bacterial lawn using ImageJ . Significance was determined by Student's t-test . Wild type , daf-2 ( e1370 ) , pdp-1 ( tm3734 ) and pdp-1 ( tm3734 ) ; daf-2 ( e1370 ) worms were maintained at 15°C . 5 L4 worms were picked onto individual plates and allowed to lay eggs at 22 . 5°C . Worms were transferred to a new plate every 12 hours . After 22 . 5 hours , the parental worms were picked off the plates , and the total number of eggs laid was scored . The number of progeny from these eggs was scored again after 38 hours . The % hatched eggs was calculated as a percentage of the average number of progeny over the average number of eggs laid . Significance was determined by Student's t-test . Statistical analyses were performed using JMP and Microsoft Excel . NIH Image J was used for quantification of locomotion and fat storage . | Cells in the body respond to a variety of on/off signals that are relayed in a defined spatial and temporal manner . These signals influence several processes such as growth , fat storage , and the repair of damaged molecules . As humans age , the onset of diseases such as Type 2 Diabetes , obesity , and cancer often results from an imbalance in the levels of on/off signals in the cell . The insulin/IGF-1 signaling pathway is an important regulator of longevity , development , and metabolism across phylogeny . While the protein kinases that activate this pathway have been well studied , less is known about the protein phosphatases that tune down the signals . The roundworm C . elegans has been an excellent model system to study the role of insulin/IGF-1 signaling in the aging process . Here , we identify a new phosphatase that negatively regulates the insulin/IGF-1 pathway to enhance longevity and stress-resistance . Interestingly , the phosphatase achieves this function by tuning down the activity of a conserved TGF-β pathway , a pathway important for development . By reducing TGF-β pathway activity , this phosphatase decreases expression of insulin molecules that may stimulate the insulin/IGF-1 pathway . Our studies not only unravel a new regulator of these pathways , but also point to how they are more linked than previously thought . Both insulin/IGF-1 and TGF-β signaling have been implicated in age-associated diseases , and understanding their connection will provide us with potential therapeutic avenues . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/animal",
"genetics",
"genetics",
"and",
"genomics/gene",
"discovery",
"developmental",
"biology/aging",
"diabetes",
"and",
"endocrinology",
"molecular",
"biology",
"developmental",
"biology/developmental",
"molecular",
"mechanisms"
] | 2011 | PDP-1 Links the TGF-β and IIS Pathways to Regulate Longevity, Development, and Metabolism |
For decades La Réunion has experienced a number of epidemics that have resulted in efforts to control the density of Aedes species on this Island . This study was conducted to assess household-level expenditure on protective measures against mosquito nuisance on the Island of La Réunion in 2012 . Data was collected during a cross-sectional survey of 1024 households and used to determine the relationship between the use of chemically-based protective measures and subjective and objective indicators of the density of Aedes albopictus . The average household expenditure in July 2012 was USD 9 . 86 and the total household-level expenditure over a one-year period was extrapolated to USD 28 . 05million ( range: USD 25 . 58 million to USD 30 . 76 million ) . Much of this money was spent on measures thought to be relatively ineffective against Aedes mosquitoes . Expenditure on protective measures was not influenced by the level of knowledge on mosquitoes or by the visual nuisance they generated at home , but rather by the perception of risk related to a future epidemic of chikungunya and socioeconomic factors . Most importantly , household spending on protective measures was found to be influenced by a measure of zone-level mosquito density ( the Breteau index ) , but not by objective indicators of the presence of mosquitoes within or around the house . Household-level expenditure on chemically-based protective measures is high when compared to the investment made by public entities to achieve vector control , and it is differentially influenced by subjective and objective measures of mosquito density . The current situation could be improved , firstly by ensuring that the public is well-informed about mosquitoes and the effectiveness of various protective measures , and secondly by implementing interventions that could either complement current vector-control strategies and improve their effectiveness on a country-level , or that would steer the population toward the appropriate behaviours .
Aedes albopictus , commonly known as the Asian tiger mosquito , is an anthropophilic , daytime biting species that rapidly establishes itself in new urban areas owing to its propensity to breed in both artificial and natural containers of stagnant water [1] . The tiger mosquito is particularly threatening owing to its potential for transmitting a wide range of arboviruses , including dengue and chikungunya viruses , yellow fever virus , and several other types of encephalitides [2] , [3] , [4] , [5] , [6] . La Réunion is one of the places in the world that has experienced a number of epidemics as a result of the favourable conditions it provides for mosquito species to thrive . Past outbreaks of malaria and dengue prompted authorities on the Island to implement strategies to control mosquito density . Following the resurgence of dengue in 2004 , the local vector control services , referred to as the ‘Services de Lutte Antivectorielle’ , or LAV , started developing a control strategy targeted at urban vectors , primarily Aedes albopictus [7] , [8] , [9] . The major chikungunya outbreak that swept through La Réunion in 2005–2006 created even stronger motivation for authorities to set up entomologic surveillance of Aedes albopictus in all urban areas . This surveillance effort continues today through monitoring of traditional stegomyia indices of immature stages ( i . e . Container Index , House Index , Breteau Index ) as are used in other control programmes [10] . The house index is defined as the percentage of houses infested by larvae and/or pupae . The container index is defined as the percentage of water-holding containers with active immature stages of mosquitoes . The Breteau index is defined as the number of positive containers per 100 houses , a positive container being one that contains larval and/or pupal stages of mosquito . The Breteau index is being used as a measure of zone-level vector density in this analysis ( see method section ) . Aedes albopictus remains the main target of the work of the LAV , a service which is provided by the Regional Health Agency ( Agence Régionale de la Santé , or ARS ) in La Réunion . The vector-control strategy integrates five core activities: vector surveillance , environmental , mechanical , and chemical control , and public health education campaigns [11] . Another aspect of the work of the vector control services is the early detection and treatment of cases of arboviral infection to prevent the spread of new epidemics . For the most part , the day-to-day activities of LAV officers involve education and promotion of vector-control at the household-level . These officers routinely visit households in La Réunion and provide education to families on the importance of eliminating sources of stagnant water around the house , such as emptying water from pots and saucerss placed under potted plants . Given the investment of both financial and human resources toward the control of Aedes albopictus in La Réunion , a study was proposed to assess the population's perceptions and behaviour related to mosquito nuisance , and identify whether the current strategy could be improved or enhanced through new vector-control measures or interventions . For this study , insight into household-level behaviour was gained using estimations of expenditure on protective measures against mosquitoes . The objective was to determine whether spending at the household-level is influenced by subjective or objective exposure to Aedes mosquitoes on the Island of La Réunion , and whether this level of expenditure warrants action by public authorities to improve current vector-control strategies .
This study was performed in urban areas of La Réunion . The Island is divided into 4 geographic sectors ( North/South/East/West ) , 24 municipalities and 273 neighbourhoods . These neighbourhoods are divided into 960 zones used by the ARS for Aedes albopictus surveillance and control . These homogenous zones are defined according to urban planning and environmental criteria , they extend over 275 km2 or 11% of the Island , and mostly cover urban areas . According to the National Institute of Statistics and Economic Studies ( or INSEE ) in France , an urban area is defined as an agglomeration of more than 2000 residents where no dwelling is separated from the next closest dwelling by >200 metres [12] . Home-owners generally allow LAV officers of the ARS to enter into their private dwellings to carry out routine vector-control activities . We commenced our sampling technique by selecting all zones controlled by the ARS on a minimum of three occasions between 2007 and 2011 . We used zones located near the coast ( less than 500 metres in altitude ) where the presence of mosquitoes from one year to the next is most likely to persist and to ensure a relative homogeneity in environmental factors . Next , we focused on zones that showed relative stability in mosquito density between 2007 and 2011and we classified these as either negative or positive zones using criteria based on the Breteau Index ( i . e . the number of positive containers per 100 houses ) . Using the historical data gathered over this 5-year period , we defined a negative zone as one that had a value for the Breteau Index at LAV routine visits that was lower than 50% of the average Breteau Index value during the same month and year in all zones . A positive zone was defined as having a value for the Breteau Index during LAV routine visits that was higher than 50% of the average Breteau Index value during the same month and year in all zones . A zone was kept for inclusion in the study if it was classified as positive or negative a majority of times during all LAV routine visits ( e . g . classified as ‘positive’ twice during three routine visits conducted between 2007 and 2011 ) . This was done to ensure a degree of stability in the classification of zones through time . A total of 184 zones , 68 positive zones and 116 negative zones , were identified using this methodology . Being that face-to-face interviews in 184 zones were not feasible , a two-stage cluster random sample was drawn from this first selection . In the first stage of this two-stage sampling technique , a random sample of 26 zones ( 13 positive and 13 negative ) was drawn taking into account the geographic distribution of the population on the Island in the four sectors ( North/South/East/West ) . Next , households were randomly selected to achieve a fixed sample of 40 households per zone . The selection of households was undertaken while LAV officers were in the field . All households were randomly chosen and surveys were conducted in locations routinely checked by vector control officers . For each zone , the officers were asked to interview residents living in alternate households . The households were selected in this way while walking through the zone . The LAV officers commenced the survey starting at the four corners of the zone and walked in varying directions that were also chosen at random . The percentage of absentees and refusals varied considerably between zones; the average percentage of absentees was 36% ( range of 5% to 50% ) , and the average percentage of refusals was 12% ( range of 4% to 24% ) . The reason most often quoted for refusing to participate in the survey was a lack of time to respond to the questionnaire . For each selected household , the LAV officers conducted both a face-to-face interview that was addressed to the head of the household as well as an observational survey of the outside of the dwelling itself . The data collected were validated by comparing key characteristics of our sample to information provided in the latest census ( e . g . number of household members by age of the head of household , level of education of the head of household , socioeconomic status ) . No significant differences were found for these key characteristics , an indicator that our sample was representative of the population in La Réunion . Conversely , a significant difference for the Breteau Index measured during the month of interviews ( i . e . July 2012 ) was found between the positive and negative zones in the sample ( Mann-Whitney test p-value<0 . 001 ) . This finding confirms that the survey provides a rather accurate picture of the long-term average density of Aedes albopictus in the selected zones ( i . e . our sampling technique resulted in the selection of zones that retained their characteristic classification of vector density through time ) . Questions were derived from existing literature on protective behaviours against mosquitoes [13] , [14] , [15] , [16] , [17] , [18] . Due to the circulation of dengue virus at the time of the study and an epidemic alert level up to 2B issued by the ARS during the month of April 2012 , questions on the risk and perception of dengue were also integrated into the questionnaire . A pilot study was launched in June 2012 to test the validity of the questionnaire . The main study was carried out in July 2012 , and was conducted according to the rules established by the National Data Protection Authority . Informed consent to answer the 40-minute questionnaire and to allow a LAV officer to conduct an observational survey of the residence was obtained verbally from all participants at the beginning of the interview . Translators were used when necessary . Analysis was performed using STATA/SE v11 . 0 ( StatCorp , College Station , TX ) . Descriptive information on behaviour , knowledge , and perceptions related to mosquitoes in our study sample was extracted using univariate and bivariate analyses . However , extrapolating from a simple average expenditure per household , as well as the relationship between expenditure per household and Breteau index , could be confounded by other household- or zone-level cofactors ( presence of mosquitoes at home , wealth index , education amongst others ) , resulting in biased estimates at the Island level . In light of this we decided to analyse the influence of household- and zone–level characteristics on expenditure on protective measures ( i . e . direct expenditure at the household level ) using multivariate regression analysis . Simple generalised linear regression analysis was first performed using household-level expenditure on protective measures as the outcome variable . All factors significantly related to household-level expenditure were then entered into the final multiple regression model , using a random effects model . Provided that random effects are uncorrelated with the fixed predictors in the model , a random effects model is preferable , as it allows for the consideration of both household- and zone-level characteristics in a single model [24] . As the number of zones is relatively low ( 26 ) , we didn't perform a mixed multilevel analysis , following Scherbaum et al recommendations ( in which case a minimum of 30 level 2 zones is recommended to perform multilevel analysis ) [25] . Lastly , univariate sensitivity analysis was conducted to determine how sensitive our estimates of household expenditure were to variation in input parameters , including perceived risk of a new epidemic of chikungunya , age , wealth , and education . Analysis was performed on wealth quintiles distribution across the population ( from 100% in poorest category to 100% in richest category ) , educational levels ( from 100% in the primary education category to 100% in college and higher category ) , age ( 20 to 99 years old ) , and perceived risk of a new epidemic of chikungunya ( from all people considering that there is a low risk to 100% considering that there is a high risk ) . We also tested the influence of excluding air-conditioning and fans from the expenditure calculation . Results are displayed graphically using a tornado diagram .
A total of 1024 households in La Réunion were interviewed over the month of July 2012 . Figure 2 provides the location of zones interviewed on the Island and Table 2 provides descriptive statistics of the main factors included in the multivariate analysis . Only about 40% percent of respondents could identify the female as the biting gender in mosquito species . When asked to identify diseases transmitted by mosquitoes , the mean score for respondents in the study sample was of 3 . 73 correct answers out of 5 . Not surprisingly , 92% of interviewees knew that chikungunya is a disease transmitted by mosquitoes , and another 78% answered correctly for malaria . More than a quarter of the respondents answered incorrectly that influenza is a disease transmitted by mosquitoes . Most respondents declared that mosquitoes were present in their homes ( 77% ) . However , when asked about the frequency of mosquito bites over the last 7 days , 93% claimed that they were seldom or never bitten by mosquitoes . In spite of the low level of mosquito bites , 90% of interviewees considered mosquitoes a nuisance and 63% stated that these insects were of no particular use . When asked why mosquitoes were a nuisance , more than 80% replied that this was due to their role in transmitting diseases , 80% stated that mosquito bites and itching were important reasons , and 65% stated that it was due to the noise they created . When questioned about the risk of epidemics , 47 . 7% of interviewees perceived that the risk of a new outbreak of chikungunya was reasonable or high . On the other hand , the risk of a dengue epidemic was perceived as being reasonable or high by 52% of respondents . This higher perceived risk of dengue could be explained by the increased attention given to this disease during the time of the study when a small epidemic of dengue was unfolding across the Island . Interestingly , however , when an open question was posed about which diseases could be transmitted by mosquitoes , only 20% of respondents spontaneously quoted “dengue” . With respect to overall perception of vector control efforts , approximately 70% of respondents were confident that it is possible to reduce the number of mosquitoes . However , 20% insisted that nothing could be done in this regard . More than 75% of interviewees think that science could make further advances in the field of vector control . In terms of the acceptability of different measures to control mosquito numbers , 97% find the elimination of stagnant water acceptable , 82% feel that measures to repel mosquitoes are acceptable , and 74% accept techniques that prevent the reproduction of mosquitoes . Insecticide spraying is deemed acceptable by 65% of the study sample . When asked whether they eliminate sources of stagnant water in and around their households , 97% of respondents declared that they did this . The frequency of this behaviour varied , however , with 17% claiming that they would eliminate these sources at least once per day , 45% declaring that they did this a few times per week , 11% stating that they did it a few times per month , and about 2% stating once per year . About 23% of the study sample claimed that they had definitively eliminated potential sources of stagnant water by removing empty containers or other potential recipients from their surroundings . Among the various measures listed in the questionnaire , mosquito coils emerged as the most commonly used protective measure in this study , with 69% quoting that they used this measure at the time of the study . Insecticide/mosquito repellent sprays for the house ( 53% ) , non-electric diffusers ( 38% ) , and repellent creams and sprays applied to the skin ( 36% ) were less frequently used by the sample during this period . Although use of these measures persists , about 50% of respondents stated that they consider them to be either reasonably or very dangerous to one's health , and another 24% felt that these products were not really effective . According to the health recommendations for travellers [20] , mosquito coils are amongst the least effective measures to repel Aedes mosquitoes , intra-domiciliary insecticides and electric diffusers were found to have limited and weak effectiveness , and repellent creams and sprays applied to the skin were judged to be of stronger effectiveness . Therefore , the utilisation pattern of protective measures in this study population appears to be directly inverse to the recommendations of the health authorities [20] in terms of product effectiveness . The lists provided by respondents of the measures they use to protect themselves against mosquitoes were used to determine the average household-level expenditure on these measures during the study period . This average expenditure per household was estimated to be USD 18 . 09 during the month of July 2012 and the median expenditure was estimated at USD 15 . 54 . We tested the robustness of these household-level expenditure estimates by verifying whether these correspond with the results of a direct and closed question on household expenditure in the questionnaire . Respondents were asked to specify whether they judged their monthly spend on protective measures against mosquitoes to be less than EUR 10 ( USD 12 . 24 ) , between EUR 10 ( USD 12 . 24 ) and EUR 20 ( USD 24 . 48 ) , EUR 21 ( USD 25 . 70 ) to EUR 40 ( USD 48 . 96 ) , or more than EUR 40 ( USD 48 . 96 ) . Just over 60% of respondents declared spending between USD 12 . 24 and USD 24 . 48 per month on protective measures against mosquitoes . Taking the middle point of the range of each category , we found an average expenditure of USD 13 . 60 , when using the results of this direct and closed question . A direct declaration of monthly expenditure ( i . e . a categorical variable ) is judged to be less reliable and accurate compared to the estimation of expenditure based on a list of used products ( continuous variable ) which is why the latter , i . e . the estimates of household-level expenditure , have been used as the dependent variable in the regression analysis . Table 3 summarises the findings of the multivariate analysis that shows household-level expenditure on protective measures against mosquitoes as the dependent variable . Wealth quintiles , age ( p = 0 . 039 ) , and an educational level above a college degree ( p = 0 . 087 ) , were found to be positively and significantly associated ( at the 10% level ) with household-level expenditure on protective measures . Gender of the head of the household was not found to influence expenditure . Knowledge on diseases transmitted by mosquitoes is shown to positively affect household expenditure on protective measures but the coefficient of this variable just misses the statistically significant threshold of 10% ( p = 0 . 141 ) . Correct knowledge of the biting gender of the mosquito species or the distance travelled by mosquitoes had no significant impact on expenditure on protective measures . In terms of perceived risk of a potential epidemic , the perceived threat of another chikungunya outbreak was found to be significantly associated with expenditure on protective measures . The potential threat of a dengue epidemic , however , did not have the same effect . Perception of the effectiveness of protective measures was found to influence household expenditure on these items only for those who stated that these measures are ‘effective’ ( p = 0 . 021 ) . The regression analysis shows that while the declaration of having mosquitoes in one's household does not appear to influence expenditure on protective measures , an estimation of the frequency of mosquito bites given by respondents , on the other hand , has a significant and positive relationship with expenditure on products . There is a clear relationship between the perceived biting frequency ( a measure of subjective exposure to mosquitoes ) and the degree of spending on protective measures ( the response to this exposure ) . In terms of objective evidence of mosquito density , the number of positive Aedes breeding sites measured in and around the household does not appear to influence expenditure , and it is in fact the vector density for the entire zone ( measured using the Breteau index ) , that has a positive impact on household spending on protective measures ( p = 0 . 004 ) . Assuming that the positive relationship between the Breteau Index and household spending exists for all seasons of the year , it was possible to extrapolate the results of this study to make household expenditure estimations for the months between July 2011 and July 2012 . A prerequisite for this extrapolation was data on the seasonal variation of the Breteau Index over this one-year time-frame , which were obtained using ARS records of monthly-measured Breteau index ( publicly available on a monthly basis on the ARS website [26] ) . When asked about mosquito nuisance 72% of participants claimed to be affected by this mostly during the Austral summer ( November to April ) , and another 42% declared that their use of protective measures would increase during the year , a plausible finding being that this study was conducted during the month of July . The indices recorded by the ARS were compared with the results from our regression analysis ( provided in Table 3 ) . We used predictions of the model at fixed values of ARS-measured Breteau index and averaged over the remaining significant covariates to estimate an annual global expenditure for households on the Island . Results for both the average expenditure per household and for all households in Réunion are given in Figure 3 and Table 4 . In this study , the definition of a household has been limited to persons living within houses . The total number of households in La Réunion in 2009 was 284 , 391 , as measured by the National Institute of Statistics and Economic Studies in France ( INSEE ) . Among them , INSEE estimates that around 71% reside in houses . This corresponds to a total of 201 , 917 households ( as defined in this study ) on the Island of La Réunion . The analysis shows that the predicted average household expenditure for July 2012 is USD 9 . 86 , which is a better estimate of household expenditure than that derived from the list of used products as explained in the descriptive statistics section . This figure is also more accurate because it takes into account confounders included in the regression analysis . The projected total annual expenditure per household was estimated to be USD 138 . 92 ( when using the survey definition of a household ) . If we include all households on the Island and assume that residents living in dwellings other than houses have no mosquito-related expenditure ( i . e . taking a conservative approach ) , the total annual expenditure per household is USD 98 . 63 ( the lower bound estimate of the total annual expenditure per household ) . Overall , the amount spent by 201 , 917 households in Réunion from July 2011 to July 2012 on personal protective measures is estimated to be USD 28 . 05 million ( 95% CI of USD 25 . 58 million to USD 30 . 76 million ) . By comparison , other authors have found that the chikungunya epidemic resulted in medical expenses of up to EUR 43 . 9 million ( USD 53 . 74 million ) [27] . The annual Gross Domestic Product ( GDP ) of La Réunion was estimated to be EUR 14 . 42 billion ( USD 17 . 65 billion ) in 2009 . Hence , the household-level expenditure related to protective measures in La Réunion amounts to 0 . 15% of its annual GDP in 2009 . The ARS already dedicates about EUR 10 million ( USD 12 . 24 million ) annually to the vector-control service or LAV , which is only a part of the overall budget being spent by municipalities , associations , and other actors to ensure vector control on the Island through various activities . The estimated population of the La Réunion Island in 2009 was about 816 , 360 inhabitants . ARS expenditure per person on vector control for 2009 was thus about USD 14 . 99/inhabitant ( ARS expenditure does not represent total public expenditure on mosquito prevention and control – this is a value that we cannot capture , due to the role played by a number of different entities/authorities that in some way affect mosquito density on the Island ) . The annual GDP ( Gross Domestic Product ) per capita of La Réunion was EUR 17 , 884/inhabitant ( USD 21 , 890/inhabitant ) in 2009 . When compared to the expenditure of the vector-control service of the ARS , the total annual expenditure on protective measures for all households in La Réunion ( i . e . USD 28 . 05 million ) appears to be disproportionately high , a finding that has been pointed out by another study [28] . In order to test the sensitivity of our expenditure estimate to variation in the main factors influencing household-level expenditure , a univariate sensitivity analysis was carried out , as stated in the methods section . The factors influencing expenditure included perceived risk of a new epidemic of chikungunya , age , wealth , and education . We also tested the influence of excluding air-conditioning and fans from the expenditure calculation . The results are displayed graphically using a Tornado diagram in Figure 4 . As guidance to interpret the findings in the diagram , for instance , age structure of the population has a large impact on estimated expenditure whereas variation in perceived risk of a new epidemic of chikungunya is unlikely to result in savings at the Island scale . The fact that the results are not strongly affected by the exclusion of air-conditioning and fans can be explained by the fact that the prices of these products were discounted in the main analysis .
The general limitations of this study are related to the fact that our expenditure estimates could be underestimated or overestimated . Being that the study was carried out in July which is a less favourable month for mosquitoes , it is probable that our household expenditure is under-estimated ( conservative bias ) . Moreover , we have excluded from the study residents living in high-rise buildings , apartments and flats . This is an additional source of underestimation for our expenditure estimate . However , it is also possible that the expenditure estimates have been overestimated as participants may have reported using more protective measures than they actually do in reality and because the measures they use target mosquito nuisance in general and not only that created by Aedes albopictus , the focus of our study . In La Réunion the name “Tiger mosquito” is not commonly used and it is probable that the majority of people cannot distinguish between different species of mosquitoes . This is a fair assumption considering the low level of general knowledge on mosquitoes we found in this study sample . This assumption should be tested in future studies in La Réunion by specifically asking participants whether they can correctly identify Aedes albopictus , which has been found to be the case in other parts of the world , including the South of France [38] . Our inclusion criteria allowed us to include zones located in specific geographical settings ( i . e . located near the coast ( less than 500 metres in altitude ) where the presence of mosquitoes from one year to the next is most likely to remain stable . This is another source for potential overestimation of household expenditure . Using these inclusion criteria could have resulted in a slight overestimation of the average Breteau index in our sample compared to the Island average . Indeed , in terms of risk indicators , the average Breteau index in our sample is 42 . 28 [37 . 92–46 . 63] . The estimate provided by the ARS in 2012 for the whole Island is 30 . 82 [27 . 19–34 . 44] and 38 [34 . 37–41 . 62] in 2011 . Nevertheless , this slight overestimation of the Breteau Index should not have much influence on the estimates for household-level expenditure obtained from the regression analyses . This is because we have a variety of situations represented through the positive and negative zones used in this study . Our household-level expenditure estimations may also be limited by the fact that we have no longitudinal survey to confirm usage patterns of products across time . In addition to this , using model predictions with average values of covariates can lead to over-estimation of expenditure when this is made on the scale of an Island . In this article , we were particularly interested in expenditure on chemically-based protective measures or repellents ( including insecticide-treated nets ) . We did not focus on ecological interventions ( such as eliminating mosquito breeding sites and stagnant water ) or other protective behaviours ( such as limiting outdoor activities ) , information which would be valuable when measuring opportunity costs due to time spent on these activities or the impact of mosquito nuisance on quality of life . This could be the subject of another study in the future . Another limitation is that this cross-sectional study cannot fully explain why people use some chemical measures that are judged to be ineffective by French public authorities . It is probable that habits or tradition continue to play a role in the use of certain measures to repel or kill mosquitoes in La Réunion , for example the use of fire . Identifying the reasons for these continuing behaviours and the use of measures that may actually prove to be ineffective requires a qualitative study approach that would capture more information than is possible through use of standard questionnaires and quantitative methods . Lastly , a key lesson from this cross-sectional survey is that longer-term research should be undertaken in order to take into account seasonal variations in protective behaviours against mosquito nuisance and disease threats in order to provide more robust conclusions . Differences in mosquito control practices at the local level involve the interplay of place , scale and politics [39] . This study is one of the first attempts to quantify household-level expenditure on protective measures against mosquitoes , a very important step considering that community involvement is considered to be at the heart of vector-control strategies in La Réunion and elsewhere . More importantly , longer-term studies on this subject , as well as studies on the effectiveness of different products , can be instrumental in determining potential savings at the household-level due to improvements in public messages and the introduction of new policies or interventions that are currently considered as being too expensive . Finally , it is evident that household-level behaviour is differentially affected by subjective and objective measures of exposure to Aedes albopictus . Both variables need to be taken into account when explaining the use of chemically-based protective measures against mosquitoes and any related variations in expenditure . | The French Ministry of Health has , for decades , dedicated numerous resources to control mosquito density on the Island of La Réunion . These efforts were strengthened following an outbreak of chikungunya , a virus transmitted by Aedes mosquitoes , in 2005–2006 . In order to understand how public perception and behaviour is affected by this vector , a study was undertaken in 2012 . Public behaviour was assessed using estimates of household expenditure on protective measures against mosquitoes . Information was gathered using a survey administered to 1024 households on the Island . Knowledge about mosquitoes was found to be poor across the sample , while perceptions of a risk from epidemics were high . The threat of a chikungunya epidemic was found to be associated with increased expenditure on protective measures , as was a zone-level measure of mosquito density , the Breateau Index . The most important finding is that overall household expenditure due to mosquitoes over a one-year period is USD 28 . 05 million , rather high when compared to the public service investment . Future vector-control in La Réunion needs to ensure that public health messages are understood by the population and that interventions are implemented that promote appropriate behaviours and reduce current spending at the household-level on protective measures . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"social",
"and",
"behavioral",
"sciences",
"economics"
] | 2014 | Household-Level Expenditure on Protective Measures Against Mosquitoes on the Island of La Réunion, France |
For most of the world , human genome structure at a population level is shaped by interplay between ancient geographic isolation and more recent demographic shifts , factors that are captured by the concepts of biogeographic ancestry and admixture , respectively . The ancestry of non-admixed individuals can often be traced to a specific population in a precise region , but current approaches for studying admixed individuals generally yield coarse information in which genome ancestry proportions are identified according to continent of origin . Here we introduce a new analytic strategy for this problem that allows fine-grained characterization of admixed individuals with respect to both geographic and genomic coordinates . Ancestry segments from different continents , identified with a probabilistic model , are used to construct and study “virtual genomes” of admixed individuals . We apply this approach to a cohort of 492 parent–offspring trios from Mexico City . The relative contributions from the three continental-level ancestral populations—Africa , Europe , and America—vary substantially between individuals , and the distribution of haplotype block length suggests an admixing time of 10–15 generations . The European and Indigenous American virtual genomes of each Mexican individual can be traced to precise regions within each continent , and they reveal a gradient of Amerindian ancestry between indigenous people of southwestern Mexico and Mayans of the Yucatan Peninsula . This contrasts sharply with the African roots of African Americans , which have been characterized by a uniform mixing of multiple West African populations . We also use the virtual European and Indigenous American genomes to search for the signatures of selection in the ancestral populations , and we identify previously known targets of selection in other populations , as well as new candidate loci . The ability to infer precise ancestral components of admixed genomes will facilitate studies of disease-related phenotypes and will allow new insight into the adaptive and demographic history of indigenous people .
During the past decade , data generated by high-throughput genotyping technologies have enabled studies probing into two central questions in human evolutionary biology: the characterization of human population genetic structure , and the search for the molecular signature of natural selection . Insights gleaned from these studies have provided important clues for understanding the phenotypic diversity of our species , and variables representing population structure are routinely incorporated as covariates in genome-wide association studies of complex traits and diseases . At a global level , as well as within a continent or even a sub-continental region , geography has been shown to act as the leading driving force in shaping the pattern of genetic variation that we observe today [1]–[5] . In parallel , analyses based on European , African and East Asian populations have revealed that recent positive selection is a prevalent phenomenon throughout the genome [6]–[8] . Using data from the Human Genome Diversity-CEPH Panel ( HGDP ) , a recent and comprehensive survey suggests that , while adaptation to local environment is a common theme throughout human evolution , the genetic loci involved in adaptation show little overlap among non-contiguous geographic regions [9] . While geography poses a significant reproductive barrier , multiple waves of massive trans-continental migration have occurred during the past centuries , giving rise to admixed populations . The ancestry of non-admixed individuals can often be traced to precise regions based solely on genetic data , but characterizing the sub-continental ancestry origins of an admixed individual has not been demonstrated to date . For example , the two largest minority groups in North America , Latinos and African Americans , both arose as a result of mating among populations that had been in historical reproductive isolation . The “Hispanic” or “Latino” populations include the ethnically diverse groups of Latin America; although significant genetic contributions can be traced to Indigenous American , European and West African populations , it has been challenging to determine whether one's Indigenous American ancestors originate from North , Central , or South America . Solving this problem has implications for both a deeper understanding of human evolution and for human disease , since genetic diversity between Latino populations is characterized both by variation in continent-level ancestry – e . g . Mexicans on average have lower African ancestry than Puerto Ricans – and by the population structure among the ancestral Indigenous American populations [4] , [10] . The assessment of the precise ancestral origin and the quantification of genetic structure within an ancestry component are limited , in part , by analytic challenges . Principal Component Analysis ( PCA ) is a classic technique for multivariate data analysis , which aims to project high-dimensional data to a much lower dimension while capturing the greatest level of variation [11] . This approach has gained popularity in genetic analyses due to both computational efficiency and interpretability: when the underlying population structure is driven mainly by reproductive isolation and subsequent genetic differentiation , the principal components ( PCs ) mirror the geographic origins of individuals [3] . By itself , however , PCA is not well suited for studying admixed populations: while leading PCs usually represent the relative contributions of continentally-divided ancestral populations , subsequent PCs may be simultaneously influenced by structures within one or more of the ancestral populations , and are consequently difficult to interpret . We tackled this problem by employing an analytic strategy that works backwards according to the temporal nature of demographic events that underlie human admixture: genomes are first separated into the major and most recent components that reflect inter-continental migration , then each of those components is further investigated separately . As described below , we apply a probabilistic method for inferring locus-specific ancestry along the chromosome , followed by a variant of PCA to further investigate each of the ancestry-specific genomic components , which we term “virtual genomes” . This hierarchical strategy yields a fine-scale view of genetic structure in admixed populations , and provides insight into the population history of nonextant ancestral populations . As an example , we study a cohort of 492 parent-offspring trios recruited from Mexico City . Our results confirm the a priori expectation that the most significant European contributors to the Mexican gene pool are populations from the Iberian Peninsula , but reveal that the Indigenous American component of the Mexican genomes is more complex . Studying the genetic structure of admixed genomes also offers the unique opportunity to probe the adaptive landscape of the ancestral populations . This is particularly powerful for studying the Indigenous American populations , for which limited genotype data is available . As proof of principle , we report a novel application of the extended haplotype homozygosity test for recent positive selection to the European and Indigenous American “virtual genomes” evident in the Mexican cohort , and identify numerous loci as potential targets of positive selection .
Our analytic strategy for studying population structure in admixed populations is shown in Figure 1; details of the approach are described in what follows , and in the Materials and Methods section . This approach first applies a model-based clustering method , frappe , to the intact genotype matrix , identifying components that correspond to variation in continental-level admixture proportions , and estimating the relative proportion of those components for each individual . Locus-specific continental ancestry along a genome is then inferred using SABER+ , an extension of a Markov-Hidden Markov Model method [12] that partitions each genome into ancestral haplotype segments or “virtual genomes” . Finally , within-continent population structure is determined by applying PCA to the virtual genomes , treating the rest of the genome as missing . To account for the large amount of the missing data resulted from the continent-specific genomes , we implement a variation of the subspace PCA ( ssPCA ) algorithm [13] . Most of the results described here are from a panel of 492 Mexican parent-offspring trios recruited from Mexico City ( MEX1 ) as part of a previous genome-wide association study using genotype data from the Illumina 550K platform [14] . For comparison , we also examined data from 23 HapMap Phase3 Mexican trios recruited from Los Angeles , California ( MEX2; http://hapmap . org ) . Reference populations for inferring continental-level ancestry were taken from HapMap ( CEU , YRI ) , and additional sources as described below and in Table S1 . Among the 984 parents of the Mexico City trios ( MEX1 ) , we used frappe to estimate median ancestry proportions of 65% Indigenous American , 31% European , and 3% African; the corresponding statistics in the 46 HapMap Mexican individuals from Los Angeles ( MEX2 ) are 45% , 49% , and 5% , respectively ( Figure 2A ) . The distribution of Indigenous American ancestry in the Mexico City population is shifted upward compared to the Los Angeles population ( Figure 2B ) , which may reflect differences in the extent of European admixture . African ancestry is low in both cohorts , although the distribution is skewed to the right , reaching over 40% for some individuals . We next used SABER+ to estimate recombination breakpoints between ancestral chromosomes and thus locus-specific ancestral origin— Indigenous American , European , or African—in individuals from the MEX1 and MEX2 cohorts . For the work described here , the primary goal of SABER+ is to partition the Mexican genomes into haplotype segments according to continental ancestry that can be used for subsequent analysis . However , the output of SABER+ can also be used as an independent means of assessing global ancestry , simply by averaging locus-specific ancestries across all markers , and yields estimates that are highly correlated ( r>0 . 99 ) with frappe ( Figure S1 ) . To facilitate the analyses of sub-continental genetic structure , we constructed virtual genomes by retaining haplotype segments from a single continental-ancestral population , while masking ( i . e . setting to missing ) segments from all other ancestral populations; for example , MEX1AMR and MEX1EUR denote the sets of Indigenous American and European haplotype segments from the Mexico City individuals , respectively . In a principal component analysis of this data that includes the YRI and CEU HapMap populations , the Indigenous American , European , and African virtual genomes mark vertices of a triangle ( Figure 2C ) in which the intact genomes of the MEX1 and MEX2 individuals are distributed broadly along an Indigenous American – European axis represented by PC1 . The exact position of the intact MEX1 and MEX2 genomes depends on admixture proportions; individuals with the greatest level of African ancestry , which corresponds to PC2 , mostly lie at intermediate positions along the Indigenous American -European axis . Importantly , the MEXEUR and MEXAFR virtual genomes ( red and blue points , respectively ) form discrete clusters whose locations coincide with those of the HapMap CEU and YRI , respectively , and , while there is no reference population in this analysis for Indigenous American , the MEXAMR virtual genomes also form a discrete cluster at a vertex of the triangle . These observations suggest that the ability of SABER+ to assign local ancestry to a specific continental origin is highly accurate , which is essential for subsequent analyses . The distribution of the length of ancestry blocks is shaped by population history since admixture . When two individuals from different parental populations mate , the first generation offspring inherits exactly one chromosome from each parental population . In subsequent generations , recombination events in an admixed individual generate mosaic chromosomes of smaller ancestry segments . Intuitively , more recent admixing gives rise to longer ancestry blocks than older admixture . Furthermore , conditioning on the time since admixing within an individual's pedigree , block length distribution also depends on the individual level ancestry proportions: e . g . , an individual with 90% European ancestry tends to have long European ancestral blocks because recombination events in the person's genealogy are likely to have joined two European haplotypes , and therefore fewer ancestry changes are expected . A likelihood-based model has been proposed that can estimate several aspects of admixture history [15] . However , the admixing rates in Mexicans from the European , Indigenous American and African ancestral populations are likely dependent and difficult to model with this likelihood-based method; therefore , we attempted to estimate admixing time using a different approach . We first computed the theoretical number of ancestry blocks for individuals according to their ancestral proportions , and carried out that computation assuming a series of different admixing times ( 5–25 generations , dotted lines in Figure 2D ) . The parabolic shape of these curves conforms to the intuitive idea outlined above that the number of block peaks at an intermediate ancestry proportion . We then superimposed the observed number of ancestry blocks in each MEX1 individual onto the theoretical curves; these results suggest an admixing time of 10–15 generations ago ( Figure 2D ) . The admixing time of the European component appears slightly longer than that for the Indigenous American component ( 15 generations vs . 12 ) ; one potential explanation is that some mixing occurred between the European and the African ancestral individuals prior to admixing with the Indigenous American populations . With the MEXEUR uncoupled from the MEXAMR genomes , we investigated structure within each of these virtual genomes separately . ( We did not investigate the MEXAFR virtual genomes due to their small sample size ) . Because there is a large amount of missing data , e . g . the virtual genome of one individual may cover very different loci from the virtual genome of other individuals , we used the ssPCA approach as described in Materials and Methods . To help evaluate the robustness of our approach , we carried out simulation experiments , in which the effects of random error in the inference of continental locus-specific ancestry were measured with regard to their impact on accuracy of within-continent substructure estimates . Results summarized in Materials and Methods and Figure S2 indicate that European substructure can be well separated in the presence of up to 5% error , i . e . Indigenous American alleles mistakenly included in the European virtual genomes , which is well above the level of uncertainty ( <2% ) associated with the SABER+ approach . In addition to the HapMap CEU , who are mostly of Northern European ancestry , we used individuals recruited from Dublin , ( Ireland ) , Warsaw ( Poland ) , Rome ( Italy ) and Porto ( Portugal ) to provide references for different areas within Europe . The first two PCs provide good separation of these reference populations , and correspond roughly to North-South and West-East gradients ( Figure 3A ) . Both the MEX1EUR and MEX2EUR virtual genomes are most closely related to intact genomes from Porto , which we interpret as a surrogate for populations from the Iberian Peninsula , [3] , consistent with the historical record that the first European migrants to Mexico were Spaniards . For analysis of the MEXAMR virtual genomes , we introduced 129 individuals representing 8 different Indigenous American populations as reference genomes ( Table S1 ) [16] . Initially , we also included the HapMap CEU based on previous results in which some Indigenous American individuals from the Human Genome Diversity Panel ( HGDP ) were observed to have non-negligible levels of European ancestry [2] . Indeed , the first two PCs for this analysis occur along European-Indigenous American and within-America axes ( Figure 3B ) , and reveal varying levels of European ancestry in the Mayan , Quechua and Colombian populations . In this analysis and subsequent ones carried out in which certain reference populations were removed ( CEU removed from Figure 3C; CEU , Surui , Karitiana and Pima removed from Figure 3D ) , the MEXAMR virtual genomes are most closely related to intact genomes of individuals from southwestern Mexican state of Guerrero ( Guerr ) , which includes Nahua , Mixtec and Tlapanec indigenous groups . Although the Guerrero individuals and the Pima individuals cluster together in Figure 3C , they are separable on PC 3 ( Figure S3 ) , along which the Guerrero , but not Pima individuals , cluster with MEXAMR . The Indigenous American virtual genomes of Mexicans from Mexico City ( MEX1AMR ) are similar to those from Los Angeles ( MEX2AMR ) ; further , we observe a gradient with varying contribution from Mayans , with some Mexicans deriving their Indigenous American ancestry predominantly from Mayans . One individual from Mexico City has an Indigenous American virtual genome that is localized with the Quechua ( arrow , Figure 3C ) and therefore is likely to have a source of Indigenous American ancestry that is distinct from that of the other Mexicans . The ability to accurately construct ancestral virtual genomes from admixed genomes provides a number of opportunities in the areas of human evolution and genetic anthropology . As an example of how such data can be used more generally , we examined the Mexican ancestral components for regions of extended haplotype homozygosity , which mark loci that have undergone recent positive selection . We used the integrated haplotype score ( iHS ) statistic [8] , with a modified normalization procedure so as to fit a standard normal distribution . For the virtual genome SNPs that show the strongest evidence of positive selection , the degree of overlap between the Europeans and Indigenous Americans is similar to that expected by chance ( Figure 4A ) . Specifically , we considered SNPs with |iHS|>2 . 5 , which represent approximately the top 1% scores in either components; 3874 and 3931 SNPs meet this criterion in MEXAMR and MEXEUR , respectively , with 57 SNPs overlap between the two sets ( expected overlap = 40 , p = 0 . 094 ) . Similarly , we found little overlap between the iHS scores in MEXAMR and those computed based on the HapMap populations [8] ( Figure 4B ) . In contrast , the correlation is much higher between the iHS in MEXEUR and those from HapMap CEU ( r = 0 . 79 ) , which reflects shared population and adaptive histories of Southern Europeans ( the MEXEUR ) and the CEU ( mostly from Northern and Central Europe ) . Specifically , of 3257 and 3460 SNPs with |iHS|>2 . 5 in MEXEUR and CEU , respectively , 655 are overlapping ( expected overlap = 32 , p<2 . 2−16 ) ( Figure 4C ) . These findings are consistent with previous observations that intact genomes from the HGDP collection exhibit histories of positive selection that differ according to continent [9] . We also asked what genes might underlie the strongest signatures of positive selection . Towards this end , we grouped SNPs into 50kb windows , selected regions with at least 20 SNPs and at least 10% of SNPs with |iHS|>2 . 5 , and ranked windows by the maximum |iHS| score . The top 10 regions within the MEXEUR and MEXAMR components are shown in Table 1 and Table 2 . The only genomic location that features in both lists is the HLA region on chr 6p , a region known to have experienced strong selection [17]; however , the precise variants that show high iHS scores differ between the Europeans and Indigenous Americans . Outside the HLA region , the most prominent signal in the European component coincides with APBA2 on chr 15q , which is in close proximity to a known pigmentation gene , OCA2 . In the Indigenous Americans component , the strongest signal occurs in chr 6p12 . 3-2; this region harbors numerous genes , including IL17A which is associated with chronic inflammatory diseases such as rheumatoid arthritis , and PKHD1 which is associated with polycystic kidney disease [18] , [19] .
We find extensive variation with respect to continental-level ancestry proportions , both between geographic regions – shown by the much higher Indigenous American ancestry in the Mexico City cohort compared to the HapMap Mexican Americans from Los Angeles – and between individuals within each cohort . This study benefits from the ability to divide the genome of a single Mexican individual into its constituent ancestral components . The ability to trace chromosomal segments to their respective ancestral populations allows us to scrutinize the ancestry origin of each individual within a continent . Within the European component of the Mexican genomes ( MexEUR ) , nearly all individuals , both from Mexico City and from Los Angeles , trace their European ancestries to a Southern European population , as represented in our study by the Portuguese . Within the Indigenous American component of the genomes ( MexAMR ) , a majority of individuals trace their ancestries to groups from the southwest coastal regions of Mexico , consistent with a previous study , which found Zapotec individuals from the State of Oaxaca to best approximate the Indigenous American ancestral population for Mestizos [23] . Importantly , we find evidence of varying levels of Mayan admixture , as well as one individual with Indigenous American ancestry from Bolivia/Peru . Of note , individuals with high levels of Mayan or South American ancestries do not stand out in the continental-level PCA , as their continental-level ancestry proportions are comparable to the rest of the Mexicans . The finding that most Mexican individuals trace their European and Indigenous American ancestry to well-defined geographic regions contrasts sharply the lack of structure in the African ancestry in the African Americans: not only did we trace each African American individual to multiple West/Central West African groups , but the relative proportions are nearly constant across all individuals [24] . This difference can be reconciled by the distinct migratory histories: the African ancestry in African-American populations is largely derived from the trans-Atlantic slave trade , which forcibly departed African individuals from various geographic regions of Western Africa , ranging from Senegal to Nigeria to Angola [25] . In contrast , no evidence suggests massive relocation of the Indigenous Americans during the colonization in North America , and hence reproductive isolation likely has been maintained between geographically separated Indigenous American populations . One limitation of the current study is the incomplete sampling of the Indigenous American populations in our reference panel , which represents two distinct regions in Mexico: the Southwest coastal State of Guerrero and the Yucatan Peninsula . Thus , while most Mexicans trace their Indigenous American ancestries to the indigenous groups from the State of Guerrero ( Guerr ) , it is possible that the true ancestors of the extant Mexicans are an un-sampled group that is genetically similar to Guerr . With the coming of whole genome sequencing data , it is possible that indigenous populations from neighboring states can be distinguished , and thus it may even be possible to detect admixture from closely related Indigenous American groups . The EHH analyses of the Southern European and the Indigenous American components of the Mexican genomes separately revealed numerous intriguing putative targets of recent positive selection . We note that many other approaches have been developed to detect specific types of selective events , and are equally applicable [26] , [27] . We chose to use the iHS test because it has been applied to both the HapMap dataset and the HGDP dataset , thus facilitating comparison . The goal of this paper is not to conduct a comprehensive survey of the selective landscape in the ancestral populations of the present day Mexicans , but rather to illustrate the potential benefits of such endeavors . Given the difficulties in recruiting large samples of non-admixed indigenous individuals from each well-defined Indigenous American group , we argue that admixed populations will provide valuable insight in future endeavors in understanding the evolutionary histories of the Indigenous American populations , some of which may have been extinct . For example , individuals with full Taíno ancestry are rare , but approximately 15% of the contemporary gene pool of Puerto Ricans may have been derived from Taínos . Hence , admixed Puerto Rican genomes can be used to learn about those of the ancestral Taínos [28] . We note that this approach of assembling an ancestral population from a mixed population has also provided important insights in the Aboriginal Australian population in a recent study [29] . Distinguishing between selective events that occurred within the ancestral populations and those that occurred post-admixing requires careful consideration of the tests and associated assumptions . In the current setting , we reasoned that , since a novel adaptive allele is unlikely to be swept to a substantial frequency within a period of less than 500 years ( since the arrival of the Europeans in Mexico ) , and since the EHH method does not have appreciable power to detect low frequency adaptive alleles [9] , most of the signals detected by the EHH had occurred prior to admixing , and hence represent selection within the ancestral populations . On the other hand , the preservation of a long haplotype excludes the possibility of very ancient selective events; this belief is also supported by the observation that there is little overlap between the signatures detected in the Southern European and the Indigenous American components . In previous studies of Puerto Ricans and African Americans , numerous genomic locations were found where locus-specific ancestry deviate from the genome-wide average , and could represent targets of selection in the admixed populations [30] , [31] . In the current analyses , the only locus showing deviation from the genome-wide average is the HLA region on chr 6 , again supporting a population-specific pattern of selection . Therefore , the adaptive history of the Indigenous American groups may vary considerably , and should be studied separately and not as a whole group . Such analyses can be achieved , for example , by examining the Indigenous American components in Mexicans versus that of Puerto Ricans . Our results have important implications for the design of genome-wide association studies based on admixed populations . Epidemiologic studies have found varying prevalence of conditions such as asthma , diabetes and alcohol-related problems across Hispanic national groups [28] , [32] , [33] . Distinct population and adaptive history among Hispanics ethnic groups can give rise to heterogeneity in complex traits . Therefore , the importance of accounting for intra-continental genetic structure in disease mapping studies , in addition to adjusting inter-continental admixture proportions , needs to be carefully evaluated .
The Mexican individuals analyzed in this project come from two sources: a panel of 492 Mexican parent-offspring trios recruited from Mexico City as part of a previous genome-wide association study ( MEX1 ) [14] , and 23 HapMap Phase3 Mexican trios recruited from Los Angeles , California ( MEX2; http://hapmap . org ) . For estimating locus-specific ancestry , we used the HapMap CEU ( N = 88 ) and YRI ( N = 100 ) individuals for the ancestral populations . To analyze the European component of the admixed genome , we augmented the Mexican datasets with individuals recruited from Dublin , Ireland ( N = 43 ) , Rome , Italy ( 45 ) , Warsaw , Poland ( N = 45 ) and Porto , Portugal ( N = 43 ) . For the Indigenous American component analyses , we combined the data generated in two previous studies [2] , [16] . Four Mayan individuals with substantial European admixture are removed . The combined set used for the subsequent analyses includes 14 individuals from Guerrero , Mexico ( two Nahua , seven Mixtec and five Tlapanec ) , 24 Mayan individuals from the Yucatan Peninsula , 24 Quechua collected in Cerro de Pasco , Peru , 25 individuals of largely Aymara ancestry collected in La Paz , Bolivia , 13 Karitiana and eight Surui from Brazil , seven Colombians , and 14 Pima . Because the sample sizes for Nahua , Mixtec and Tlapnec are small , and all individuals were recruited from the same state , we considered these individuals as one group . Table S1 summarizes the individuals used for each analysis . Genotyping and quality control procedures have been described in the primary publications for each dataset , except for the dataset of 176 European individuals . Briefly , MEX1 and the HGDP individuals were genotyped on Illumina 550K and on 650K Beadchip , respectively . The Indigenous American individuals from Bigham et al . ( 2009 ) were genotyped on Affymetrix 1M SNP arrays [16] . The set of 176 European individuals were genotyped using Illumina HumanHap300 arrays; this dataset originally included 180 individuals; four individuals were found with non-negligible non-European ancestry and were excluded . SNPs with a call rate of less than 95% were excluded . The number of individuals and markers used for each analysis is summarized in Table S1 . We used BEAGLE to construct haplotypes for Mexican trios [34] . As children provide no additional information regarding population structure or adaptation , they are not used in subsequent analyses . Continental-level admixture proportions were estimated two ways: ( 1 ) a model-based clustering algorithm implemented in frappe [35] , and ( 2 ) average locus-specific ancestries across all markers . Locus-specific ancestry was estimated with SABER+ , an extension of a previously described approach , SABER , that uses a Markov-Hidden Markov Model [12] . SABER+ differs from SABER in implementation of a new algorithm , an Autoregressive Hidden Markov Model ( ARHMM ) , in which haplotype structure within the ancestral populations is adaptively constructed using a binary decision tree based on as many as 15 markers , and which therefore does not require a priori knowledge of genome-wide ancestry proportions ( Johnson et al . , in preparation ) . In simulation studies , the ARHMM achieves accuracy comparable to HapMix [36] but is more flexible in modeling the three-way admixture in the Mexican population and does not require information about the recombination rate . HapMap CEU and YRI individuals were used as the reference ancestral populations . Based on frappe and supported by PCA , 50 individuals in MEX1 set have more than 95% Indigenous American ancestry . These individuals were initially used to approximate the Indigenous American ancestors in the locus-specific ancestry analyses; an iterative procedure is used to identify and correct for the non-Indigenous American segments in these individuals . Accuracy of the locus-specific ancestry is verified by performing a PC analysis , treating each individual as three non-admixed genomes , MexEUR , MexAMR , and MexAFR ( see section “subspace PCA” below ) . We implemented this algorithm to accommodate the large amount of missing genotype data in partially masked virtual genomes , and used it to derive all the PCA results reported here . The statistical theory of the algorithm in a general data mining context can be found in [13]; however , various modifications are required for the current setting , as described below . Let Gh ( h = 1 , 2 ) be two N×M matrices , in which denote the unordered pair of alleles at SNP m ( m = 1 , … , M ) in individual n ( n = 1 , … , N ) ; the columns of Gh are standardized to have mean 0 and variance 1 . To compute the subspace spanned by the first k principal components ( PC ) , we begin by finding a matrix decomposition , , which minimizes the reconstruction error , R , defined as:subject to the constraints that the column vectors of A are of unit norm and mutually orthogonal and the row vectors of S are also mutually orthogonal . Here A is a N×d matrix , S is a M×d matrix , and d<N≤M represents the desired number of leading PC's . The algorithm we use is a generalized instance of the coordinate descent approach [37] , which iteratively optimizes matrix A for fixed S and then optimizes S fixing A according to the rules:where λ is a learning rate , the superscripts , r , indicate iteration , and the subscripts , j , denote the j-th column of a matrix . It can be shown that the columns of A and S span the subspace of the first d PCs , and that the leading PCs can be computed by orthogonalizing the columns of A and S [13] . To evaluate the accuracy of our modified ssPCA approach , we applied it in parallel with Eigenstrat [21] to the intact Mexican genomes , and found the leading PCs produced by the two algorithms were virtually identical , up to a permutation of signs . We carried out two simulation experiments to evaluate the impact of statistical uncertainties associated with estimating locus-specific ancestry , and to investigate the performance of the ssPCA approach . In the first set of simulations , we created 10 datasets in which 400 admixed genomes were modeled to mimic a Latino population: each individual draws chromosomal segments from European and Indigenous American ancestry , and the proportion of Indigenous American ancestry in each individual matches what we observed in MEX1 . For 200 individuals , European-derived segments were sampled from the HapMap CEU haplotypes , representing Northern and Western European ancestry , while for the remaining 200 individuals , the European-derived segments were sampled from MexEUR inferred from the actual Mexican genotype data , representing Southern European ancestry . The chromosomal segments from CEU and MexEUR in the admixed individuals were treated as the true European virtual genomes . To evaluate the potential impact of statistical uncertainty , we introduced random errors in which the true identities of European vs . Indigenous American segments were switched with probability ε . The top PC for each set of simulated virtual genomes ( at each of 8 error rates , ε = 0 . 01–0 . 20 ) was computed with ssPCA . We evaluate the effect of these errors by calculating a confusion fraction , ξ , that quantifies the accuracy with which the estimated first PC separates individuals with Northern vs . Southern European ancestry , and is defined as the proportion of individuals that lie on the “wrong” side of a threshold that best separates the two groups . Thus , ξ can range from 0 ( perfect separation ) to nearly 50% ( complete confusion as would be observed for genetically homogenous groups ) . Finally , we analyze each of the 10 datasets using SABER+ , exactly as was done for real data: apply ssPCA to estimate substructure , and calculate a confusion fraction . The results for this simulation experiment are depicted in Figure S2 , and show that the confusion fraction increases substantially , from a mean of 2 . 14% to 17 . 5% , at error rates between 0 . 03 and 0 . 05 . Using SABER+ on these same 10 datasets yields a mean confusion fraction of 1 . 58% , which corresponds to an error rate <0 . 02 ( indicated by the arrow in Figure S2 ) . In a second set of simulations to investigate the ability of the ssPCA approach to deal with missing data , we created five datasets in which the proportion of genome-wide European ancestry in each of 400 admixed genomes was fixed at either 50% or 30% , respectively . Applying SABER+ and ssPCA to these datasets yields mean confusion fractions of 0 and 0 . 7% , respectively , indicating that our approach performs well for situations such as the one described here , where mean genome-wide continental ancestry proportions are above 30% for both the European and the Indigenous American components . We used the number of ancestry blocks in an individual as summary statistics . Tracing through a pedigree of T generations , the expected number of recombination events in a haploid genome is 0 . 01×TL , where L is the total genome length ( taken to be 3435cM [38] ) . Under a hybrid-isolation model and assuming a genome-wide ancestry proportion of z , a fraction of 2×z ( 1−z ) of the recombination events occurs between two haplotypes of opposite ancestry and thus leads to transitions in ancestry . When we count the number of ancestry blocks in the real data , we do not observe recombination events that occur between two haplotypes of the same ancestry . Hence the expected number of ancestry switches in a diploid genome is B = ( 2×2×0 . 01 ) ×TL×z ( 1−z ) , and each ancestry switch creates one additional ancestry block . When there is no ancestry switch in a genome , the number of ancestry blocks is defined to be the same as the number of chromosomes . Therefore , for each specific time of admixing , T , we computed the expected number of ancestry blocks as B+2×22 , with the genome-wide ancestry proportion , z , varying from 0 to 1 at 100 equally spaced grid points . Each curve in Figure 2D shows the expected number of ancestry blocks as a function of admixture proportions for a specific admixing time . The estimated numbers of European and non-European ancestry blocks from the Mexican individuals were tallied and compared to the expected values . To assess the impact of uncertainty associated with estimating the number of ancestry blocks , we note that errors in estimating locus-specific ancestry often create very short ancestry blocks . Hence , we simulated admixed genomes according to the hybrid-isolation model , but removed extremely short blocks ( segments with <10 SNPs ) from both simulated genomes and real data . The estimated admixing time remained the same under this alternative analysis , suggesting the estimated admixing time is relatively robust . The hybrid-isolation model was chosen because of the mathematical simplicity; under a more realistic continuous gene-flow model , the estimated times of admixing should be interpreted as an approximation of average admixing time , weighted by the relative level of gene-flow in each generation . To assess the sub-continental population structure , each Mexican genome was partitioned into three non-admixed genomes , by masking ( i . e . setting to missing ) alleles from all but one ancestral population . In other words , the European component of a Mexican's genome ( MexEUR ) was derived by treating as missing all alleles whose origins were inferred as African or Indigenous American . For within European analysis , we applied ssPCA on the dataset consisting of MexEUR ( including both Mexico City and HapMap samples ) , 88 HapMap CEU and 176 European individuals from four cities: Dublin ( Ireland ) , Warsaw ( Poland ) , Rome ( Italy ) and Porto ( Portugal ) . Because of the limited number of informative haplotype segments , Mexican individuals with less than 25% European ancestry were excluded from this analysis . In an analogous fashion , we analyzed the Mexican component of the genome , MexAMR , along with 129 indigenous Indigenous American individuals representing 8 populations ( Table S1 ) . Because the African ancestry is low in both Mexican cohorts ( 3% and 5% , respectively ) , we did not analyze the within-Africa population structure . or all SNPs with frequencies between . 05 and . 95 in the respective populations , iHS was calculated following Voight et al . [8] with two modifications . First , haplotype homozygosity scores for a core SNP is computed on the subset of haplotypes in which the core SNPs are derived from a specific population . If a haplotype is truncated because of an ancestry change , the haplotype beyond the ancestry switch point is considered different from all other haplotypes in the corresponding interval . We have also considered an alternative strategy in which only haplotypes that do not have an ancestry change within 400 SNPs from the core SNPs are included in the calculation; the results are virtually identical . Second , instead of binning SNPs by the inferred ancestral allele frequencies and calculating the standard deviation of iHS in each bin , we used a quantile regression to estimate the 25th- and 75th-percentile of the empirical null distribution as a function of the minor allele frequency . The raw iHS scores were then normalized by the estimated inter-quartile range within each chromosome; the resulting standardized iHS scores fit a standard normal distribution well . To define regions that may harbor recently adaptive alleles , we seeded a region by a window of 50kb around a SNP with extreme iHS scores; we then successively scanned to the left and to the right , 50kb a time , merging neighboring regions in which at least one SNPs has an |iHS|>2 . 5 ( which represents the 99th-percentile of the scores ) ; finally , the top 10 list in Table 1 requires that at least 10% of the SNPs in the region have |iHS|>2 . 5 . The proportion of SNPs with high |iHS| was the criterion used by Pickrell et al . [9] . It has been suggested that genome-wide and locus-specific ancestries may show particular poor correlation at loci under selection compared to neutrally evolving loci [39] . We did not observe such a trend at loci with the highest iHS scores in either MEXAMR or MEXEUR . | Admixed individuals , such as African Americans and Latinos , arise from mating between individuals from different continents . Detailed knowledge about the ancestral origin of an admixed population not only provides insight regarding the history of the population itself , but also affords opportunities to study the evolutionary biology of the ancestral populations . Applying novel statistical methods , we analyzed the high-density genotype data of nearly 1 , 500 Mexican individuals from Mexico City , who are admixed among Indigenous Americans , Europeans , and Africans . The relative contributions from the three continental-level ancestral populations vary substantially between individuals . The European ancestors of these Mexican individuals genetically resemble Southern Europeans , such as the Spaniard and the Portuguese . The Indigenous American ancestry of the Mexicans in our study is largely attributed to the indigenous groups residing in the southwestern region of Mexico , although some individuals have inherited varying degrees of ancestry from the Mayans of the Yucatan Peninsula and other indigenous American populations . A search for signatures of selection , focusing on the parts of the genomes derived from an ancestral population ( e . g . Indigenous American ) , identifies regions in which a genetic variant may have been favored by natural selection in that ancestral population . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genome-wide",
"association",
"studies",
"genomics",
"mathematics",
"evolutionary",
"biology",
"statistics",
"genetics",
"population",
"genetics",
"biology",
"computational",
"biology",
"statistical",
"methods",
"genetics",
"and",
"genomics"
] | 2011 | Ancestral Components of Admixed Genomes in a Mexican Cohort |
Mutator phenotypes accelerate the evolutionary process of neoplastic transformation . Historically , the measurement of mutation rates has relied on scoring the occurrence of rare mutations in target genes in large populations of cells . Averaging mutation rates over large cell populations assumes that new mutations arise at a constant rate during each cell division . If the mutation rate is not constant , an expanding mutator population may contain subclones with widely divergent rates of evolution . Here , we report mutation rate measurements of individual cell divisions of mutator yeast deficient in DNA polymerase ε proofreading and base-base mismatch repair . Our data are best fit by a model in which cells can assume one of two distinct mutator states , with mutation rates that differ by an order of magnitude . In error-prone cell divisions , mutations occurred on the same chromosome more frequently than expected by chance , often in DNA with similar predicted replication timing , consistent with a spatiotemporal dimension to the hypermutator state . Mapping of mutations onto predicted replicons revealed that mutations were enriched in the first half of the replicon as well as near termination zones . Taken together , our findings show that individual genome replication events exhibit an unexpected volatility that may deepen our understanding of the evolution of mutator-driven malignancies .
A network of DNA metabolic activities maintains genomic integrity during each cell division [1] , ensuring that eukaryotic mutation rates remain less than one mutation per billion base-pairs synthesized . Defects to these activities can lead to mutator phenotypes that increase the rate of mutation [2] . As the mutator population expands , genetic diversity increases , fueling evolution . In multi-cellular organisms , mutator phenotypes accelerate tumorigenesis by generating mutations that overcome the genetic and environmental barriers to unrestrained proliferation [3 , 4] . In tumors that are not initially mutator-driven , chemotherapeutic treatment provides selection pressure for sub-clonal mutator cell lineages to emerge , which more easily evolve drug-resistance . Thus , mutator phenotypes may pose substantial challenges to cancer therapy , necessitating a greater understanding of their inherent vulnerabilities . The most abundant source of potential mutations in dividing cells are polymerase errors , which are corrected by the synergistic activities of polymerase proofreading and mismatch repair ( MMR ) [2] . Pol ε and Pol δ perform the bulk of leading and lagging strand DNA replication in eukaryotes , respectively [5] , and contain intrinsic proofreading exonucleases that excise the vast majority of polymerase errors . Mismatches that escape proofreading are recognized by Msh2•Msh6 ( base-base mismatches , insertion/deletion mispairs ) or Msh2•Msh3 ( primarily insertion/deletion mispairs ) [2] . These complexes recruit the endonucleases Mlh1•Pms1 ( Pms2 in mammals ) or Mlh1•Mlh3 , which initiate processing and re-synthesis of the DNA [2] . Defects to proofreading or MMR increase mutation rates in microbes and mammalian cells [2] . In humans , mutations that compromise Pol ε or Pol δ proofreading or MMR lead to colorectal ( CRC ) and endometrial cancers ( EC ) [6–11] , supporting the hypothesis that maintenance of DNA replication fidelity restrains neoplasia [3 , 4 , 12–14] . Synergistic defects in both MMR and proofreading greatly accelerate tumorigenesis [15] . Since proofreading and MMR act in series on the same pool of errors , concomitant defects in these activities elevate mutation rates 10 , 000-fold in diploid yeast [16–18] . In haploid yeast , this level of mutagenesis causes error-induced extinction [16 , 17 , 19 , 20] . Not all proofreading and MMR defects are synthetically lethal to haploids . Yeast cells with mutant alleles for Pol ε proofreading deficiency ( pol2-4 ) and Msh6 ( msh6Δ ) exhibit mutation rates of only 1000-fold greater than background [21 , 22] , just below the critical level at which haploid colony forming capacity declines [20] . Strong mutator phenotypes may be more volatile than commonly appreciated . The first hints of hypermutability came from differences observed between haploid and diploid yeast in the rates of base-analogue 6-hydroxylaminopurine ( HAP ) induced mutagenesis [23 , 24] . Subsequent studies revealed a wide variability in the mutagenesis induced in diploids by HAP or AID/APOBEC cytosine deaminase expression: clones selected for a mutant phenotype had much higher genome-wide mutational loads than unselected clones exposed to the same mutagenic treatment [25] . A similar hypermutable state has been advanced to explain why diploid strains deficient in Pol δ proofreading display mutation rates 3 to 20-fold greater than isogenic haploid yeast strains [26] . These results are consistent with the hypothesis that some Pol δ proofreading-deficient cells enter a “hypermutator” state , which is lethal to haploids but tolerated by diploids [26] . If mutational processes are similarly volatile during tumorigenesis , they may influence the rate of tumor evolution and the nature of genetic diversity present in the growing tumor clone . Testing for mutator volatility has proven technically challenging . Historically , mutator phenotypes have been measured by scoring the frequency of rare mutations in selectable genes in thousands or even millions of cells . Analysis of the fluctuation in mutation frequencies in multiple independent cultures yields the mutation rate of the target gene during clonal expansion [27] . In an alternative approach , individual cell lines are propagated through bottlenecks over hundreds or thousands of generations and then analyzed by whole genome sequencing to derive the generational mutation rate [28 , 29] . Both of these methods can only report the average mutation rate of the entire population , which obscures the actual mutation rate for any given replication event . Overcoming this limitation requires the measurement of mutation rates at single cell resolution . As an experimental approach , single cell DNA sequencing holds promise , but current methods require in vitro enzymatic amplification of the genome [30 , 31] . Because DNA polymerases are used to amplify the DNA , base misincorporation events can lead to the scoring of thousands of false mutations . Additionally , the spatial and temporal relationship between cells is lost in these experiments; thus , it is impossible to know precisely how many cell divisions occurred between any two related cells . We devised an alternative approach , which is to sequence clones of cells derived from sequential cell divisions of the same cell lineage . Each clone contains the mutational history of the replication event , as well as all previous genome replications . By comparing clones derived from sequential cell divisions , it is possible to determine precisely when each mutation arose . Here , we apply this strategy to determine the fidelity of individual genome replication events of pol2-4 msh60Δ mutator yeast cells .
We used budding yeast cells to investigate mutation rates of individual cell divisions because they divide asymmetrically into “mother” and “daughter” cells that are easily separated by micromanipulation . The daughter cells readily expand into clones , which can then be subjected to whole genome sequencing to ascertain mutations . During the first division of the mother cell ( M0→M1 ) , new DNA replication errors retained by the mother ( M1 ) in the form of mismatches become mutations in the next round of replication and segregate to the daughter ( D2 ) or are retained by the mother ( M2 ) ( Fig 1A ) . Mutations and mismatches segregated to the daughter will be unique to her clone of cells , whereas mutations retained by the mother cell ( maternal mutations ) will appear in the third daughter ( D3 ) clone and all subsequent daughter clones ( e . g . D4 , D5 , etc ) . The number of new maternal mutations in each daughter clone can be used to determine the mutation rates of individual maternal cell divisions . As a source of mother cells , we used haploid spores freshly dissected from tetrads derived from meiosis of a diploid strain that was heterozygous for both alleles ( POL2/URA3::pol2-4 MSH6/msh6Δ::LEU2 ) . The four haploid genotypes from this strain are: 1 ) wild-type ( WT ) with respect to replication fidelity ( POL2 MSH6 ) , 2 ) proofreading defective ( pol2-4 MSH6 ) , 3 ) MMR defective ( POL2 msh6Δ ) , or 4 ) proofreading and MMR defective ( pol2-4 msh6Δ ) . Canavanine-resistance ( Canr ) mutation rates determined by fluctuation analysis [32–34] revealed that the pol2-4 and msh6Δ alleles individually increased mutation rates 3 and 10 times above background , respectively ( S1A Fig ) [21 , 22] . The pol2-4 msh6Δ cells have mutation rates of 1 . 7 x 10-4 Canr mutants per cell division ( S1A Fig ) , which corresponds to 4 . 3 x 10-7 mutations/bp/cell division using the method of Drake [35] that we employed previously [20] . Assuming at least 80% sequencing coverage of the haploid yeast genome ( 1 . 2 x 107 bp ) and that the mutability of CAN1 is representative of other genes , we estimated we would observe an average of 4 to 5 mutations for each division of pol2-4 msh6Δ cells [ ( 4 . 3 x 10–7 mutations/bp/cell division ) x ( 1 . 2 x 107 bp ) x 0 . 8 = 4 mutations] . To establish single cell lineages , we randomly chose spores to serve as mother cells and moved them to a defined location on an agar plate . When the mother cells began dividing , we moved all daughters to unique positions on the plate where they formed colonies ( S1B Fig ) . We then sequenced the genomes of the viable daughter clones , scoring only mutations at genomic positions accurately called in all members of a lineage . In all lineages , we assessed greater than 80% of the yeast genome ( 1 . 05 ( ±0 . 05 ) x 107 base-pairs; Table 1 ) . We then compared the sequence of the last viable daughter clone to the sequences of all earlier daughter clones to define mutations fixed in the maternal lineage at each cell division . We observed no mutations in three WT control lineages , while two mutations were fixed during 9 cell divisions of an msh6Δ mother cell ( Lineage A in S1 Dataset ) . In contrast , we observed an average of 30±7 mutations during 11±3 divisions of pol2-4 msh6Δ mother cells ( Lineages B-H in S1 Dataset , summarized in S1 Table ) . All told , 237 mutations accumulated over 87 divisions of pol2-4 msh6Δ cells ( Table 1 ) . The average mutation rate was 2 . 6 x 10-7 mutations/bp/cell division , determined by dividing the total number of mutations in all lineages by the total base-pairs scored ( Table 1 ) . The mutations were distributed across all 16 chromosomes ( Fig 1B ) , although mutation rates for individual chromosomes varied six-fold ( S1C Fig ) . The mutation spectra , characterized by high numbers of GC→AT and GC→TA mutations , corresponded well with the combined published ura3 and can1 mutation spectra of pol2-4 msh6 cells ( Fig 1C , S2 Table ) , as well as with spectra obtained with purified proofreading-deficient Pol ε [21 , 22] . AT→TA mutations , also frequently observed in vitro , were relatively less abundant in the whole genome spectra , but this can be explained by the preferential repair of these mismatches by the Msh2•Msh3 complex [36] , which remains active in pol2-4 msh6Δ cells . Individual cell divisions exhibited considerable variation in mutation counts ( Fig 2A ) that did not correlate with the replicative age of the mother cell ( r = 0 . 1 , p = 0 . 4 , Pearson ) . We modeled mutagenesis in pol2-4 msh6Δ cells as a Poisson process to test whether this variation could be explained by a single overall mutation rate . Long used to model random mutagenesis [27 , 37] , the Poisson function calculates the probability of a defined number of rare events ( k ) occurring within a fixed interval of time at a given rate parameter ( λ ) . In Eq 1 , Pk is the probability of cells fixing k mutations in a single cell division . For our purposes , λ is defined as μ , the average per-base-pair mutation rate ( 2 . 6 x 10-7 mutations/bp/cell division from Table 1 ) , times G , the size of the sequenced genome in base pairs; e is the base of the natural logarithm . The resulting probability for each value k , multiplied by the number of cell divisions scored , gives the expected number of divisions with k mutations . We separately modeled Poisson distributions to the data of each lineage to account for differences in the number of base-pairs sequenced and cell divisions scored ( S2A Fig ) and then summed the Poisson distributions together . The resulting “Summed-Poisson model” poorly fit the combined observations from all lineages due to overdispersion of the data ( Fig 2B , S2B Fig ) . We considered the possibility that overdispersion could be due to zero-inflation from under-sampling of genomic sites in lineages with the smallest sequenced genome sizes ( Lineages C , D , G1 , G2 , Table 1 ) . However , we found no correlation between the number of samples in a lineage with 0 mutations and the size of the sequenced genome , suggesting that under-sampling is not the source of the overdispersion ( S3 Fig ) . We reasoned that overdispersion could result from a mixture of underlying distributions generated by two or more mutation rates . To test this hypothesis , we first generated a simplified model of the data that grouped all cell divisions into one set , utilizing the average genome size ( 1 . 02 x 107 bp , Table 1 ) . The Simplified and Summed Poisson distributions were virtually identical ( S2C Fig ) , suggesting that the data could indeed be modeled as a single set . To compare how well single-distribution and distribution-mixture models fit the data , we used finite mixture modeling ( FMM ) , which is a computational approach that fits mixtures of parametric distributions to data [38] . Because the Poisson distribution is described by the single parameter λ ( μG in eq 1 ) , the number of parameters in each model is equal to the number of Poisson distributions underlying the composite probability mass function . Fitted models included one to five parameters ( i . e . , mutator states ) . Because parameter dimensionality increases by one between each of the five fitted models , a best-fit model was selected by comparison using maximum-likelihood-ratio tests of nested hypotheses with one degree of freedom for each test . The best-fit model described the data significantly better than models with fewer parameters , and not significantly worse than models with more parameters . The best fit model by maximum likelihood estimation was a Two-Poisson distribution with values for λ of 0 . 402 and 3 . 897 . The difference between the Poisson and Two-Poisson distributions by the likelihood ratio test was highly significant ( Likelihood Ratio , Chi-Square Test = 40 . 58 , Degrees of Freedom = 1 , p < 1 . 9 x 10–10 ) . The difference between the best Two-Poisson and Three-Poisson models was insignificant ( Likelihood Ratio , Chi-Square Test = 0 . 80 , Degrees of Freedom = 1 , p = 0 . 37 ) , indicating that increasing the number of Poisson distributions in the mixture does not improve the fit to the data . Thus , the best-fit model was a mixture of two Poisson distributions . To study the relative contribution of these two distributions to the observed mutation count , we constructed a “Summed Two-Poisson” model that took into account differences between the lineages ( S2 Dataset ) . We calculated the expected single Poisson distributions of mutation counts for each lineage assuming a hypo- ( 0 . 4 x 10-7 mutations/bp/cell division ) or hypermutator state ( 4 x 10-7 mutations/bp/cell division ) . We summed the Poisson distributions from all lineages to obtain the expected distribution of mutation counts across the entire experiment for cells with a hypo- or hypermutator state . We then combined the hypo- and hypermutator Poisson distributions , with each contributing 50% to the final mixture . After comparing this mixture to the observed distribution of the data , we adjusted the contribution of each mixture component , ultimately finding that a model with 35% hypomutator divisions and 65% hypermutator divisions provided the best fit ( Fig 2B ) . Thus , on the strength of the above hypothesis testing and modeling , we propose pol2-4 msh6Δ mutator mother cells assume either a hypomutator state or a hypermutator state as they pass through S-phase , with mutation rates that differ by an order of magnitude . In the Two-Poisson model , the bulk of observed mutations would arise in cell divisions with a hypermutator state . Only 13 of the 237 mutations would arise during hypomutator cell divisions . Close examination of the mutations in error prone cell divisions revealed numerous instances in which pairs or trios of mutations occurred on the same chromosome . To determine the significance of this pattern , we computationally simulated the experiment 10 , 000 times , assuming random mutagenesis . Each round of the simulation returned a value for the number of random co-occurrences of 2 or 3 mutations on the same chromosome . Plotting the values from all 10 , 000 iterations gave 95% confidence intervals of 18 to 33 observations of 2 or more mutations on the same chromosome in the same cell division and 0 to 6 observations of 3 mutations . We observed 39 instances of 2 or more mutations and 8 observations of 3 or more mutations ( Fig 3 ) , suggesting that the hypermutator state may be expressed in only a portion of the genome in a given cell cycle . To explore the relationship between mutagenesis and replication dynamics , we mapped mutations onto the yeast DNA replication profiles from Raghuraman and colleagues [39] ( S4 Fig ) . These replication profiles , generated using isotopic labeling time-course experiments and high density microarrays , report the timing ( trep ) of 50% DNA replication within a sliding 10kb window , quantified every 500 base-pairs across the genome . Local maxima and minima represent putative locations for origins and termination zones and the line between the two denotes replicons . We found the chromosomal positions of mutation pairs occurring in the same chromosome and cell division were not correlated ( r = 0 . 24 , p = 0 . 2 , S5A Fig , S3 Dataset ) . Only five sets of co-occurring mutations reside in the same replicon ( S4 Fig , see Chrs . 4 , 12 , 13 , and 16 ) , with two pairs affecting the same replicon on chromosome 12 ( see purple and teal triangles ) . The remaining co-occurring mutation pairs were separated by multiple replicons , consistent with the hypothesis that they arose from independent Pol ε complexes . Intriguingly , co-occurring mutation pairs frequently reside in DNA on the same chromosome with similar trep values ( r = 0 . 47 , p = 0 . 006 , 2-tailed , Pearson , S5B Fig , S3 Dataset ) [39] , suggesting that the polymerase errors may have been committed at a similar time during S-phase . The same correlation is not apparent when all pairwise combinations of co-occurring mutation trios are considered; however , in 6 out of 8 trios , two of the three mutations did occur in DNA with similar trep values , consistent with a temporal relationship . We also examined the predicted replication timing of mutations occurring on different chromosomes in the same cell division . As with the mutation trios , no correlation is observed when all pair-wise combinations are considered . These observations suggest that while the hypermutator state may have periods of increased mutagenesis affecting a fraction of the genome , it may not be temporally constrained . Mapping the mutations onto the replication profiles revealed an enrichment of mutations near origins and termination zones ( Fig 4 ) . To examine the overall distribution of mutations within replicons , we determined the distance of each mutation to the nearest origin and termination zone [39] , and binned mutations by their fractional distance to the origin in 0 . 1 increments ( Fig 4 ) . More mutations occurred in bins closer to the origin ( 130 ) than in bins closer to the termination zone ( 107 ) . The disparity would have been even greater were it not for a dramatic increase in the number of mutations in the bin closest to the termination zone ( Fig 4 ) . This distribution of mutations significantly deviated from that expected by chance ( Chi-Square Test , p = 0 . 005 ) , suggesting that mutagenesis in pol2-4 msh6Δ cells may be influenced by the dynamics of replication .
Mutagenesis has long been modeled as a Poisson process under the assumption that mutations occur independently in each cell division with a constant mutation rate parameter [27 , 37] . We set out to test the hypothesis that mutation accumulation in pol2-4 msh6Δ cells results from a single Poisson process guided by a constant mutation rate . Using finite mixture modeling and likelihood ratio tests we found that a Two-Poisson model fit the mutation count data significantly better than a single Poisson model . Thus , a single mutation rate does not appear to underlie the generation of mutations in pol2-4 msh6Δ cells . As an alternative to the simple idea of two mutator states , we also considered a negative binomial model in which each cell division adopts a different mutation rate , with the rates being gamma distributed . This scenario produces a Poisson model . We compared the negative binomial and Two-Poisson models using Akaike’s Information Criterion ( AIC ) [40] , which distinguishes between classes of models on the basis of both goodness-of-fit and parsimony . The AIC for the negative binomial model is 368 . 49 , whereas the AIC for the Two-Poisson model is 363 . 93 ( lower is better ) . The relative likelihood ratio of the negative binomial model over the Two-Poisson model is 0 . 102 , indicating that the Two-Poisson model provides a better fit to the data . In the Two-Poisson model , we have proposed that the distribution of mutation counts is caused by two underlying mutation rates . However , a second statistical process determines how many new mutations appear in a mother cell besides mutation rate: namely the segregation of mutations between mother and daughter cells . It is conceivable that there is a single underlying mutation rate , but two types of biased segregation patterns . In one segregation pattern , the mother cell would retain the mutations with a high probability . In the other pattern , the mother would have a low probability of retaining the mutations . By varying the frequency of the two segregation patterns and the degree of segregation bias , it is possible to produce two overlapping Poisson distributions identical to the two-state mutator model . While we cannot formally distinguish between the two models , we find it easier to imagine how the error rates of mutator polymerases may be modulated than how there would be two distinct unequal segregation patterns of mutations . The distinct mutator states proposed for pol2-4 msh6Δ cells could derive from a process that influences the mutator activity of proofreading-defective Pol ε . The modeled per-genome mutation rate of the hypomutator state ( 0 . 4 ) matches the rates recently described for MMR-deficient haploid ( 0 . 71 ) and diploid ( 0 . 38 ) yeast mutation accumulation lines [41] . This correspondence suggests that cell divisions with a hypomutator state do not appreciably express the Pol ε proofreading-deficient phenotype . There may be a regulatory switch that influences Pol ε rates of misincorporation or mispair extension , the action of alternative repair mechanisms that edit Pol ε errors , or the extent to which proofreading-deficient Pol ε contributes to the overall replication of the genome . Our proposal for two distinct mutation rates in mutator cells contrasts previous work with Escherichia coli , which found that mutation count data in individual mutator cells conforms to a single Poisson distribution [42] . This work elegantly followed the occurrence of likely mutations by the formation of persistent , fluorescently labeled MMR foci that form when there is a failure to repair a mismatch . The contrasting results between the two studies could stem from either technical or biological differences between the two experimental systems . A notable limitation of counting fluorescent foci is the potential for undercounting . Specifically , it may be difficult to resolve high numbers of foci in cells with a hypermutator state , especially if the mutations occur in close proximity . In addition , if the hyper-mutator state saturates MMR , not all mismatches would lead to fluorescent foci . Genetically , we relied on tandem deficiencies in MMR and polymerase proofreading to raise mutation rates to an appreciable level , whereas the work in bacteria used strains deficient in either MMR ( mutH ) or proofreading . Of course , intrinsic differences between prokaryotic and eukaryotic DNA replication and repair could also explain the contrasting data sets . In particular , prokaryotes utilize a single replicative polymerase ( Pol III ) for both leading and lagging strand bulk DNA synthesis whereas eukaryotes divide the labor between Pol δ and Pol ε . As discussed below , Pol δ may replace Pol ε at some point during leading strand replication , providing a potential avenue by which the contribution of proofreading-deficient Pol ε to genome replication and mutagenesis may be modulated . An important clue to the mutator volatility of pol2-4 msh6Δ cells lies in the observation that pairs of mutations occurred on the same chromosome more frequently than expected by chance ( Fig 3 ) . The correlation in predicted replication timing of co-occurring mutation pairs suggests that the hypermutator state may be linked to replication dynamics of individual chromosomes . Chromosomes occupy distinct regions within the nucleus [43] and evidence exists for replication factories consisting of multiple active replisomes acting on distinct origins with similar replication timing [44] . Thus , changes in replication fidelity could be factory-specific , restricting the mutator phenotype to limited regions of the genome in a given cell cycle . If the phenomenon that underlies focal expression of the hypermutator state were to extend to the entire genome , much higher genome-wide mutation rates may result . Our ability to detect much higher genome-wide mutation rates in the current study is limited by the likelihood that extreme mutator states would result in haploid lethality . Future studies with diploid cells , which are buffered against recessive deleterious mutation accumulation , will be required to explore the full extent of mutator volatility . We investigated the link between replication dynamics and mutagenesis by mapping mutations onto published replication profiles [39] . This approach has some limitations worth noting . Replication profiles average the replication timing of large numbers of cells , yet replication initiation events are probabilistic phenomena that vary between cells [45] . Variation in origin firing likely leads to variation in the termination of DNA replication as most termination occurs independently of sequence context within zones located between adjacent firing origins [46] . Since no two cells follow the same temporal order of origin firing , there may be substantial variation in replicons . Someday it may be possible to monitor replication fidelity and dynamics simultaneously . Until then , we feel the best approach is to map mutations onto replicons that incorporate the probability of origin firing , recognizing that in some cases , these “probabilistic” replicons may differ from the actual replicons in which the mutations occurred . In our study , we normalized the positions of mutations within the probabilistic replicons using fractional distances ( Fig 4 ) . In cases where the probabilistic and actual replicons differ , the fractional distances would be inaccurate and likely diminish any signal for the enrichment of mutations near termination zones . In our view , this makes evidence for enrichment even more compelling . Ample evidence supports the hypothesis that Pol ε performs leading strand DNA replication near origins of replication [5 , 47] . Whether Pol ε remains the leading strand polymerase through the end of each replicon continues to be debated [48] . Replacement of Pol ε with Pol δ may be important for joining the leading strand with the downstream Okazaki fragment [49] . We found more mutations in the first half of the replicon than in the second half . This uneven distribution is consistent with a model in which proofreading-deficient Pol ε is replaced by Pol δ with increasing probability as replication proceeds [48] . If replacement of Pol ε with Pol δ after initiation is subject to regulation , the hypomutator state could be explained by hyper-activation of a mechanism that replaces proofreading-deficient Pol ε with Pol δ . We also found an unexpected concentration of mutations near the termination zones , suggesting that proofreading-deficient Pol ε replisomes that do make it to the termination zone may become especially error prone . The relationship between mutagenesis and replication dynamics has also been explored in a recent mutation accumulation study that utilized MMR-deficient strains expressing either WT Pol ε or a mutant variant with the M644G substitution in the polymerase active site [50] . In this case , replicons were defined as the distance between origins and the inter-origin midpoints , which serve as proxies for termination zones [50] . Mutation frequency was constant across the defined replicons for Pol ε-M644G-expressing MMR-deficient cells , but increased significantly near the inter-origin midpoints in WT Pol ε MMR-deficient controls . The nature of each Pol ε variant may account for these distinct patterns of mutagenesis . Pol ε-M644G retains proofreading activity , but elevates the rates of both misinsertion and mispair extension [47] . Whereas polymerase pausing at mispaired primer termini may trigger the replacement of proofreading-deficient Pol ε with Pol δ , the propensity of Pol ε-M644G to extend mispairs may limit this exchange , ensuring mutagenesis extends to the end of the replicon . The enrichment of mutations near termination zones in MMR-deficient cells with either WT or proofreading-deficient Pol ε suggests that replication fork convergence is an error-prone process , monitored by MMR . The absence of a signal in the Pol ε-M644G data may indicate Pol ε-M644G does not become any more error-prone near termination zones . However , a subtle but significant enrichment of mutations near termination zones may lie hidden within the Pol ε-M644G data—probabilistic differences in the firing of adjacent origins mean that termination zones are not always at the inter-origin midpoint . Consistent with this , when we analyze our data using inter-origin midpoints rather than the defined termination zones , we find only a muted enrichment of mutations near the midpoint ( S6 Fig ) . The ability of mutator cells to assume distinct mutagenic states may have important implications for understanding the remarkable mutation accumulation observed in tumors with Pol ε proofreading defects [7–11 , 51] . POLE tumors are often microsatellite-stable , suggesting that the high mutation burden of POLE cells does not simply result from synergy between proofreading and MMR defects [7 , 51] The POLE cancer alleles are usually heterozygous in the tumor clones [51] . Modeling of the most common allele ( POLE-P286R ) in diploid yeast ( pol2-P301R ) reveals a strong semi-dominant mutator phenotype [52] that contrasts the weak semi-dominant mutator phenotype of pol2-4 ( pol2-D290A , E292A ) [52] . It is conceivable that if the POLE cancer alleles confer volatile mutator phenotypes , the spread between the highest and lowest mutator states may be much larger than we observed with pol2-4 . POLE cells that pass through a hypermutator state would acquire adaptive mutations more readily . Over multiple rounds of selection during tumor evolution this would lead to an extremely rapid accumulation of mutations within the dominant tumor clone . The average mutation frequency in the exomes of POLE tumors ( 235 x 10-6 mutations/bp ) [11] appear to be near the lethal limit for diploids [15 , 53] . Thus , once a POLE tumor clone escapes the restraints on growth , there may be strong selection pressure to limit mutation accumulation , giving an advantage to cells in a hypo-mutator state . Single cell resolution replication studies of human POLE mutant cells are needed to test the hypothesis of mutator volatility in cancer . Understanding the source of volatility may lead to treatments that directly target the mutator phenotype for cancer therapy .
Yeast were grown at 30°C using YPD , synthetic complete ( SC ) media or SC “drop-out” media deficient in defined amino acids to select for prototrophic genetic markers [54] . Premade nutrient supplements for SC and SC lacking uracil ( SC-Ura ) and leucine ( SC-Leu ) were purchased from Bufferad . Other drop-out nutrient supplements were made as described [54] from individual components purchased from Sigma-Aldrich or Fisher Scientific . Canavanine-resistant ( Canr ) mutants for mutation rate assays were selected on SC plates lacking arginine that contained 60 μg/ml canavanine ( Sigma-Aldrich ) . To construct AH2801 , the POL2/URA3::pol2-4 MSH6/msh6Δ::LEU2 diploid used in these experiments , we first deleted one of the two copies of MSH6 in AH0401 [53] to obtain AH0604 . We transformed [55] AH0401 with a LEU2 DNA fragment amplified from pRS415 [56] with the MSH6GU ( TTTAATTGGAGCAACTAGTTAATTTTGACAAAGCCAATTTGAACTCCAAAAGATTGTACTGAG AGTGCAC ) and MSH6GD ( ACTTTAAAAAAAATAAGTAAAAATCTTACATACATCGTAAATGAA AATACCTGTGCGGTATTTCACACCG ) primers and Phusion Polymerase ( New England Biolabs ) [98°C for 1 minute followed by 30 cycles of ( 98°C , 10 sec . ; 54°C , 30 sec . ; 72°C , 90 sec . ) ] . We then integrated the pol2-4 allele into AH0604 using a URA3::pol2-4 chimeric DNA fragment . To generate the URA3::pol2-4 chimeric DNA fragment we first amplified three overlapping DNA fragments . Using the same amplification conditions as above , we amplified URA3 from pRS416-POL2 [21] with the YIF1KIrp ( AGTAAATAGAAAATTTATGACGTAGGAATAAAAGTATATAAGTATTTAACAAATTGGAACAA CACTCAACCCTATCTCGGTCTA ) and YIF1KIfp ( GAAGAGATCAAAGAGAGGATTTAAT TTCATGCGCATTATTATTATCTACGGTCCAGAGCAGATTGTACTGAGAGTGCACCA ) primers , the POL2 promoter from genomic DNA with the pol2-6376 ( GACCGTAGATAATAATAATGCGCATG ) and pol2-S3 ( CTCAGGAGTTTCCTGGCCTCG ) primers , and the pol2-4 fragment from pRS415pol2-4 [21] with the pol2-seq1F ( GGTGGGAGCT TCAAGTCG ) and pol2-8752rp ( CTCCGGTTTCGGTGTATA CTCAAAGTC ) primers . The three fragments were combined in equal-molar ratios and subjected to chimeric PCR [98°C for 1 minute followed by 15 cycles of ( 98°C , 10 sec . ; 72°C , 30 sec . ; 58°C , 20 sec . ; 72°C , 3 min . ) ] using YIFKIrp and pol2-8752rp . The entire POL2 sequence was confirmed by Sanger and Illumina whole genome sequencing . To isolate haploid mutator mother cells , we first sporulated AH2801 by diluting an overnight culture of the strain 1:50 in YPD and growing the cells until they reached a concentration of 1–2 x 107 cells/ml . The cells were recovered by centrifugation , washed with sterile water , re-suspended in sporulation media ( 1% potassium acetate , 0 . 1% yeast extract ( Difco ) , 0 . 05% Dextrose ) at a concentration of 1 . 5–3 x 107 cell/ml , and then grown for four days at 30°C with shaking . For tetrad dissection , 50 μl of sporulated culture were spun down and re-suspended in 1 M sorbitol with 5 μls of Zymolyase 20T ( 25μg/μl ) ( MP Biomedicals ) and then incubated for 10 minutes at 30°C to digest the asci walls . Ice-cold sterile water ( 0 . 8 ml ) was added to suspension and 5 μls were pipetted onto agar plates . For Canr mutation rate measurements , 40 AH2801 tetrads were dissected on SC media . After ~2 . 5 days of growth , individual colonies ( 2–3 x 106 cells ) were scrapped from the plates and re-suspended in 100 μl of sterile water . 90μl was plated on canavanine-selection plates . The remaining suspension was used for 10-fold serial dilutions , which were plated on SC to determine the total number of cells in each colony ( Nt ) as well as on canavanine selection plates to accurately count the number of Canr mutants in pol2-4 msh6Δ colonies . Since we were blind to the genotypes of the spore clones , at the same time we plated cells on SC-Leu to identify those that carried the msh6Δ::LEU2 allele and on SC-Ura to identify cells carrying URA3::pol2-4 . The AH0401 strain from which AH2801 is derived was designed to facilitate mutation rate measurements in diploids and is heterozygous at the CAN1 locus ( CAN1::natMX/can1Δ::HIS3 ) [53] . Thus , we also assessed the ability of AH2801 spore clones to grow on SC-His . Clones unable to grow on SC-His carried the CAN1::natMX allele and were sensitive to canavanine . Mutation counts from these clones were used for mutation rate calculations . After 3 days of growth , colonies on SC and canavanine selection plates were counted and the data grouped according to genotype . Each spore clone was treated as an independent replica culture for fluctuation analyses [27] . Mutation rates were calculated from the mutant counts in each replica culture by estimating the likely number of mutational events ( m ) by maximum likelihood using newtonLDplating in Salvador 2 . 3 with Mathematica 8 . 0 and then divided by the average number of cell divisions inferred from the Nt counts [33 , 34 , 57] . Confidence intervals ( 95% ) were calculated with CILDplating in Salvador 2 . 3 . We estimated the per-base-pair mutation rate from the Canr mutation rate of pol2-4 msh6Δ cells using the approach of Drake [35 , 58 , 59] as previously described [20] . In all lineage experiments , cells were grown during the day and stored overnight at 4°C as described [60] . During all incubation steps the plates were wrapped in parafilm . In the first experiments ( Lineages A , C , F , G1 , G2 , and H ) , tetrads were dissected on non-selective SC media in order to monitor fixation of mutations in WT and msh6Δ control cells as well as pol2-4 msh6Δ mother cells . One tetrad was dissected per plate and two spores were chosen at random for lineage analysis . Genotyping assays for pol2-4 and msh6Δ alleles were performed as described [21] and verified during whole genome sequencing analysis . During the lineage isolation , mother cells were incubated at 30°C and examined every two hours using a Zeiss Axioskop 40 Tetrad Dissection microscope fitted with a 50μm fiber optic needle . After each cell division , the mother and daughter cells were manually separated using the micromanipulator as previously described [54] . The daughter cells , distinguished by a smaller diameter , were moved to a defined coordinate on the agar plate . In later experiments ( Lineages B , D , and E ) , we focused solely on double mutant spores by dissecting tetrads on media lacking leucine ( to select for msh6Δ::LEU2 ) and uracil ( to select for URA3::pol2-4 ) . Several tetrads were dissected per plate and a single double mutant spore per plate was selected for analysis by its ability to divide in the absence of leucine and uracil . As before , all daughter cells were moved to a defined coordinate on the agar surface . In addition , the first granddaughter cell born to each daughter cell was also moved to a defined coordinate to serve as a back-up in case the daughter cell died . Dissection continued until double mutant mother cells ceased dividing , whereupon all daughter and granddaughter clones were allowed to form colonies . For sequencing , the entire daughter colony was used to inoculate overnight YPD cultures . Glycerol stocks of each daughter culture were archived and genomic DNA was purified with a ZR Fungal/bacterial purification kit ( Zymo Research ) . The purified DNA was simultaneously fragmented and ligated to Illumina DNA adapters using the Nextera V2 Kit ( Illumina ) , post-indexed by PCR ( primer sequences available upon request ) , and sequenced using 101 bp , paired-end reads on an Illumina 2500 platform . Once all members of a lineage had been sequenced , we used the Burrows-Wheeler Aligner ( v0 . 6 . 2 ) [61] to align the reads against a copy of the S288C S . cerevisiae genome ( Assembly R64-1-1 ) in which low complexity and highly repetitive sequences have been removed ( <0 . 5% of the genome ) with RepeatMasker . After the initial alignment , unmapped and ambiguously mapped reads were filtered out . PCR duplicates were evaluated using the MarkDuplicates option in the Picard suite of programs ( http://picard . sourceforge . net ) . To further reduce false variant calls , the Genome Analysis Tool Kit ( GATK ) suite of programs was used for local realignment and base quality score recalibration [62] . We used VarScan2 [63] for variant calling with a minimum read depth of 15 , a minimum variant frequency of 0 . 8 , and a quality score of 15 . We then filtered the resulting variants to remove strain-specific single nucleotide polymorphisms segregating in our genetic background using a database of putative SNPs segregating in the BY4743 strain background that we previously compiled [53] . We normalized coverage to ensure that we only scored mutations in sequences shared by all members of a lineage . Finally , to determine maternal and daughter specific variants , we compared the mutations found in the last daughter to those found in all preceding daughters . Variants found in the last daughter and shared by one or more of the preceding daughters were designated maternally fixed mutations ( See S1 Dataset ) . Those not found in the maternal lineage were designated daughter-specific mutations and were not evaluated further for this study . Called mutations were then visually confirmed using the Integrated Genome Browser . Models of expected mutation count data from different Poisson distributions were calculated in Excel . We also used the finite mixture modeling ( FMM ) procedure of the SAS system ( v . 9 . 4 ) to compute and compare single-distribution and distribution-mixture models of the mutation count data . A Poisson response distribution and log link function were specified , and parameter estimation was by maximum likelihood . To investigate whether the data could be described better by a negative binomial distribution than by a mixture of two Poisson distributions , we compared the fit of a single negative binomial distribution against the fit of the two-Poisson mixture . Because this comparison cannot be treated as a nested set of hypotheses , we used information theory to characterize the fit , choosing Akaike’s Information Criterion ( AIC ) for assessment . We again used the FMM procedure of the SAS system ( v . 9 . 4 ) , specifying a Poisson or negative binomial response distribution , and a log link function , for model fitting . Quantification of relative likelihood was after Burnham and Anderson ( 2002 ) . We used 10 , 000 iterations of a simulation to determine the likelihood that multiple mutations would occur on the same chromosome . To simulate the random distribution of 237 mutations over 87 cell divisions , we generated a sequence file of 87 yeast genomes—each chromosome with a unique identifier . In each iteration of the simulation we randomly selected 237 bases within these 87 yeast genomes and then counted the frequency at which the same chromosome was “mutated” two or three times . From the resulting histogram , we determined the 95% confidence intervals . The locations of all mutations were mapped onto the replication profiles from Raghuraman et al . [39] . The positions of all putative replication origins and termination zones in the replication profiles were identified by noting chromosomal positions where trep values were at a local maxima or minima . Segments between adjacent maxima and minima were used to define probabilistic replicons to which all mutations were then assigned . The fractional distance of a mutation between any origin/termination pair was calculated by dividing the distance of a mutation to the closest origin by the total distance between the origin and termination zone . The resulting fractional distances were then grouped into bins corresponding to fractional distances of 0 . 1 . To test the significance of the observed distribution , we assumed that random mutagenesis would produce bins of equal size ( 23 . 7 mutations/bin ) . We then compared the observed and expected distributions using a Chi-Squared test . We also performed a similar analysis using a different way of defining replicons outlined by Lujan et al . [50] . In this approach , a replicon is defined as the DNA segment between a confirmed origin ( see OriDB , [64] ) and halfway to the adjacent confirmed origin ( inter-origin midpoint ) . After assigning mutations to these DNA segments , we then determined the fractional distance of each mutation to the closest origin by dividing the distance of the mutation to the origin by the total distance between the origin and the inter-origin mid-point . As above , we binned mutations by their fractional distance to the origin and determined the significance of the distribution using a Chi-Squared test . | Mutations fuel microbial evolution and cancer . Cells with an increased rate of mutation are said to have a “mutator phenotype” and adapt more rapidly than non-mutator cells . Our study utilizes a novel way of measuring mutation rates of individual cell divisions to show that mutator cells can adopt one of two mutation rates that differ tenfold in magnitude . This mutator volatility suggests that the rates of mutation accumulation may vary widely within the same clone of mutator cells . Understanding how to modulate the mutator state may provide an avenue to treat certain cancers . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Volatility of Mutator Phenotypes at Single Cell Resolution |
Identifying driver mutations in cancer is notoriously difficult . To date , recurrence of a mutation in patients remains one of the most reliable markers of mutation driver status . However , some mutations are more likely to occur than others due to differences in background mutation rates arising from various forms of infidelity of DNA replication and repair machinery , endogenous , and exogenous mutagens . We calculated nucleotide and codon mutability to study the contribution of background processes in shaping the observed mutational spectrum in cancer . We developed and tested probabilistic pan-cancer and cancer-specific models that adjust the number of mutation recurrences in patients by background mutability in order to find mutations which may be under selection in cancer . We showed that mutations with higher mutability values had higher observed recurrence frequency , especially in tumor suppressor genes . This trend was prominent for nonsense and silent mutations or mutations with neutral functional impact . In oncogenes , however , highly recurring mutations were characterized by relatively low mutability , resulting in an inversed U-shaped trend . Mutations not yet observed in any tumor had relatively low mutability values , indicating that background mutability might limit mutation occurrence . We compiled a dataset of missense mutations from 58 genes with experimentally validated functional and transforming impacts from various studies . We found that mutability of driver mutations was lower than that of passengers and consequently adjusting mutation recurrence frequency by mutability significantly improved ranking of mutations and driver mutation prediction . Even though no training on existing data was involved , our approach performed similarly or better to the state-of-the-art methods .
Cancer is driven by changes at the nucleotide , gene , chromatin , and cellular levels . Somatic cells may rapidly acquire mutations , one or two orders of magnitude faster than germline cells [1] . The majority of these mutations are largely neutral ( passenger mutations ) in comparison to a few driver mutations that give cells the selective advantage leading to their proliferation [2] . Such a binary driver-passenger model can be adjusted by taking into account additive pleiotropic effect of mutations [3 , 4] . Mutations might have different functional consequences in various cancer types and patients , they can lead to activation or deactivation of proteins and dysregulation of a variety of cellular processes . This gives rise to high mutational , biochemical , and histological intra- and inter-tumor heterogeneity that may explain the resistance to therapies and complicates the identification of driving events in cancer [5 , 6] . Point DNA mutations can arise from various forms of infidelity of DNA replication and repair machinery , endogenous , and exogenous mutagens [6–9] . There is an interplay between processes leading to DNA damage and those maintaining genome integrity . The resulting mutation rate can vary throughout the genome by more than two orders of magnitude [10 , 11] due to many factors operating on local and global scales [12–14] . Many studies support point mutation rate dependence on the local DNA sequence context for various types of germline and somatic mutations [9 , 11 , 13 , 15] . For both germline and somatic mutations , local DNA sequence context has been identified as a dominant factor explaining the largest proportion of mutation rate variation [10 , 16] . Additionally , differences in mutational burden between cancer types suggest tissue type and mutagen exposure as important confounding factors contributing to tumor heterogeneity . Assessing background mutation rate is crucial for identifying significantly mutated genes [17 , 18] , sub-gene regions [19 , 20] , mutational hotspots [21 , 22] , or prioritizing mutations [23] . This is especially important considering that the functional impact of the majority of changes observed in cancer is poorly understood , in particular for rarely mutated genes [24] . Despite this need , there is a persistent lack of quantitative information on per-nucleotide and per-codon background rates in various cancer types and tissues . There are many computational methods that aim to detect driver genes and fewer methods trying to rank mutations with respect to their potential carcinogenicity . As many new approaches to address this issue have been developed [25] [26] , it still remains an extremely difficult task . As a consequence , many driver mutations , especially in oncogenes , are not annotated as high impact or disease related [27] even though cancer mutations harbor the largest proportion of harmful variants [28] . In this study we utilize probabilistic models that estimate background mutability per nucleotide or codon substitution to rank mutations and help distinguish driver from passenger mutations . The mutability concept has been used in many evolutionary and cancer studies ( although it has been estimated in different ways ) and is defined as a probability to obtain a nucleotide or codon substitution based on the underlying background processes of mutagenesis and repair that are devoid of cancer selection component affecting a specific genomic ( or protein ) site . The mutability can be calculated using background models ( mutational profiles ) , mutational signatures or mutations motifs that are constructed under the assumption that vast majority of cancer context-dependent mutations have neutral effects , while only a small number of these mutations in specific sites are under positive or negative selection . To assure this , we removed all recurrent mutations as these mutations might be under selection in cancer . Mutational profiles are calculated by sampling the frequency data on types of mutations and their trinucleotide ( for nucleotide mutations ) and pentanucleotide ( for codon substitutions ) contexts regardless of their genomic locations . These models can be used to estimate the expected mutation rate in a given genomic site as a result of different local or long-range context-dependent mutational processes . In this paper we try to decipher the contribution of background DNA mutability in the observed mutational spectrum in cancer for missense , nonsense , and silent mutations . We compiled a set of cancer driver and neutral missense mutations with experimentally validated impacts collected from multiple studies and used this set to verify our approach and compare it with other existing methods . Our approach has been implemented online as part of the MutaGene web-server and as a stand-alone Python package: https://www . ncbi . nlm . nih . gov/research/mutagene/gene .
We analyzed all theoretically possible codon substitutions that could have occurred by single point mutations in 520 cancer census genes and calculated their mutability values based on their genomic context . We found that only about one percent of all theoretically possible codon substitutions were observed in the surveyed 12 , 013 tumor samples derived from the COSMIC v85 cohort ( S1 Table ) . Using the pan-cancer model , across all analyzed possible codon substitutions produced by single point mutation , mutability ranged from 1 . 61 x 10−7 to 1 . 80 x 10−5 ( mean = 1 . 34 x 10−6 ) . Lower and upper boundaries for mutability are dependent on the cancer model selection , and cancer models with higher mutational burdens like melanomas ( 1 . 92 x 10−7 to 1 . 35 x 10−4 , mean = 7 . 00 x 10−6 ) have higher mutability values compared to cancers such as prostate adenocarcinoma ( 5 . 12 x 10−8 to 7 . 31 x 10−6 , mean = 3 . 95 x 10−7 ) . We found that across codon substitutions which were not observed in the COSMIC v85 cohort , the mean mutability ( 1 . 29 x 10−6 ) was found to be three-fold lower compared to the mutability of observed codon substitutions ( 3 . 88 x 10−6 ) using pan-cancer background model , Mann-Whitney-Wilcoxon test p < 0 . 01 ( Fig 1A ) . This finding also holds true for different cancer-specific models ( the list of cancer-specific mutational profiles can be found in https://www . ncbi . nlm . nih . gov/research/mutagene/signatures#mutational_profiles ) . The same result is confirmed for per-nucleotide mutability ( 1 . 04 x 10−6 versus 3 . 36 x 10−6 , Mann-Whitney-Wilcoxon test p < 0 . 01 ) . In addition , we validated our result on a set of observed mutations from 9 , 228 patients who had undergone prospective sequencing of MSK-IMPACT gene panel . Looking at mutations in the genes which were sequenced in all patients in the MSK-IMPACT cohort , the same pattern remains that observed codon substitutions had a higher mutability ( 3 . 41 x 10−6 ) , compared to those which were theoretically possible , but did not occur in cancer patients ( 1 . 30 x 10−6 , Mann-Whitney-Wilcoxon test , p < 0 . 01 ) ( Fig 1B ) . S1 Fig shows cumulative and probability density distributions of nucleotide mutability values for all observed mutations in patients , for theoretically possible mutations in all cancer census genes and for two genes in particular , CASP8 and TP53 . While there are many theoretically possible mutations with low mutability values , the observed cancer spectrum is dominated by mutations with high mutability . A similar pattern is seen for cancer-specific cases ( Fig 2 ) . Fig 3A and 3B show the distributions of codon mutability values for all possible missense , nonsense , and silent mutations accessible by single nucleotide base substitutions in the protein-coding sequences of 520 cancer census genes calculated with the pan-cancer background model . Codon mutability spans two orders of magnitude , and silent mutations have significantly higher average mutability values ( mean = 5 . 68 x 10−6 ) than nonsense ( mean = 3 . 44 x 10−6 ) or missense mutations ( mean = 3 . 29 x 10−6 ) ( Kruskal-Wallis test p < 0 . 01 and Dunn’s post hoc test p < 0 . 01 for all comparisons ) . These differences in codon mutabilities could be a reflection of the degeneracy of genetic code , where multiple silent nucleotide substitutions in the same codon may increase its mutability . However , degeneracy of genetic code should not affect the calculation of nucleotide mutability . While the differences between types of mutations are less pronounced for nucleotide mutability ( Fig 3C ) , silent mutations are still characterized by the highest nucleotide mutability values ( mean = 3 . 91 x 10−6 for silent , 3 . 10 x 10−6 for nonsense and 3 . 17 x 10−6 for missense mutations , Kruskal-Wallis test p < 0 . 01 and Dunn’s post hoc test p < 0 . 01 for all comparisons ) . Under the null model of all mutations arising as a result of neutral background mutational processes , somatic mutations should accumulate with respect to their mutation rate and one would expect a positive correlation between mutability and observed mutational frequency of individual mutations . As Fig 3B and 3D show , there is a trend for silent and nonsense mutations . To further investigate this relationship , in the pan-cancer COSMIC v85 cohort we calculated both Spearman’s rank , a non-parametric test taking into account that mutability is not normally distributed , and Pearson linear correlation coefficients between codon mutability and frequencies of mutations across all 520 cancer census genes . We also explored this association for each gene with at least ten unique mutations of each type: silent , nonsense , and missense ( Fig 4 ) . Overall , we found 84 and 137 genes with significant ( p < 0 . 01 ) positive Spearman and Pearson correlations , respectively , for at least one mutation type ( S2 Table ) . Among the genes with significant correlations , 41 belong to tumor suppressor genes , 28 are oncogenes , and 15 genes are classified as either fusion genes or both oncogene and tumor suppressor . For some genes , including TP53 ( first column , Fig 4E ) and tumor suppressor CASP8 ( second column , Fig 4E ) , a rather strong linear relationship between mutability and recurrence frequency of observed mutations ( R2 > 0 . 5 ) was observed . Breaking up all codon changes into silent , nonsense and missense reveals the highest correlations for silent ( ρ = 0 . 15 , r = 0 . 1 , p < 0 . 01 ) and nonsense ( ρ = 0 . 20 , r = 0 . 15 , p < 0 . 01 ) mutations ( S2 Fig ) . The effects of mutations on protein function , with respect to their cancer transforming ability , can drastically differ in tumor suppressor genes ( TSG ) and oncogenes , therefore we performed our analysis separately for these two categories ( Fig 5 ) . In general , mutations in TSG can cause cancer through the inactivation of their products , whereas mutations in oncogenes may result in protein activation . We used COSMIC gene classification separating genes into tumor suppressors and oncogenes . Genes which were annotated as both TSG and oncogenes were excluded from this analysis . Gene ontology ( GO ) analysis found that top GO annotations in TSG for cellular compartments were “nucleus” , “chromosome” , and “nuclear part” and for molecular functions were “protein” , “DNA” , and “enzyme binding” . For the oncogenes , the top GO annotations for cellular components were “nucleoplasm” , “nucleus” , and “nuclear lumen” and for molecular function “heterocyclic compound binding” , “organic cyclic compound biding” and “sequence-specific DNA binding” . A full list of genes and the associated GO terms is available in Supplemental S3 Table . In addition , we used COSMIC classification into genes with dominant or recessive mutations , but overall results were similar to the ones produced using classification into TSG and oncogenes ( S3 Fig ) . We observed a weak but statistically significant correlation between codon mutability and recurrence frequency in TSG ( ρ = 0 . 17 , r = 0 . 13 , p < 0 . 01 ) while oncogenes showed a weaker Spearman correlation and no significant Pearson correlation ( ρ = 0 . 13 , p < 0 . 01; r = 0 , p = 0 . 61 ) ( S2B and S2C Fig ) . This correlation mostly arises from neutral mutations as shown in the following section . An inverse U-shaped trend was detected for missense and silent mutations in oncogenes: highly recurrent mutations ( observed in three and more samples ) were characterized by low average mutability values ( Fig 5 ) . In the latter case , selection may be a more important factor compared to background mutation rate explaining reoccurrence of these mutations . Functionally conserved sites overall were found to be more frequently mutated in oncogenes [29] , and our analysis did not find a straightforward association between mutability and evolutionary conservation . We complied a combined dataset of experimentally annotated missense mutations in cancer genes from several sources . Mutations were categorized as ‘non-neutral’ or ‘neutral’ based on their experimental effects on protein function , transforming effects , and other characteristics ( see Methods and S4 Table ) . For all mutations in combined dataset , whether they were observed in MSK-IMPACT or the COSMIC v85 cohorts , the codon mutability values of neutral mutations were significantly higher ( mean = 2 . 71 x 10−6 ) ( Mann-Whitney-Wilcoxon test , p < 0 . 01 ) than for non-neutral mutations ( mean = 1 . 74 x 10−6 ) ( Fig 6A ) . Binning the mutations by their reoccurrence frequency also showed differences between ‘neutral’ and ‘non-neutral’ , with the frequency of neutral mutation depending on their mutability . For neutral mutations , mutations that were observed in three or more samples had higher background mutability ( meanMSK = 6 . 39 x 10−6 , meanCOSMIC = 6 . 22 x 10−6 ) compared to mutations which were not observed ( meanMSK = 2 . 46 x 10−6 , meanCOSMIC = 2 . 54 x 10−6 ) . In contrast , the background mutability of non-neutral mutations did not vary with the reoccurrence frequency ( Fig 6B ) , suggesting that background mutability was much less important in driving reoccurrence of non-neutral mutations . In the previous sections we explored the contribution of background mutational processes in understanding the observed mutational patterns in cancer . With our finding that background mutability differs between neutral mutations and non-neutral mutations , we explored if background mutability could be used to facilitate the detection of cancer driver mutations or provide a reasonable ranking in terms of their potential carcinogenic effects . We tested different ways to calculate codon mutability and if it could help to differentiate between experimentally annotated neutral , or putatively passenger mutations , and non-neutral driver mutations . We found that a simple and intuitive measure , B-score , calculated ( see next section ) performed the best on the combined experimental test set . A similar measure was used previously to identify mutational hot spots [21 , 30] . Hotspots are defined for sites , whereas our approach assesses specific mutations , and different mutations from the same hotspot can be drivers or passengers . For instance , TP53 Tyr236 site is annotated as a hotspot in [21 , 30] , however p . Tyr236Phe mutation in this site is experimentally characterized as neutral in the IARC database . We compared the performance of B-score to six state-of-the-art computational methods which distinguish driver from passenger mutations in cancer: CHASM [31] , CHASMplus [32] , VEST[33] , REVEL[34] , CanDrAplus[35]and FatHMM[36] . Table 1 shows the performance of the various computational predictors at classifying mutations from the combined dataset observed in two sets of cancer cohorts . To compare across methods , which use different thresholds for calling neutral versus non-neutral mutations , we calculated the Matthew’s correlation coefficient ( MCC ) across a range of thresholds for each method and reported the maximal MCC value . Based on the MCC , the best classifiers are CHASMplus , B-score and CanDrAplus ( MCC = 0 . 64 , 0 . 61 , and 0 . 58 respectively ) ( Table 1 ) . Surprisingly , mutation reoccurrence frequency alone performs very well , with MCC of 0 . 49 in the COSMIC v85 cohort and 0 . 51 in the MSK-Impact cohort . B-Score is able to provide a correction to reoccurrence frequency using codon mutability and yields a much better performance than frequency alone . Intriguingly , inverse mutability alone performs better than random , emphasizing the fundamental quality of non-neutral mutations in cancer: mutability of driver mutations is lower than the mutability of passengers ( Fig 6 ) . We also explored the performance of methods in classifying mutations that were not observed or observed only once in the COSMIC v85 cohort or MSK-Impact cohort ( S6 Table ) . For mutations which were not observed in the COSMIC v85 cohort B-Score classification performance is low but better than random ( AUC = 0 . 65 ) . On mutations which were observed in only one cancer sample in the cohort ( 207 passenger and 157 driver mutations ) , B-Score still performed better than VEST and CHASM ( MCC = 0 . 46 , 0 . 42 , and 0 . 36 respectively ) . On the combined set which includes all experimentally verified mutations , whether they were observed or not observed in cancer patients , B-score ranks fourth after CHASMplus , REVEL and FatHMM ( S7 Table ) . B-score also allows to break ties for mutations observed in the same number of patients . For example in the TP53 gene , mutations p . Glu11Lys and p . Cys135Gly have been observed in two patients each in the COSMIC v85 cohort . However , p . Glu11Lys ( mutability of 1 . 18 x 10−5 ) is predicted a passenger mutation and p . Cys135Gly ( mutability of 2 . 20 x 10−7 ) is predicted as a driver mutation which is consistent with the annotations from the experimental combined dataset . Even though our probabilistic model indirectly incorporates different factors affecting mutation rate , we checked explicitly if large-scale factors , allowing mutations of the same type to have different mutational probabilities in different genes , affected retrieval performance on the combined test set . Several methods have been developed to estimate gene weights ( see Methods ) , which consider the overall number of mutations or the number of silent mutations affecting a gene . Additionally , we estimated the gene weights based on the number of SNPs in the vicinity of a gene . We also examined the effects of several large-scale confounding factors such as gene expression levels , replication timing , and chromatin accessibility ( provided in the gene covariates in MutSigCV [37] ) on gene weights . We used gene weights to adjust mutability values and explored whether any of the gene weight models were helpful in distinguishing between experimentally determined neutral and non-neutral mutations . We found that “no-outlier”-based weight ( r = 0 . 66 , p = 0 . 004 ) and “silent mutation”-based weight ( r = 0 . 65 , p = 0 . 004 ) significantly correlated with the gene expression levels . No other correlations of gene weights with confounding factors were found . Overall , using gene weight as an adjustment for varying background mutational rates across genes did not improve classification performance of mutations in the experimental benchmark . Only a SNP-based weight affected the AUC-ROC , but the gain was minimal , and no gene weight affected the MCC ( S8 Table ) . It is consistent with the previous studies that found local DNA sequence context as a dominant factor explaining the largest proportion of mutation rate variation [10 , 16] . MutaGene webserver provides a collection of cancer-specific context-dependent mutational profiles [38] . It allows to calculate nucleotide and codon mutability and B-Score for missense , nonsense and silent mutations for any given protein coding DNA sequence and background mutagenesis model using the “Analyze gene” option . Following the analysis presented in this study , we added options to provide a ranking of mutations observed in cancer samples based on the B-Score or the multiple-testing adjusted q-values . Using the combined dataset as a performance benchmark ( Table 1 , S7 Table ) , we calibrated two thresholds: the first corresponds to the maximum of MCC , and the second corresponds to 10% FPR . Mutations with the B-Score below the first threshold are predicted to be “cancer drivers” , whereas mutations with scores in between two thresholds are predicted to be “potential drivers” . All mutations with scores above the second threshold are predicted as “passengers” . Importantly , calculations are not limited to pan-cancer and can be performed using a mutational profile for any particular cancer type , the latter would result in a cancer-specific ranking of mutations and could be useful for identification of driver mutations in a particular type of cancer . An example of prediction of driver mutations status for EGFR is shown in Fig 7 . MutaGene Python package allows to rank mutations in a given sample or cohort in a batch mode using pre-calculated or user-provided mutational profiles or signatures and is available at https://www . ncbi . nlm . nih . gov/research/mutagene/gene .
To understand what processes drive point mutation accumulation in cancer , we used DNA context-dependent probabilistic models to estimate the baseline mutability for nucleotide mutation or codon substitution in specific genomic sites . Passenger mutations , constituting the majority of all observed mutations , may have largely neutral functional impacts and are unlikely to be under selection pressure . For passenger mutations one would expect that mutations with lower DNA mutability would have lower observed mutational frequency and vice versa . In a recent study the fraction of sites harboring SNPs in the human genome was indeed found to correlate very well with the mutability although the later was estimated differently from our study [39] . We detected a significant positive correlation between background mutability , which is proportional to per-site neutral mutation rate , and observed reoccurrence frequencies of mutations in cancer patients . In accordance with this trend , we also found that mutations that were not observed in cancer cohorts were marked by a lower background mutability . For some genes , such as TP53 or CASP8 , mutations and their frequencies can be predicted from their mutability values . Outliers of this association trend , mutations which reoccur at high frequencies but have low mutabilities might be important for inferring mutations under positive selection , as illustrated especially for missense mutations in oncogenes . In this respect , reoccurring synonymous mutations with low mutability may represent interesting cases for further investigation of potential synonymous drivers . Mutability of synonymous mutations was found to be the highest among other types of mutations . Observed mutational frequency of synonymous mutations scales with their mutability , therefore it is important to correct for mutability while ranking these mutations with respect to their driver status . Overall , B-score predicted 102 synonymous driver mutations in 64 out of 520 cancer-associated genes . It has been previously shown that some synonymous mutations might be under positive selection and can affect the speed and accuracy of transcription and translation , protein folding rate , and splicing [40] . Some recurrent highly mutable synonymous mutations might not represent relevant candidates of drivers , whereas some rare mutations with relatively low mutability are predicted to be drivers by our approach ( e . g . KDR gene p . Leu355 = , NTRK1 gene p . Asn270 = ) . In this paper we developed and tested a probabilistic model , implemented as B-Score , to adjust the reoccurrence frequency of a mutation ( a measure commonly used in clinical research to identify genes and mutations under selection ) by its expected background mutability . B-Score is able to provide a correction to reoccurrence frequency using mutability and improves the classification of cancer driver and passenger mutations by up to 20% compared to reoccurrence frequency alone . The advantages of B-score are that: ( i ) it is intuitive and interpretable , ( ii ) does not rely on many parameters , and ( iii ) does not involve explicit training on driver and passenger mutation sets . One of the disadvantages is that it requires the knowledge of a total number of patients tested . We found that B-Score performed comparably or better to many of state-of-the-art methods even for rare mutations observed in two large cancer cohorts . However , it underperformed for those mutations from combined experimental set that were not observed in cancer patients . These latter mutations might either constitute functionally disruptive mutations not directly connected with the carcinogenesis or might represent rare cancer mutations not yet detected in large cancer cohorts . A lot of efforts have been focused so far on developing a comprehensive set of cancer driver mutations verified at the levels of functional assays or animal models [26 , 41 , 42] . However , existing sets often contain predictions and very few neutral cancer passenger mutations . The vast majority of computational prediction methods rely on machine learning algorithms trained on mutations from a few genes and/or on recurrent mutations as estimates of driver events or use germline SNPs or silent mutations as the presumed “neutral” set . In many cases , the performance is evaluated on similarly generated synthetic benchmarks . As a result , methods can be trained on incorrectly labeled data and even if trained on correct data , can exhibit a well-known overfitting effect . While mutational processes vary widely among cancer types , and different driver mutations have been shown to be preferentially associated with specific mutational processes [39 , 40] , there remains a lack of cancer-specific driver/passenger datasets . In our combined dataset , the effects of mutations were determined using experimental assays , which were not linked to any particular cancer type , therefore a pan-cancer model was used for calculation of B-score and other methods tested . However , we provide the ability to apply cancer-specific B-Score ranking of mutations using the models available via the MutaGene package and website ( see Methods ) . Additionally , for some cancer types , the background mutational processes may differ greatly between subsets of cancer patients . For these highly heterogenous cancer types rather than using cancer type specific , it may be more appropriate to use background mutational profiles/models specific for a given cohort . In this study , we restricted our test dataset to only missense mutations that have been experimentally assessed , with several thousands of driver and passenger mutations from 58 genes . Intriguingly , we found that experimentally annotated driver mutations had a lower background mutability than neutral mutations , suggesting possible action of context-dependent codon bias towards less mutable codons at critical sites for these genes , although more studies would have to be conducted to further investigate this observation . This important difference in mutability between drivers and passengers may explain the outstanding performance of the simple measure B-score which enables an understanding of the differential roles that background mutation rate and selection play in shaping the cancer mutational spectrum .
We assembled a combined dataset that included mutations from the five datasets described below . First we obtained missense mutations for TP53 gene with experimentally determined functional transactivation activities from IARC P53 database where they were classified as functional , partially-functional , and non-functional[43] . The second dataset contained experimental evidence collected from the literature[44] . The experimental evidence of impact of mutations included changes in enzymatic activity , response to ligand binding , impacts on downstream pathways , an ability to transform human or murine cells , tumor induction in vivo , or changes in the rates of progression-free or overall survival in pre-clinical models . Mutations were considered “damaging” if there was literature evidence to support their impact on at least one of the above-mentioned categories . Mutations with no significant impacts on the wild-type protein function were classified as “neutral” . Mutations with no reliable functional evidence were regarded as “uncertain” and were not used in this study . The third dataset included experimentally verified BRCA1 mutations and was originally collected by using deep mutational scanning to measure the effects of missense mutations on the ability of BRCA1 to participate in homology-directed repair . In this dataset missense mutations were categorized as either “neutral” or “damaging” [45 , 46] . Noteworthy , BRCA1 set contained inherited germline as well as somatic mutations . The fourth dataset explored over 81 , 000 tumors to identify drivers of hypermutation in DNA polymerase epsilon and polymerase delta genes ( POLE/POLD1 ) . “Drivers of hypermutation” were variants which occurred in a minimum of two hypermutant tumors , which were never found in lowly mutated tumors , and did not co-occur with an existing known driver mutation in the same tumor . Other variants in these genes were considered “passengers” with respect to hypermutation[25] . The fifth dataset consisted of missense mutations annotated based on their effects on cell-viability in Ba/FC and MCF10A models[47] . “Activating mutations” were mutations where the cell viability was higher than the wild-type gene , and “neutral mutations” were those mutations where cell-viability was similar to the wild-type . Ng et al . used these consensus functional annotations to compare the performance of 21 different computational tools in classifying between activating and neutral mutations using ROC analysis , with activating mutations acting as the positive set and neutral as the negative set . The authors found that the tools yielding best performance were CanDrAplus and CHASM . We included 743 mutations ( 488 neutral and 255 activating ) in 50 genes accessible through single nucleotide substitutions out of the 816 activating and neutral mutations that Ng et al tested [47] . Finally , we removed redundant and conflicting entries when mutations annotated as non-functional or neutral in one dataset were also annotated as damaging or benign in another . As a result , all mutations in the combined data set were categorized as “non-neutral” ( affecting function , binding or transforming ) and “neutral” ( other mutations ) . We treated “functional” and “partially -functional” mutations in IARC TP53 dataset as “neutral” , and “non-functional” as “non-neutral” . Overall , the combined dataset contains 5 , 276 mutations ( 4 , 137 neutral and 1 , 139 non-neutral ) from 58 genes ( S4 Table ) and is available on MutaGene website at https://www . ncbi . nlm . nih . gov/research/mutagene/benchmark . The Catalogue of Somatic Mutations in Cancer ( COSMIC ) database stores data on somatic cancer mutations and integrates the experimental data from the full-genome ( WGS ) and whole-exome ( WES ) sequencing studies [48] . Cancer census genes ( 520 genes ) were defined according to COSMIC release v84 . For each of these genes , we explored all theoretically possible nucleotide mutations along the DNA sequence of the principal transcripts . This resulted in 4 , 129 , 461 possible nucleotide substitutions , and 3 , 293 , 538 codon substitutions . For analyses comparing oncogenes and tumor suppressor genes ( TSG ) , genes classified as only fusion genes or those with both oncogenic and TSG activities were not used . This resulted in 205 oncogenes and 167 TSG ( S3 Table ) . For gene ontology ( GO ) enrichment analysis we used the R package “GOfuncR” . For enrichment analysis , the genes annotated as either TSG or oncogenes were compared to all other genes in the “Homo . sapiens” gene annotation package in R . 98% of all COSMIC v85 samples contained less than 1000 mutations so were not hypermutated . COSMIC v85 samples which came from cell-lines , xenografts , or organoid cultures were excluded . Only mutations with somatic status of “Confirmed somatic variant” were included and mutations which were flagged as SNPs were excluded . For each cancer patient , a single sample from a single tumor was used . Additionally , it is possible that the same patient may be assigned different unique identifiers in different papers , and duplicate tumor samples are sometimes erroneously added to COSMIC database during manual curation . These samples may affect the recurrence counts of mutations . We applied clustering method in order to detect and remove any redundant tumor samples . Each sample was represented as a binary vector with 1 if a sample had a mutation in a particular genomic location and 0 otherwise . The binary vectors were compared with Jaccard distance metric , J=|A∪B|−|A∩B||A∪B| , where identical samples have J = 0 , followed by agglomerative clustering with complete linkage . Non-singleton clusters with pairwise distance cutoff of J ≥ 0 . 3 were extracted and only one representative of each cluster was used , whereas other samples were discarded . Because of these relatively stringent criteria for inclusion , it is likely that some small number of non-duplicate samples were discarded in this process . MSK-IMPACT cohort was obtained from cBioPortal [49] . We ensured that no mutations were counted multiple times for each patient; if there were multiple tumor samples per patient , primary and metastatic , the primary tumor was kept , and the metastatic discarded . Only those tumors which were sequenced against a matched normal sample were kept to ensure validity of somatic mutations . In the 520 genes we explored , we investigated if these genes were expressed in cancer cell lines from multiple tissue types using RNAseq data from the January 2019 release of the Cancer Cell Line Encyclopedia [50] . Using RNAseq data of 1 , 019 unique cancer cell lines from 26 different tissue types and a cutoff for expression at 0 . 5 RPKM ( Reads Per Kilobase of transcript , per Million mapped reads ) , we found that 512 genes were expressed in at least one tissue . Context-dependent mutational profiles were constructed previously from the pools of mutations from different cancer samples by counting mutations observed in a specific trinucleotide context [38] . Altogether , there are 64 different types of trinucleotides and three types of mutations x≫y ( for example C≫A , C≫T , C≫G and so on ) in the central position of each trinucleotide which results in 192 trinucleotide context-dependent mutation types . In a mutated double-stranded DNA both complementary purine-pyrimidine nucleotides are substituted , and therefore we considered only substitutions in pyrimidine ( C or T ) bases , resulting in 96 possible context-dependent mutation types m = a[x ≫ y]b , where a , b , x , y ∈ {A , T , C , G} , x ≠ y . Thus , mutational profile can be expressed as a vector of a number of mutations of certain type ( f1 , … , f96 ) or a number of mutations of certain type per sample ( r1 , … , r96 ) . Profiles were constructed under the assumption that vast majority of cancer context-dependent mutations have neutral effects , while only a negligible number of these mutations in specific sites are under selection . To assure this , we removed recurrent mutations ( observed twice or more times in the same site ) as these mutations might be under selection in cancer . In the current study we used pan-cancer and cancer-specific mutational profiles for breast , lung adenocarcinoma , colon adenocarcinoma , and skin melanoma derived from MutaGene [38] . We calculated mutability that described baseline DNA mutagenesis per nucleotide or per codon . Mutability was defined as a probability to obtain a context-dependent nucleotide mutation purely from the baseline mutagenic processes operating in a given group of samples . Mutability is proportional to the expected mutation rate of a certain type of context-dependent mutation regardless of the genomic site it occurs . For exome mutations , given the number of different trinucleotides of type t in a diploid human exome , nt , the nucleotide mutability is calculated as: pmnuc=rmnt ( 1 ) In protein-coding sequences it is practical to calculate mutation probability for a codon in its local pentanucleotide context , given trinucleotide contexts of each nucleotide in the codon . For a given transcript of a protein , at exon boundaries the local context of the nucleotides was taken from the genomic context . The COSMIC consensus transcript was chosen for the transcript for each protein . Changes in codons can lead to amino acid substitutions , synonymous and nonsense mutations . Therefore , codon mutability was calculated as the probability to observe a specific type of codon change which can be realized by single nucleotide mutations at each codon position i as: pMcodon=1−∏i3 ( 1−∑jkpijnuc ) ( 2 ) Where k denotes a number of mutually exclusive mutations at codon position i . For example , for Phe codon “TTT” in a given context 5’-A and G-3’ three single nucleotide mutations can lead to Phe→Leu substitution ( to codons “TTG” , “TTA” and “CTT” for Leu ) : A[T≫C]TTG in the first codon position or mutually exclusive ATT[T≫G]G and ATT[T>>A]G in the third codon position . In this case the probability of Phe→Leu substitution in the ATTTG context can be calculated as pPhe→Leucodon=1− ( 1−pA[T≫C]T ) ( 1−pT[T≫A]G−pT[T≫G]G ) where trinucleotide frequencies were taken from the mutational profile . Amino acid substitutions corresponding to each missense mutation are calculated by translating the mutated and wild type codons using a standard codon table . Codon mutability strongly depends on the neighboring codons as illustrated in S5 Fig . Gene weights estimate a relative probability of a gene compared to other genes to be mutated in cancer through somatic mutagenesis . There are multiple ways the gene weights can be calculated: SNP-based weight was calculated using the number of SNPs in the vicinity of the gene of interest . We used the “EnsDb . Hsapiens . v86” database to find genomic coordinates of a gene , including introns , and extended the range in both 3’ and 5’ directions according to the window size ( S7 Table ) . We then counted the number of common SNPs from dbSNP database[51] within the genomic region . Gene weight was calculated as: ωgSNP=nSNPLwindow , where nSNP is the number of SNPs and Lwindow is the length of the genomic region in base pairs . We tested several window sizes for defining the genomic regions around the gene of interest ( S6 Table ) . Mutation-based weight was calculated using the number of nucleotide sites with reoccurring mutations counted only once to avoid the bias that may be present due to selection on individual sites: ωgsites=kgnk . Here kg is the number of mutated sites and nk is the number of base pairs in the gene transcript . Silent mutation-based weight was introduced previously and was shown to be superior in assessment of significant non-synonymous mutations across genes [52] . This weight can be calculated by taking into account only silent somatic mutations: ωgsilent=sgNLg . Here sg is the total number of somatic silent mutations within the gene , N is the number of tumor samples and Lg is the number of codons in the gene transcript . No-outlier-based weight introduced previously [21] takes into account the number of all codon mutations within a gene , Cg , excluding mutations in outlier codon sites bearing more than the 99th percentile of mutations of the gene: ωgout=CgNLg , normalized by gene length Lg in amino acids and the total number of samples N . Using gene weights , an adjusted probability per codon can be then expressed as: pM′=ωgpM ( 3 ) Similarly , per nucleotide probability can be calculated adjusted by gene weight . B-score uses the binomial model to calculate the probability of observing a certain type of mutation in a given site more frequently than k: Bscore=∑k=n+1N ( Nk ) pk ( 1−p ) N−k ( 4 ) where p=pM′ or p=pm′ and k is the number of observed mutations of a given type at a particular nucleotide or codon , N is a total number of cancer samples in a cohort . Depending on the dataset chosen or a particular cohort of patients ( for instance , corresponding to one cancer type ) , the total number of samples N and the numbers of observed mutations k will change . While ranking mutations in a given gene , Bscore can further be adjusted for multiple-testing with Benjamini-Hochberg correction as implemented on the MutaGene website . CanDrAplus34 program was downloaded and ran using default specifications with the “Cancer-in-General” annotation data file . REVEL33 predictions were obtained from dbNSFP database[53] . CHASMplus predictions were obtained using CRAVAT[54] . The pan-cancer model was used for CHASMplus . FatHMM35 cancer-associated scores were obtained from their webserver . Differences between various groups were tested with the Kruskal-Wallis , Dunn test , and Mann-Whitney-Wilcoxon tests implemented in R software . Dunn’s test is a non-parametric pairwise multiple comparisons procedure based on rank sums; it is used to infer difference between means in multiple groups and was used because it is relatively conservative post-hoc test for Kruskal-Wallis . Associations between mutability and observed frequency ( the number of individuals with a mutation in whole-exome/genome studies from COSMIC ) , was tested using Pearson as well as Spearman correlation tests since the variables were not normally distributed . To quantify the performance of scores , we performed Receiver Operating Characteristics ( ROC ) and precision-recall analyses . Sensitivity or true positive rate was defined as TPR = TP/ ( TP + FN ) and specificity was defined as SPC = TN/ ( FP+TN ) . Additionally , in order to account for imbalances in the labeled dataset , the quality of the predictions was described by Matthew’s correlation coefficient [55]: MCC=TP*TN−FP*FN√ ( TP+FP ) * ( TP+FN ) * ( TN+FP ) * ( TN+FN ) In order to compare across tools , the threshold which gave the maximum MCC was chosen for each tool to calculate TP , TN , FP , and FN . | Cancer development and progression is associated with accumulation of mutations . However , only a small fraction of mutations identified in a patient is responsible for cellular transformations leading to cancer . These so-called drivers characterize molecular profiles of tumors and could be helpful in predicting clinical outcomes for the patients . One of the major problems in cancer research is prioritizing mutations . Recurrence of a mutation in patients remains one of the most reliable markers of its driver status . However , DNA damage and repair processes do not affect the genome uniformly , and some mutations are more likely to occur than others . Moreover , mutational probability ( mutability ) varies with the cancer type . We developed models that adjust the number of mutation recurrences in patients by cancer-type specific background mutability in order to prioritize cancer mutations . Using a comprehensive experimental dataset , we found that mutability of driver mutations was lower than that of passengers , and consequently adjusting mutation recurrence frequency by mutability significantly improved ranking of mutations and driver mutation prediction . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"mutation",
"substitution",
"mutation",
"point",
"mutation",
"cancer",
"genetics",
"nonsense",
"mutation",
"database",
"and",
"informatics",
"methods",
"silent",
"mutation",
"genetics",
"biological",
"databases",
"biology",
"and",
"life",
"sciences",
"mutation",
"databa... | 2019 | Finding driver mutations in cancer: Elucidating the role of background mutational processes |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.